Dissertations / Theses on the topic 'Cognitive algorithm'

The heterogeneity and complexity of modern communication networks demands coupling network nodes with intelligence to perceive and adapt to different network conditions autonomously. Cognitive Networking is an emerging networking research area that aims to achieve this goal by applying distributed reasoning and learning across the protocol stack and throughout the network. Various cognitive node and cognitive network architectures with different levels of maturity have been proposed in the literature. All of them adopt the idea of coupling network devices with sensors to sense network conditions, artificial intelligence algorithms to solve problems, and a reconfigurable platform to apply solutions. However, little further research has investigated suitable reasoning and learning algorithms. In this dissertation, we take cognitive network research a step further by investigating the reasoning component of cognitive networks. In a deviation from previous suggestions, we suggest the use of a single flexible distributed reasoning algorithm for cognitive networks. We first propose an architecture for a cognitive node in a cognitive network that is general enough to apply to future networking challenges. We then introduce and justify our choice of the island genetic algorithm (iGA) as the distributed reasoning algorithm. Having introduced our cognitive node architecture, we then focus on the applicability of the island genetic algorithm as a single reasoning algorithm for cognitive networks. Our approach is to apply the island genetic algorithm to different single and cross layer communication and networking problems and to evaluate its performance through simulation. A proof of concept cognitive network is implemented to understand the implementation challenges and assess the island genetic algorithm performance in a real network environment. We apply the island genetic algorithm to three problems: channel allocation, joint power and channel allocation, and flow routing. The channel allocation problem is a major challenge for dynamic spectrum access which, in turn, has been the focal application for cognitive radios and cognitive networks. The other problems are examples of hard cross layer problems. We first apply the standard island genetic algorithm to a channel allocation problem formulated for the dynamic spectrum cognitive network environment. We also describe the details for implementing a cognitive network prototype using the universal software radio peripheral integrated with our extended implementation of the GNU radio software package and our island genetic algorithm implementation for the dynamic spectrum channel allocation problem. We then develop a localized variation of the island genetic algorithm, denoted LiGA, that allows the standard island genetic algorithm to scale and apply it to the joint power and channel allocation problem. In this context, we also investigate the importance of power control for cognitive networks and study the effect of non-cooperative behavior on the performance of the LiGA. The localized variation of the island genetic algorithm, LiGA, is powerful in solving node-centric problems and problems that requires only limited knowledge about network status. However, not every communication and networking problems can be solved efficiently in localized fashion. Thus, we propose a generalized version of the LiGA, namely the K-hop island genetic algorithm, as our final distributed reasoning algorithm proposal for cognitive networks. The K-hop island genetic algorithm is a promising algorithm to solve a large class of communication and networking problems with controllable cooperation and migration scope that allows for a tradeoff between performance and cost. We apply it to a flow routing problem that includes both power control and channel allocation. For all problems simulation results are provided to quantify the performance of the island genetic algorithm variation. In most cases, simulation and experimental results reveal promising performance for the island genetic algorithm. We conclude our work with a discussion of the shortcomings of island genetic algorithms without guidance from a learning mechanism and propose the incorporation of two learning processes into the cognitive node architecture to solve slow convergence and manual configuration problems. We suggest the cultural algorithm framework and reinforcement learning techniques as candidate leaning techniques for implementing the learning processes. However, further investigation and implementation is left as future work.
Ph. D.

Radio Communication technologies are undergoing drastic demand over the past two decades. The precious radio resource, electromagnetic radio spectrum, is in vain as technology advances. It is required to come up with a solution to improve its wise uses. Cognitive Radio enabled by Software-Defined Radio brings an intelligent solution to efficiently use the Radio Spectrum. It is a method to aware the radio communication system to be able to adapt to its radio environment like signal power and free spectrum holes. The approach will pose a question on how to efficiently detect a signal. In this thesis different spectrum sensing algorithm will be explained and a special concentration will be on new sensing algorithm based on the Eigenvalues of received signal. The proposed method adapts blind signal detection approach for applications that lacks knowledge about signal, noise and channel property. There are two methods, one is ratio of the Maximum Eigenvalue to Minimum Eigenvalue and the second is ratio of Signal Power to Minimum Eigenvalue. Random Matrix theory (RMT) is a branch of mathematics and it is capable in analyzing large set of data or in a conclusive approach it provides a correlation points in signals or waveforms. In the context of this thesis, RMT is used to overcome both noise and channel uncertainties that are common in wireless communication. Simulations in MATLAB and real-time measurements in LabVIEW are implemented to test the proposed detection algorithms. The measurements were performed based on received signal from an IF-5641R Transceiver obtained from National Instruments.

The internet and communication industry continue to develop new technologies rapidly, which has caused a boom in smart and networking device manufacturing. With new trends, operators are constantly battling towards deploying multiple systems to cater for the need of all users. The higher bandwidth utilization and flexibility demanded new networking solutions which paved way for Software Defined Network (SDN). SDN is centralized platform which works with other technologies such as Network Function Virtualization (NFV) to offer reliable, flexible and centrally controllable network solutions. It offers remote access control with logical design of the system, security and resource management. Traditional and new developing networks despite their advantages present numerous security challenges. With growing users worldwide, bandwidth related security risks such as Distributed Denial of Service (DDOS) are of grave concern. This encourages towards reliable and rapid response solutions such as Cognitive Algorithms (CA) which can adapt to a threat in real time environment. This dissertation proposes the use of CA to deploy security and mitigation measures against potential DDOS flooding attack to avoid network failure and memory depletion in SDN. The experiment done in proof of concept (PoC) provided proof of greater network resource utilization by limiting the attack while mitigation policies are implemented. It also shows that CA can adapt to growing and evolving network attack strength to counter as much as possible without the intervention of the operator. The work for future solutions based on CA and Artificial Intelligence (AI) for security have been established.

For complex computer networks with many tunable parameters and network performance objectives, the task of selecting the ideal network operating state is difficult. To improve the performance of these kinds of networks, this research proposes the idea of the cognitive network. A cognitive network is a network composed of elements that, through learning and reasoning, dynamically adapt to varying network conditions in order to optimize end-to-end performance. In a cognitive network, decisions are made to meet the requirements of the network as a whole, rather than the individual network components. We examine the cognitive network concept by first providing a definition and then outlining the difference between it and other cognitive and cross-layer technologies. From this definition, we develop a general, three-layer cognitive network framework, based loosely on the framework used for cognitive radio. In this framework, we consider the possibility of a cognitive process consisting of one or more cognitive elements, software agents that operate somewhere between autonomy and cooperation. To understand how to design a cognitive network within this framework we identify three critical design decisions that affect the performance of the cognitive network: the selfishness of the cognitive elements, their degree of ignorance, and the amount of control they have over the network. To evaluate the impact of these decisions, we created a metric called the price of a feature, defined as the ratio of the network performance with a certain design decision to the performance without the feature. To further aid in the design of cognitive networks, we identify classes of cognitive networks that are structurally similar to one another. We examined two of these classes: the potential class and the quasi-concave class. Both classes of networks will converge to Nash Equilibrium under selfish behavior and in the quasi-concave class this equilibrium is both Pareto and globally optimal. Furthermore, we found the quasi-concave class has other desirable properties, reacting well to the absence of certain kinds of information and degrading gracefully under reduced network control. In addition to these analytical, high level contributions, we develop cognitive networks for two open problems in resource management for self-organizing networks, validating and illustrating the cognitive network approach. For the first problem, a cognitive network is shown to increase the lifetime of a wireless multicast route by up to 125\%. For this problem, we show that the price of selfishness and control are more significant than the price of ignorance. For the second problem, a cognitive network minimizes the transmission power and spectral impact of a wireless network topology under static and dynamic conditions. The cognitive network, utilizing a distributed, selfish approach, minimizes the maximum power in the topology and reduces (on average) the channel usage to within 12\% of the minimum channel assignment. For this problem, we investigate the price of ignorance under dynamic networks and the cost of maintaining knowledge in the network. Today's computer networking technology will not be able to solve the complex problems that arise from increasingly bandwidth-intensive applications competing for scarce resources. Cognitive networks have the potential to change this trend by adding intelligence to the network. This work introduces the concept and provides a foundation for future investigation and implementation.
Ph. D.

With the increase of wireless devices, the wireless spectrum is becoming overloaded causing users to experience delays and performance degradation. Typically, a device will start transmitting data on a frequency and continue transmitting on that frequency regardless of the channel being overloaded or not. Some smarter devices such as routers are able to sense when their channel is becoming overloaded by observing delays and amount of devices transmitting on that frequency. Spectrum analyzers are usually very expensive and usually do not provide many functionalities other than analysis. Utilizing newer alternatives for sensing the spectrum such as Software Defined Radios (SDR) can address frequency allocation problems and allow users to decide the best frequency to use for communication. A promising SDR such as GNU Radio will be covered in this thesis, as well as the hardware components needed for its functionality. In this thesis, a cognitive radio approach is taken in designing a channel selection algorithm by scanning and monitoring the wireless spectrum on IEEE 802.11 b/g through the use of GNU Radio and USRP. Tests are performed as a proof of concept and to help future research with the use of cognitive radios.
Thesis (M.S.)--Wichita State University, College of Engineering, Dept. of Electrical Engineering.

With the ever increasing demand for wireless applications, current wireless systems are challenged to meet the higher data rate and higher reliability requirements. Although the current and future technological developments allow making these requirements reachable, some other resources remain limited. The radio spectrum is one such natural resource. Previous studies have shown that the radio spectrum is not efficiently utilized. Therefore, recent studies are focused on fully utilizing this unexpandable radio spectrum. Cognitive radio (CR) has emerged as a possible solution to improve the spectrum utilization by opportunistically exploiting the licenced users transmit spectrum in dynamically changing environments. On the other hand, the development of CR technology raises new challenges of proper design of transmission and receive schemes for CR to facilitate high data rate access and better performance along with high spectral efficiency. To achieve these objectives, in this thesis, advanced transceiver algorithms for CR physical layer are designed to improve the throughput and the error rate performance in hostile wireless channels. We first designed a linear precoder for orthogonal space-time block coded, orthogonal frequency division multiplexing (OFDM)-based multiple-input multiple-output antenna CR when operating in correlated Rayleigh fading channels. The linear precoder is designed by minimizing an upper bound on the average pairwise error probability, constrained to a set of per subcarrier power constraints at CR transmitter and a set of primary users interference power thresholds. An efficient algorithm is proposed to obtain the optimal precoder matrices. We then proposed a power allocation policy to achieve a lower-bound on the ergodic sum capacity of single-input single-output opportunistic spectrum sharing multiple access channel with imperfect channel estimates. An efficient algorithm is proposed to obtain the optimal power allocation for each CR transmitter. Finally, we proposed a blind parameter estimation algorithm for OFDM signal affected by a time-dispersive channel, carrier phase, timing offset, carrier frequency offset and additive Gaussian noise. The cyclostationarity properties of received OFDM signal in time-dispersive channel is exploited to estimate the OFDM parameters. These parameters includes OFDM symbol period, useful symbol period, cyclic prefix factor, number of subcarriers and carrier frequency offset.

Processing requirement of a complex autonomous robotic vehicle demands high efficiency in algorithmic and software execution. Today's advanced computer hardware technology provides processing capabilities that were not available a decade ago. There are still major space and time limitations on these technologies for autonomous robotic applications. Increasingly, small to miniature mobile robots are required for reconnaissance, surveillance, and hazardous material detections for military and industrial applications. The small sized autonomous mobile robotic applications have limited power capacity as well as memory and processing resources. A number of algorithms exist for producing optimal traverses given changing arc costs. One algorithm stands out as the most used algorithm in simple path finding applications such as games, named the A * algorithm. This dissertation investigated the hypothesis that cognitive based adaptive path planning algorithms are efficient. This assumption is based on the observed capability of biological systems, which ignore irrelevant information and quickly process non-optimum but efficient paths. Path planning function for all organisms from insects to humans is a critical function of survival, and living organisms perform it with graceful accuracy and efficiency. This hypothesis was tested by developing a Cognitive Based Adaptive Path Planning Algorithm (CBAPPA) and a limited simulation program to test the theory of the algorithm, and comparing the result with other known approaches. This dissertation presented a new cognitive based approach in solving the path planning problems for autonomous robotic applications. The goal of this paper was to show that adaptive cognitive based techniques are more efficient by comparing this paper's path planning approach to analytical and heuristic algorithms. This study presented a two-step methodology of Primary Path and Refined Path. Each step was implemented by a number of heuristic algorithms. This paper illustrated that the CBAPPA’s path-finding efficiency exceeds the efficiency of some popular analytical and heuristic approaches. This research paper concluded that the hypothesis was verified and cognitive based path planning algorithm is efficient and is a viable approach for autonomous robotic applications.

Cette thèse présente la formalisation et le développement d’un logiciel d'analyse musicale appelé "Multiscale Oracle Representations For Organized Sounds" (MORFOS). Ce logiciel vise à mettre en œuvre un modèle multi-échelle de la forme musicale basé sur l’Algorithme Cognitif de Jean-Marc Chouvel. Les travaux de cette thèse s'inscrivent dans la continuité de l'analyse cognitive en musicologie, visant à comprendre les processus cognitifs qui interviennent lors de l'écoute musicale. Nous étudions une représentation hiérarchisée de la musique et explorons l’influence de cette hiérarchisation dans l’organisation des événements musicaux dans le temps et la compréhension de la musique. Nous formalisons ainsi les concepts de matériau, d’objet et de diagramme formel, et nous introduisons le Diagramme Formel Multi-échelle qui décrit la structure musicale à différentes échelles temporelles et niveaux d'analyse. Celui-ci est composé de trois plans que nous introduisons : la forme, la structure et l’organisation. L'implémentation informatique de MORFOS a été réalisée en Python et accepte des représentations audio, symboliques ou vectorielles. Ce logiciel présente une architecture modulaire intégrant différents modules de traitement audio, de classification et de segmentation : nous présentons ainsi différentes mesures implémentées sous forme d’un ensemble de règles et discutons des contraintes associées à l’étude de la classification et de la segmentation à partir d’une représentation audio. Nous présentons ainsi la notion d’Agenda, qui correspond au choix par l’utilisateur d’un ensemble de règles permettant de représenter un modèle « d’écoute » pour l’analyse d’une œuvre musicale par le logiciel. La thèse explore également la question de la complexité de la structure musicale : nous proposons l’expression d’un coût associé à la description de l’objet musical acquis en fonction de son contexte, selon la définition de Kolmogorov. Nous cherchons également à comparer le comportement du logiciel MORFOS avec les phénomènes d'attention et la charge cognitive lors de l'écoute musicale. Une expérience visant à mesurer la charge cognitive pendant la tâche de segmentation musicale a ainsi été conçue. Cette thèse présente par ailleurs des réflexions sur la visualisation des diagrammes formels multi-échelles. A cette occasion, nous avons développé une interface permettant de rendre le logiciel accessible à tous les utilisateurs. Enfin, des exemples d'analyses musicales réalisées avec MORFOS sont présentées, sur une base de données musicales pop ainsi qu’un corpus d'œuvres classiques
This thesis presents the formalization and development of a music analysis software called "Multiscale Oracle Representations For Organized Sounds" (MORFOS). This software aims to implement a multi-scale model of musical form based on Jean-Marc Chouvel's Cognitive Algorithm. The work in this thesis is part of the cognitive analysis in musicology, aimed at understanding the cognitive processes involved in listening to music. We study a hierarchical representation of music and explore the influence of this hierarchy on the organization of musical events over time and on musical comprehension. We formalize the concepts of material, object, and formal diagram, and introduce the Multi-scale Formal Diagram, which describes musical structure at different temporal scales and levels of analysis. This comprises three planes, which we introduce: form, structure, and organization. MORFOS has been implemented in Python and accepts audio, symbolic, and vector representations. This software features a modular architecture integrating different modules for audio processing, classification, and segmentation: we present different measures implemented in the form of a set of rules and discuss the constraints associated with the study of classification and segmentation based on an audio representation. We also introduce the notion of Agenda, which corresponds to the user's choice of a set of rules to represent a "listening" model for the software's analysis of a musical work. The thesis also explores the question of the complexity of the musical structure: we propose the expression of a cost associated with the description of the acquired musical object depending on its context, according to Kolmogorov's definition. We also seek to compare the behavior of MORFOS software with attentional phenomena and cognitive load during musical listening. An experiment designed to measure cognitive load during the musical segmentation task has thus been devised. This thesis also presents reflections on the visualization of multi-scale formal diagrams. To this end, we have developed an interface to make the software accessible to all users. Finally, examples of musical analyses carried out with MORFOS are presented, on a pop music database and a corpus of classical works

Dynamic Spectrum Access (DSA) and multi-input multi-output (MIMO) communications are among the most promising solutions to address the ever-increasing wireless demand. Cognitive radio (CR) is the enabling technology for DSA. In this dissertation, we propose several resource allocation strategies for multiuser and cooperative MIMO communications in the context of DSA/CR systems and wireless sensor networks (WSNs). First, to maximize the Cognitive MIMO (CMIMO) network throughput, we develop a low-complexity distributed algorithm that configures the transmit antenna radiation directions and allocates power to all data streams over both frequency and space/antenna dimensions. We formulate the joint power, spectrum allocation, and MIMO beamforming problem as a noncooperative game. We prove that the game always admits at least one Nash Equilibrium (NE). To improve the efficiency of this NE (i.e., network throughput), we derive user-dependent pricing policies that force MIMO transmitters to steer their beams away from nearby unintended receivers. Second, we propose beamforming games (with and without pricing policies) that jointly improve the power and spectrum efficiency while meeting various rate demands. We derive sufficient conditions under which a given rate-demand profile can be supported. To account for user fairness, we develop a channel assignment and power allocation mechanism based on the Nash Bargaining solution. The proposed scheme allows CMIMO links to first propose their rate demands, and then cooperate and bargain in the process of determining their channel assignment, power allocation, and "precoding" matrices. In the context of WSNs where energy efficiency is a key design metric, we propose a cooperative MIMO framework. The framework partitions a WSN into various clusters in which several single-antenna sensors cooperate and form a virtual MIMO node so as to conserve power through harvesting MIMO's diversity gain. Extensive simulations show that our proposed schemes achieve significant throughput and energy efficiency improvement compared with state-of-the-art designs.

A fuzzy cognitive map (FCM) is made up of factors and direct impacts. In graph theory, a bipolar weighted digraph is used to model an FCM; its vertices represent the factors, and the arcs represent the direct impacts. Each direct impact is either positive or negative, and is assigned a weight; in the model considered in this thesis, each weight is interpreted as the probability of the impact. A directed walk from factor F to factor F' is interpreted as an indirect impact of F on F'. The probabilistic transitive closure (PTC) of an FCM (or bipolar weighted digraph) is a bipolar weighted digraph with the same set of factors, but with arcs corresponding to the indirect impacts in the given FCM. Fuzzy cognitive maps can be used to represent structured knowledge in diverse fields, which include science, engineering, and the social sciences. In [P. Niesink, K. Poulin, M. Sajna, Computing transitive closure of bipolar weighted digraphs, Discrete Appl. Math. 161 (2013), 217-243], it was shown that the transitive closure provides valuable new information for its corresponding FCM. In particular, it gives the total impact of each factor on each other factor, which includes both direct and indirect impacts. Furthermore, several algorithms were developed to compute the transitive closure of an FCM. Unfortunately, computing the PTC of an FCM is computationally hard and the implemented algorithms are not successful for large FCMs. Hence, the Reduction-Recovery Algorithm was proposed to make other (direct) algorithms more efficient. However, this algorithm has never been implemented before. In this thesis, we code the Reduction-Recovery Algorithm and compare its running time with the existing software. Also, we propose a new enhancement on the existing PTC algorithms, which we call the Separation-Reduction Algorithm. In particular, we state and prove a new theorem that describes how to reduce the input digraph to smaller components by using a separating vertex. In the application part of the thesis, we show how the PTC of an FCM can be used to compare different standpoints on the issue of work-integrated learning.

This essay aims to study the relationship between cosplay content creators and TikTok’s algorithm. To study this relationship the essay will conduct a qualitative semi-structured interviews with creators from the cosplay community on TikTok. Since the rise of digital plattforms the media and the role of producer as well as consumer has changed drastically. TikTok has been growing rapidly in popularity since its entry on the market, and by 2020 it had 500 million active users. Since many of today's digital platforms have consumer produced content, the consumer of today has taken on a mixed role between consuming and creating content, which creates a new relationship. The content consumers produce vary vastly on TikTok but one kind that has been present in much of TikTok’s existence is cosplay content. Cosplayers are creators who design costumes to already established characters or franchises. Since a discourse has started in the cosplay community on TikTok about the algorithm suppressing their content the study found it to be a good way to start examining content creators as individuals and how they behave towards an algorithm in their content creation process. The study aims to use algorithmic “folk” theory to examine what theories have been created in the community and how the theories affect the creators. The study also applies gatekeeping theory and social cognitive theory (SCT) to paint a clearer picture in how these creators view the algorithm. Seven interviews with cosplay content creators were conducted and with the help of a thematic analysis method the study found several themes in how the creators view and behave in relation to TikTok and its algorithm. The results of our study shows that there’s a definite present of “folk” theories created inside of the community. The most distinct behaviour relating to “folk” theory among the creators was that they can’t use the hashtag cosplay in the belief that the algorithm would suppress the content. This study concludes that the creators are more aware of the algorithm then they themself know and have different ways of working with and around it.

This research provides a robust solution based on artificial intelligence (AI) techniques to overcome the challenges in Software Defined Networks (SDNs) that can jeopardise the overall performance of the network. The proposed approach, presented in the form of an intelligent agent appended to the SDN network, comprises of a new hybrid intelligent mechanism that optimises the performance of SDN based on heuristic optimisation methods under an Artificial Neural Network (ANN) paradigm. Evolutionary optimisation techniques, including Particle Swarm Optimisation (PSO) and Genetic Algorithms (GAs) are deployed to find the best set of inputs that give the maximum performance of an SDN-based network. The ANN model is trained and applied as a predictor of SDN behaviour according to effective traffic parameters. The parameters that were used in this study include round-trip time and throughput, which were obtained from the flow table rules of each switch. A POX controller and OpenFlow switches, which characterise the behaviour of an SDN, have been modelled with three different topologies. Generalisation of the prediction model has been tested with new raw data that were unseen in the training stage. The simulation results show a reasonably good performance of the network in terms of obtaining a Mean Square Error (MSE) that is less than 10−6 [superscript]. Following the attainment of the predicted ANN model, utilisation with PSO and GA optimisers was conducted to achieve the best performance of the SDN-based network. The PSO approach combined with the predicted SDN model was identified as being comparatively better than the GA approach in terms of their performance indices and computational efficiency. Overall, this research demonstrates that building an intelligent agent will enhance the overall performance of the SDN network. Three different SDN topologies have been implemented to study the impact of the proposed approach with the findings demonstrating a reduction in the packets dropped ratio (PDR) by 28-31%. Moreover, the packets sent to the SDN controller were also reduced by 35-36%, depending on the generated traffic. The developed approach minimised the round-trip time (RTT) by 23% and enhanced the throughput by 10%. Finally, in the event where SDN controller fails, the optimised intelligent agent can immediately take over and control of the entire network.

Cognitive Radio is an innovative technology that allows the secondary unlicensed users to share the spectrum with licensed primary users to utilize the spectrum. For maximum utilization of spectrum, in cognitive radio network spectrum sensing is an important issue. Cognitive user under extreme shadowing and channel fading can‟t sense the primary licensed user signal correctly and thus to improve the performance of spectrum sensing, collaboration between secondary unlicensed users is required. In collaborative spectrum sensing the observation of each secondary user is received by a base station acting as a central entity, where a final conclusion about the presence or absence of the primary user signal is made using a particular decision and fusion rule. Due to spatially correlated shadowing the collaborative spectrum sensing performance decreases, and thus optimum secondary users must be selected to, not only improve spectrum sensing performance but also lessen the processing overhead of the central entity. A particular situation is depicted in the project where according to some performance parameters, first those optimum secondary users that have enough spatial separation and high average received SNR are selected using Genetic Algorithm, and then collaboration among these optimum secondary users is done to evaluate the performance. The collaboration of optimal secondary user providing high probability of detection and low probability of false alarm, for sensing the spectrum is compared with the collaboration of all the available secondary users in that radio environment. At the end a conclusion has been made that collaboration of selected optimum secondary users provides better performance, then the collaboration of all the secondary users available.
Cognitive Radio is an innovative technology that allows the secondary unlicensed users to share the spectrum with licensed primary users to utilize the spectrum. For maximum utilization of spectrum, in cognitive radio network spectrum sensing is an important issue. Cognitive user under extreme shadowing and channel fading can‟t sense the primary licensed user signal correctly and thus to improve the performance of spectrum sensing, collaboration between secondary unlicensed users is required. In collaborative spectrum sensing the observation of each secondary user is received by a base station acting as a central entity, where a final conclusion about the presence or absence of the primary user signal is made using a particular decision and fusion rule. Due to spatially correlated shadowing the collaborative spectrum sensing performance decreases, and thus optimum secondary users must be selected to, not only improve spectrum sensing performance but also lessen the processing overhead of the central entity. A particular situation is depicted in the project where according to some performance parameters, first those optimum secondary users that have enough spatial separation and high average received SNR are selected using Genetic Algorithm, and then collaboration among these optimum secondary users is done to evaluate the performance. The collaboration of optimal secondary user providing high probability of detection and low probability of false alarm, for sensing the spectrum is compared with the collaboration of all the available secondary users in that radio environment. At the end a conclusion has been made that collaboration of selected optimum secondary users provides better performance, then the collaboration of all the secondary users available.

Fueled by the rapid advancement in digital and wireless technologies, the ever-increasing capabilities of wireless devices have placed upon us a tremendous challenge - how to put all of this capability to effective use. Individually, wireless devices have outpaced the ability of users to optimally configure them. Collectively, the complexity is far more daunting. Research in cognitive networks seeks to provide a solution to the diffculty of effectively using the expanding capabilities of wireless networks by embedding greater degrees of intelligence within the network itself. In this dissertation, we address some fundamental questions related to cognitive networks, such as "What is a cognitive network?" and "What methods may be used to design a cognitive network?" We relate cognitive networks to a common artificial intelligence (AI) framework, the multi-agent system (MAS). We also discuss the key elements of learning and reasoning, with the ability to learn being the primary differentiator for a cognitive network. Having discussed some of the fundamentals, we proceed to further illustrate the cognitive networking principle by applying it to two problems: multichannel topology control for dynamic spectrum access (DSA) and routing in a mobile ad hoc network (MANET). The multichannel topology control problem involves confguring secondary network parameters to minimize the probability that the secondary network will cause an outage to a primary user in the future. This requires the secondary network to estimate an outage potential map, essentially a spatial map of predicted primary user density, which must be learned using prior observations of spectral occupancy made by secondary nodes. Due to the complexity of the objective function, we provide a suboptimal heuristic and compare its performance against heuristics targeting power-based and interference-based topology control objectives. We also develop a genetic algorithm to provide reference solutions since obtaining optimal solutions is impractical. We show how our approach to this problem qualifies as a cognitive network. In presenting our second application, we address the role of network state observations in cognitive networking. Essentially, we need a way to quantify how much information is needed regarding the state of the network to achieve a desired level of performance. This question is applicable to networking in general, but becomes increasingly important in the cognitive network context because of the potential volume of information that may be desired for decision-making. In this case, the application is routing in MANETs. Current MANET routing protocols are largely adapted from routing algorithms developed for wired networks. Although optimal routing in wired networks is grounded in dynamic programming, the critical assumption, static link costs and states, that enables the use of dynamic programming for wired networks need not apply to MANETs. We present a link-level model of a MANET, which models the network as a stochastically varying graph that possesses the Markov property. We present the Markov decision process as the appropriate framework for computing optimal routing policies for such networks. We then proceed to analyze the relationship between optimal policy and link state information as a function of minimum distance from the forwarding node. The applications that we focus on are quite different, both in their models as well as their objectives. This difference is intentional and signficant because it disassociates the technology, i.e. cognitive networks, from the application of the technology. As a consequence, the versatility of the cognitive networks concept is demonstrated. Simultaneously, we are able to address two open problems and provide useful results, as well as new perspective, on both multichannel topology control and MANET routing. This material is posted here with permission from the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of Virginia Tech library's products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to pubs-permissions@ieee.org. By choosing to view this material, you agree to all provisions of the copyright laws protecting it.
Ph. D.

This thesis aims to clearly describe the cognitive radio and its components and operations. Moreover, it aims on describing the expected outcome from the most common techniques that are proposed for use in cognitive radios. In addition, it describes the basic principles of game theory and some simple game models that can be used to analyze the efficiency of the optimization algorithms. Furthermore, it investigates the use of load balancing algorithm and genetic algorithm in optimizing the decision making operation in cognitive radios. Matlab software simulations were carried out and the results show the promising benefit of using those two algorithms along with game theory in optimizing the dynamic spectrum allocation process.

Recent advancements in frequency-agile radio technology and dynamic spectrum access network have created a huge space for improving the utilization efficiency of wireless spectrum. Existing algorithms and protocols, however, have not taken full advantage of the new technologies due to obsolete network design ideologies inherited from conventional network design, such as static spectrum access and static channelization. In this dissertation, we propose new resource management models and algorithms that capitalize on the frequencyagility of next generation radios and the dynamic spectrum access concepts to increase the utilization efficiency of wireless spectrum. We first propose a new analytical model for Dynamic Spectrum Access (DSA) networks. Compared to previous models, the new model is able to include essential DSA mechanisms such as spectrum sensing and primary interference avoidance into solid mathematical representation and thus drastically increase the accuracy of our model. The subsequent numerical study conforms well with existing empirical studies and provides fundamental insights on the design of future DSA networks. We then take advantage of partially overlapped channel in frequency-agile radio networks and propose simple joint channel scheduling and flow routing optimization algorithm that maximizes network throughput. The model quantifies the impact of fundamental network settings, such as node density and traffic load, on the performance of partially overlapped channel based networks. We then propose a cross-layer radio resource allocation algorithm JSSRC (Joint Spectrum Sharing and end-to-end data Rate Control) that iteratively adapts a frequency-agile radio network to optimum with regard to aggregate network spectrum utilization. Subsequently, we extend JSSRC to include routing and present TRSS (joint Transport, Routing and Spectrum Sharing) to solve the much more complex joint transport, routing and spectrum sharing optimization problem. Both JSSRC and TRSS enjoy theoretical convergence and achieve optimum with appropriate scheduling algorithms. The works together strive to improve efficiency of spectrum utilization in frequency-agile radio networks. Numerical and simulation studies show the effectiveness of our designs to reduce the so-called spectrum shortage problem.
Ph. D.

The federal communications commission (FCC) has predicted that, under the current regulatory environment, a spectrum shortage may be faced in the near future. This impending spectrum shortage is in part due to a rapidly increasing demand for wireless services and in part due to inefficient usage of currently licensed bands. A new paradigm pertaining to wireless spectrum allocation, known as cognitive radio (CR), has been proposed as a potential solution to this problem. This dissertation seeks to contribute to research in the field of CR through an investigation into the effect that a primary user (PU) channel occupancy model will have on the performance of a secondary user (SU) in a CR network. The model assumes that PU channel occupancy can be described as a binary process and a two state Hidden Markov Model (HMM) was thus chosen for this investigation. Traditional algorithms for training the model were compared with certain evolutionary-based training algorithms in terms of their resulting prediction accuracy and computational complexity. The performance of this model is important since it provides SUs with a basis for channel switching and future channel allocations. A CR simulation platform was developed and the results gained illustrated the effect that the model had on channel switching and the subsequently achievable performance of a SU operating within a CR network. Performance with regard to achievable SU data throughput, PU disruption rate and SU power consumption, were examined for both theoretical test data as well as data obtained from real world spectrum measurements (taken in Pretoria, South Africa). The results show that a trade-off exists between the achievable SU throughput and the average PU disruption rate. Significant SU performance improvements were observed when prediction modelling was employed and it was found that the performance and complexity of the model were influenced by the algorithm employed to train it. SU performance was also affected by the length of the quick sensing interval employed. Results obtained from measured occupancy data were comparable with those obtained from theoretical occupancy data with an average percentage similarity score of 96% for prediction accuracy (using the Viterbi training algorithm), 90% for SU throughput, 83% for SU power consumption and 71% for PU disruption rate.
Dissertation (MEng)--University of Pretoria, 2012.
Electrical, Electronic and Computer Engineering
unrestricted

Cognitive radio (CR) technology introduces a revolutionary wireless communication mechanism in terminals and network segments, so that they are able to learn their environment and adapt intelligently to the most appropriate way of providing the service for the user's exact need. By supporting multi-band, mode-mode cognitive applications, the cognitive radio addresses an interactive way of managing the spectrum that harmonizes technology, market and regulation. This dissertation gives a complete story of building a cognitive radio. It goes through concept clarification, architecture definition, functional block building, system integration, and finally to the implementation of a fully-functional cognitive radio node prototype that can be directly packaged for application use. This dissertation starts with a comprehensive review of CR research from its origin to today. Several fundamental research issues are then addressed to let the reader know what makes CR a challenging and interesting research area. Then the CR system solution is introduced with the details of its hierarchical functional architecture called the Egg Model, modular software system called the cognitive engine, and the kernel machine learning mechanism called the cognition cycle. Next, this dissertation discusses the design of specific functional building blocks which incorporate environment awareness, solution making, and adaptation. These building blocks are designed to focus on the radio domain that mainly concerns the radio environment and the radio platform. Awareness of the radio environment is achieved by extracting the key environmental features and applying statistical pattern recognition methods including artificial neural networks and k-nearest neighbor clustering. Solutions for the radio behavior are made according to the recognized environment and the previous knowledge through case based reasoning, and further adapted or optimized through genetic algorithm solution search. New experiences are gained through the practice of the new solution, and thus the CR's knowledge evolves for future use; therefore, the CR's performance continues improving with this reinforcement learning approach. To deploy the solved solution in terms of the radio's parameters, a platform independent radio interface is designed. With this general radio interface, the algorithms in the cognitive engine software system can be applied to various radio hardware platforms. To support and verify designed cognitive algorithms and cognitive functionalities, a complete reconfigurable SDR platform, called the CWT2 waveform framework, is designed in this dissertation. In this waveform framework, a hierarchical configuration and control system is constructed to support flexible, real-time waveform reconfigurability. Integrating all the building blocks described above allows a complete CR node system. Based on this general CR node structure, a fully-functional Public Safety Cognitive Radio (PSCR) node is prototyped to provide the universal interoperability for public safety communications. Although the complete PSCR node software system has been packaged to an official release including installation guide and user/developer manuals, the process of building a cognitive radio from concept to a functional prototype is not the end of the CR research; on-going and future research issues are addressed in the last chapter of the dissertation.
Ph. D.

In the literature, Cognitive radio (CR) as well as full-duplex (FD) communication technologies are proposed to increase the spectrum efficiency. The main contribution of this thesis is to introduce prediction and estimation techniques with low control overhead, and use the predicted or estimated information in resource allocation in CR networks, both in the overlay networks and the underlay networks. Prediction and estimation are important in increasing the data rate and keeping the interference at a low level. In the overlay scheme, I modeled the primary user (PU) traffic characteristics of the channels using the Probabilistic Suffix Tree (PST) algorithm. Then using this PST algorithm, I introduced a frequency hopping based control channel and derived its theoretical properties. Then I proposed two methods for selecting a channel set for transmission, which took into account both the PU channel usage statistics and, secondary user (SU) channel usage statistics as perceived by an SU of interest. The first scheme selected channels having the highest probability of successful transmission, while the second calculated a net reward using a marked Markov chain. Then using simulations, I showed that our scheme caused acceptable interference to the PUs and has better throughput performance, compared to a scheme selecting channels randomly. Then I proposed two joint channel assignment and power allocation schemes for a bi-directional FD underlay CR network with network assistance. The first scheme used the information on the number of total SU pairs present in the network. In the second scheme, I used least squares based estimation and Kalman filtering to estimate the interference at the monitoring stations using the local interference. It reduced the control overhead of keeping track of active SUs. In both of these schemes each SU pair decided on the channels to be used in the half-duplex mode and the full-duplex mode using local information. This joint optimization was done running channel assignment and power allocation algorithms alternatively. In the power allocation problem, I used a technique called monotonic optimization. After simulating both of these schemes I showed that the scheme based on estimation performs satisfactorily given that it has less control overhead.
October 2016

[Truncated abstract] Online games are an exciting new trend in the consumption of entertainment and provide the opportunity to examine selected antecedents of online game-playing based on studying the cognitive, experiential and habitual factors. This study was divided into two parts. The first part analysed the structural relations among research variables that might explain online game-playing using the Structural Equation Modeling (SEM) techniques. These analyses were conducted on a final sample of 218 online gamers. Specific issues examined were: If the variables of Perceived Game Performance, Satisfaction, Hedonic Responses, Flow and Habit Strength influence the Intention to Replay an online game. The importance of factors such as Hedonic Responses and Flow on Satisfaction in online game play. In addition to the SEM, analyses of the participants? reported past playing behaviour were conducted to test whether past game play was simply a matter of random frequency of past behaviour, or followed the specific pattern of the Negative Binomial Distribution (NBD). … The playing-time distribution was not significantly different to the Gamma distribution, in which the largest number of gamers plays for a short time (light gamers) and only a few gamers account for a large proportion of playing time (heavy gamers). Therefore, the reported time play followed a simple and predictable NBD pattern (Chisquare=. 390; p>.05). This study contributes to knowledge in the immediate field of online games and to the wider body of literature on consumer research. The findings demonstrate that gamers tend to act habitually in their playing behaviour. These findings support the argument that past behaviour (habit) is a better explanation of future behaviour than possible cognitive and affective explanations, especially for the apparent routinesed behaviour pattern on online games. The pattern of online game-playing is consistent with the finding of the NBD pattern in television viewing, in which the generalisability of the NBD model has been found in stable environments of repetitive behaviour. This supports the application of the NBD to areas beyond those of patterns in gambling and the purchase of consumer items. The findings have implications both for managerial and public policy decision-making.

Comment l'accès au savoir peut-il être impacté par la technologie ? Il suffit d'observer le virage intenté par les outils de communication au début des années 2000 pour se rendre compte : convergence des médias, pratiques participatives et numérisation massive des données. Dans ce contexte, on imagine que l'accès au savoir tend à se démocratiser. En effet, les individus semblent se réapproprier les espaces de vie, en inversant le modèle de transmission top-down, qui va du producteur vers le consommateur, au profit de processus de transfert basés sur l'intelligence collective. Pourtant, on aurait tort de réduire cette réorganisation à un simple renversement du modèle. Car l'intelligence collective est encline à divers biais cognitifs et socio-cognitifs, amenant parfois vers des situations irrationnelles. Autrefois, on s’accommodait de ces mécaniques sociales aux conséquences limitées, aujourd'hui les savoirs numérisés constituent des ensembles massivement communiquant, donnant naissance à de nouvelles voies d'accès et à de nouveaux clivages. Pourquoi ce savoir qui n'a jamais été aussi massif et ouvert, se révèle-t-il si sélectif ? Je propose d'explorer ce paradoxe. L'enregistrement massif et constant de nos traces numériques et l'hyper-connexion des individus, participent à la construction de structures organisationnelles, où se retrouvent numérisées de manière complexe, une partie des dynamiques sociales. En formalisant de la sorte les voies navigables, ces structures organisationnelles façonnent nos trajectoires. Sur cette base, les informaticiens ont mis au point des algorithmes de parcours individualisés, ayant pour objectifs de prédire et de recommander. Ainsi, on propose d'automatiser l'accès au savoir. Se pose alors la question de la gouvernance des individus, dans un contexte où l'intelligence collective est soumise à l'infrastructure : enregistrement des traces, composition des structures organisationnelles et algorithmes de parcours
How access to knowledge can be impacted by Information Technology? In the earlier 2000s, communication tools caused a significant turn : media convergence, participative practices and massive data. In this way, free access to knowledge might tend to be democratized. People seem to regain spaces, reversing traditional top-down model, going from producer to consumer, for the benefit of an horizontal model based on collective intelligence. However, it should not automatically be assumed that this leads to a simple model reversing. Collective intelligence is subject to cognitive biases, leading to potential irrational situations. Formerly, those social mechanisms had limited consequences. Nowadays, digital knowledge are massive communicating spaces, giving birth to new access paths and new cleavages. Why this massive and open knowledge, is actually so selective? I propose to explore this paradox. Massive and constant tracking of traces and individuals hyper-connection, these two facts help organizational structures design, where social dynamics are digitalized in a complex way. These structures formalize human trajectories. On this basis, computer scientists set up prediction algorithms and recommender engines. This way, knowledge access is automatized. It can then be asked about people governance, in this context of infrastructure submission: recording traces, designing knowledge structure and automating algorithms

Les futures applications de radio cognitive requièrent des systèmes de numérisation capables de convertir alternativement ou simultanément soit une bande très large avec une faible résolution soit une bande plus étroite avec une meilleure résolution, ceci de manière versatile (i.e. par contrôle logiciel). Pour cela, les systèmes de numérisation basés sur les Bancs de Filtres Hybrides (BFH) sont une solution attractive. Ils se composent d'un banc de filtres analogiques, un banc de convertisseurs analogique-numérique et un banc de filtres numériques. Cependant, ils sont très sensibles aux imperfections analogiques. L'objectif de cette thèse était de proposer et d'étudier une méthode de calibration qui permette de corriger les erreurs analogiques dans la partie numérique. De plus, la méthode devait être implémentable dans un système embarqué. Ce travail a abouti à une nouvelle méthode de calibration de BFH utilisant une technique d'Égalisation Adaptative Multi-Voies (EAMV) qui ajuste les coefficients des filtres numériques par rapport aux filtres analogiques réels. Cette méthode requiert d'injecter un signal de test connu à l'entrée du BFH et d'adapter la partie numérique afin de reconstruire le signal de référence correspondant. Selon le type de reconstruction souhaité (d'une large-bande, d'une sous-bande ou d'une bande étroite particulière), nous avons proposé plusieurs signaux de test et de référence. Ces signaux ont été validés en calculant les filtres numériques optimaux par la méthode de Wiener-Hopf et en évaluant leurs performances de ces derniers dans le domaine fréquentiel. Afin d'approcher les filtres numériques optimaux avec une complexité calculatoire minimum, nous avons implémenté un algorithme du gradient stochastique. La robustesse de la méthode a été évaluée en présence de bruit dans la partie analogique et de en tenant compte de la quantification dans la partie numérique. Un signal de test plus robuste au bruit analogique a été proposé. Les nombres de bits nécessaires pour coder les différentes données dans la partie numérique ont été dimensionnés pour atteindre les performances visées (à savoir 14 bits de résolution). Ce travail de thèse a permis d'avancer vers la réalisation des futurs systèmes de numérisation basés sur les BFH.

Ad hoc networks are emerging as a cost-effective, yet, powerful tool for communication. These systems, where networks can emerge and converge on-the-fly, are guided by the forward-looking goals of providing ubiquitous connectivity and constant access to information. Due to power and bandwidth constraints, the vulnerability of the wireless medium, and the multi-hop nature of ad hoc networks, these networks are becoming increasingly complex dynamic systems. Besides, modern radios are empowered to be reconfigurable, which harbors the temptation to exploit the system. To understand the implications of these issues, some of which pose significant challenges to efficient network design, we study topology control using game theory. We develop a game-theoretic framework of topology control that broadly captures the radio parameters, one or more of which can be tuned under the purview of topology control. In this dissertation, we consider two parameters, viz. transmit power and channel, and study the impact of controlling these on the emergent topologies. We first examine the impact of node selfishness on the network connectivity and energy efficiency under two levels of selfishness: (a) nodes cooperate and forward packets for one another, but selfishly minimize transmit power levels and; (b) nodes selectively forward packets and selfishly control transmit powers. In the former case, we characterize all the Nash Equilibria of the game and evaluate the energy efficiency of the induced topologies. We develop a better-response-based dynamic that guarantees convergence to the minimal maximum power topology. We extend our analysis to dynamic networks where nodes have limited knowledge about network connectivity, and examine the tradeoff between network performance and the cost of obtaining knowledge. Due to the high cost of maintaining knowledge in networks that are dynamic, mobility actually helps in information-constrained networks. In the latter case, nodes selfishly adapt their transmit powers to minimize their energy consumption, taking into account partial packet forwarding in the network. This work quantifies the energy efficiency gains obtained by cooperation and corroborates the need for incentivizing nodes to forward packets in decentralized, energy-limited networks. We then examine the impact of selfish behavior on spectral efficiency and interference minimization in multi-channel systems. We develop a distributed channel assignment algorithm to minimize the spectral footprint of a network while establishing an interference-free connected network. In spite of selfish channel selections, the network spectrum utilization is shown to be within 12% of the minimum on average. We then extend the analysis to dynamic networks where nodes have incomplete network state knowledge, and quantify the price of ignorance. Under the limitations on the number of available channels and radio interfaces, we analyze the channel assignment game with respect to interference minimization and network connectivity goals. By quantifying the interference in multi-channel networks, we illuminate the interference reduction that can be achieved by utilizing orthogonal channels and by distributing interference over multiple channels. In spite of the non-cooperative behavior of nodes, we observe that the selfish channel selection algorithm achieves load balancing. Distributing the network control to autonomous agents leaves open the possibility that nodes can act selfishly and the overall system is compromised. We advance the need for considering selfish behavior from the outset, during protocol design. To overcome the effects of selfishness, we show that the performance of a non-cooperative network can be enhanced by appropriately incentivizing selfish nodes.
Ph. D.

[EN] Nowadays there are many works which analyze and seek to improve the performance of Wireless Local Area Networks (WLANs) from different perspectives. A great deal of them is focused on design aspects, such as frequency distribution or channel assignment. Therefore, as these features have already been widely studied, my efforts have been directed to study other conditions that also could affect their performance and that have not been analyzed in depth yet. The main goal of this Ph.D. dissertation has been to perform a detailed study that researches the weather's impact on the performance of WLANs IEEE 802.11b/g. Two different WLAN scenarios have been analyzed to validate the results and to find precise relations. From conclusions of these previous analysis, the second objective has been to design a cognitive protocol that based on weather conditions and network performance parameters, allows networks to adjust their transmission features in order to overcome such impact. In order to conduct this study, firstly it was necessary to study which statistical methods could be used to extract the level of correlation between performance parameters of networks and weather conditions running at the same time. Secondly, I had to know which performance parameters the outdoor WLAN of Universitat Politècnica de València (UPV) could provide, and select them according to my objective. Then, I defined the period of time in which these parameters were gathered periodically. The next step was to select and collect the weather conditions from a close weather station during the same period of time. Finally, I had to perform a detailed pre-processing to put all of the volume of data in order and data were statistically analyzed. Results were successful; however there were several problems due to the variability derived from a real WLAN scenario. Therefore, an experimental setup was required in order to check the obtained results. It entailed to design and to develop an outdoor point-to-multipoint IEEE 802.11b/g link and to analyze again the weather's impact. Multiple points were considered in order to take into account different distances in the performed evaluation and to examine the behavior of different modulation schemes working under the same weather conditions. From these results, a cognitive algorithm was designed in order to reduce the weather's impact on IEEE 802.11b/g networks. One key aspect was to ensure it was energy efficient. This protocol was simulated and the obtained results were satisfactory in terms of both energy efficiency and network performance. To conclude, other external factor to WLANs studied in this Ph.D thesis has been the specific absorption rate. It deals with a current public health worry because it is used to measure the body tissue exposure to electromagnetic fields. Obviously, signal absorption by human bodies affects to the performance of WLANs and so, this parameter should be also taken into account when deploying efficient networks. For this reason, this study has been also included in this thesis.
[ES] Hoy en día existen muchos trabajos que analizan e intentan mejorar el rendimiento de las redes de área local inalámbricas desde diferentes perspectivas. Gran parte de estos trabajos se centran en aspectos de diseño, como son la distribución de frecuencias o la asignación de canales. Por lo tanto, como estos aspectos ya han sido ampliamente estudiados, los esfuerzos de esta tesis se han dirigido a estudiar otros factores que también podrían afectar a su rendimiento y que no han sido analizadas en profundidad todavía. El objetivo principal de esta tesis doctoral ha sido realizar un estudio detallado que analice el impacto de las condiciones meteorológicas sobre el rendimiento de las redes IEEE 802.11b/g. Para realizar este estudio, se han analizado dos escenarios reales con el fin de verificar los resultados y encontrar relaciones precisas. A partir de las conclusiones de estos análisis previos, el segundo objetivo ha sido diseñar un algoritmo cognitivo que, en base a las condiciones meteorológicas y a los parámetros de rendimiento de red, permita a las redes ajustar sus características de transmisión con el fin de superar tal impacto. Con el fin de llevar a cabo este estudio, primero fue necesario estudiar qué métodos estadísticos podían ser utilizados para extraer el nivel de correlación entre los parámetros de rendimiento de las redes y las condiciones meteorológicas del entorno. En segundo lugar, se tuvo que analizar qué parámetros de rendimiento de red se podían extraer de la red exterior de la UPV y seleccionarlos de acuerdo con el objetivo perseguido. A continuación, se definió el periodo de tiempo durante el cual se almacenarían los parámetros seleccionados de forma periódica. El siguiente paso fue seleccionar y almacenar las condiciones meteorológicas de una estación cercana durante el mismo periodo de tiempo. Finalmente, se realizó un preprocesado detallado con el fin de poner en orden todo el volumen de datos y se analizaron estadísticamente. Los resultados fueron exitosos, sin embargo aparecieron varios problemas por el hecho de estudiar una red real muy variable. Por lo tanto, se tuvo que desarrollar un escenario experimental con el fin de verificar los resultados. Para ello se diseñó y desarrolló un enlace exterior IEEE 802.11b/g punto a multipunto, y se analizó de nuevo el impacto de las condiciones meteorológicas. Se consideró un enlace multipunto para analizar también cómo influía el impacto del tiempo según la distancia y los diferentes esquemas de modulación. A partir de los resultados, se diseñó un algoritmo cognitivo energéticamente eficiente con el fin de reducir el impacto de los fenómenos meteorológicos en las redes IEEE 802.11b/g. Dicho algoritmo ha sido simulado y los resultados obtenidos han sido satisfactorios, tanto en términos de eficiencia energética como de rendimiento de la red. Para concluir, otro factor externo que se ha estudiado en esta tesis ha sido la tasa de absorción específica. Este parámetro está relacionado con una de las grandes preocupaciones actuales en cuanto a salud pública, ya que se utiliza para medir la exposición de los tejidos del cuerpo humano a los campos electromagnéticos. Obviamente, la absorción de señal por parte del cuerpo humano afecta a las redes y, por lo tanto, este parámetro se debería tener en cuenta a la hora de diseñar redes eficientes. Por esta razón se ha incluido en esta tesis doctoral.
[CAT] Actualment hi ha molts treballs que analitzen i intenten millorar el rendiment de les xarxes d'àrea local sense fils des de diferents perspectives. Gran part d'aquests treballs es focalitzen en aspectes de disseny, com són la distribució de freqüències o l'assignació de canals. Per tant, com aquests aspectes ja han sigut àmpliament estudiats, els esforços d'aquesta tesi doctoral s'han dirigit a estudiar altres factors que també podrien afectar al seu rendiment i que encara no han sigut analitzats amb profunditat. L'objectiu principal d'aquesta tesi doctoral ha sigut realitzar un estudi minuciós per analitzar l'impacte de les condicions meteorològiques sobre el rendiment de les xarxes IEEE 802.11b/g. Per a realitzar aquest estudi s'han analitzat dos escenaris reals per tal de verificar els resultats i trobar relacions precises. A partir de les conclusions d'aquests anàlisis previ, el següent objectiu ha sigut dissenyar un algoritme cognitiu que, en base a les condicions meteorològiques i als paràmetres de rendiment de la xarxa, permeti a les xarxes ajustar les seues característiques de transmissió per tal de superar tal impacte. Per tal de dur a terme aquest estudi, primer va ser necessari estudiar quins mètodes estadístics podien ser utilitzats per extraure el nivell de correlació entre els paràmetres de rendiment de les xarxes i les condicions meteorològiques de l'entorn. En segon lloc, es va haver d'analitzar quins paràmetres de rendiment es podien extraure de la xarxa exterior de la UPV i es van seleccionar d'acord a l'objectiu plantejat. A continuació, es va definir el període temporal al llarg del qual s'emmagatzemarien els paràmetres seleccionats de manera periòdica. El següent pas va ser seleccionar i emmagatzemar les condicions meteorològiques d'una estació propera durant el mateix període de temps. Finalment, es va realitzar un preprocessat per tal de posar en ordre tot el volum de dades i es van analitzar estadísticament. Els resultats van ser exitosos, però van aparèixer diversos problemes pel fet d'estudiar una xarxa real molt variable. Per tant, es va haver de desenvolupar un escenari experimental amb l'objectiu de verificar els resultats. Per aquesta raó es va dissenyar i implementar un enllaç exterior IEEE 802.11b/g punt a multipunt, i es va analitzar de nou l'impacte de les condicions meteorològiques. Es va considerar un enllaç multipunt per tal de d'analitzar també com influïa el impacte del temps segons la distància i els diferents esquemes de modulació. A partir d'aquests resultats, es va dissenyar un algoritme cognitiu energèticament eficient per tal de reduir l'impacte dels fenòmens meteorològics sobre les xarxes IEEE 802.11b/g. Aquest algoritme va ser simulat i els resultats obtinguts van ser satisfactoris, tant en termes d'eficiència energètica com de rendiment de la xarxa. va comprovar que la proposta aporta millores. Per concloure, un altre factor extern que s'ha estudiat en aquesta tesi doctoral ha sigut la taxa d'absorció específica. Aquest paràmetre està relacionat amb una de les preocupacions actuals pel que fa a la salut pública, ja que s'utilitza per a mesurar l'exposició dels teixits del cos humà als camps electromagnètics. Òbviament, aquesta absorció de la senyal afecta el rendiment de les xarxes i, per això, aquest paràmetre s'hauria de tenir en compte a l'hora d'implementar futures xarxes sense fils eficients. Per aquesta raó s'ha inclòs en aquesta tesi doctoral.
Bri Molinero, D. (2015). ESTUDIO DEL EFECTO DE FACTORES EXTERNOS SOBRE LAS REDES WLAN Y DISEÑO DE UN ALGORITMO COGNITIVO ENERGÉTICAMENTE EFICIENTE [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/53450
TESIS

In this thesis, Genetic Granular Cognitive Fuzzy Neural Networks (GGCFNN), combining genetic algorithms (GA) and granular cognitive fuzzy neural networks (GCFNN), is proposed for pattern recognition problems. According to cognitive patterns, biological neural networks in the human brain can recognize different patterns. Since GA and neural networks represent two learning methods based on biological science, it is indispensable and valuable to investigate how biological neural networks and artificial neural networks recognize different patterns. The new GGCFNN, based on granular computing, soft computing and cognitive science, is used in the pattern recognition problems. The hybrid forward-wave-backward-wave learning algorithm, as a main learning technology in GCFNN, is used to enhance learning quality. GA optimizes parameters to make GGCFNN get better learning results. Both pattern recognition results generated by human persons and those by GGCFNN are analyzed in terms of computer science and cognitive science.

Esta tese de doutorado trata de algoritmos de detecção cooperativa aplicados ao problema de sensoriamento espectral em sistemas de rádios cognitivos. O problema de detecção cooperativa é abordado sob dois paradigmas distintos: detecção centralizada e distribuída. No primeiro caso, considera-se que o sistema conta com um centro de fusão responsável pela tomada de decisão no processo de detecção. Já no segundo caso, considera-se que os rádios cognitivos da rede trocam informações entre si e as decisões são tomadas localmente. No que concerne ao sensoriamento espectral centralizado, são estudados os casos em que os rádios cognitivos enviam apenas um bit de decisão para o centro de fusão (decisão do tipo hard) e também o caso em que o detector envia a própria estatística de teste ao centro de fusão (decisão do tipo soft). No âmbito de sensoriamento espectral cooperativo com detecção distribuída, são tratados três cenários diferentes. No primeiro, considera-se o caso em que os rádios cognitivos têm conhecimento a priori do sinal enviado pelo usuário primário do sistema e do canal entre eles e o usuário primário. No segundo caso, há conhecimento apenas do sinal enviado pelo usuário primário. Já no terceiro, os rádios cognitivos não dispõem de qualquer informação a priori do sinal enviado pelo usuário primário. Além do problema de detecção distribuída, a tese também apresenta um capítulo dedicado ao problema de estimação, diretamente associado ao de detecção. Esse último problema é abordado utilizando algoritmos derivados da teoria clássica de filtragem adaptativa.
This doctorate thesis deals with cooperative detection algorithms applied to the spectral sensing problem. The cooperative detection problem is approached under two different paradigms: centralized and distributed detection. In the first case, is considered that a fusion center responsible for detection decision is presented in the system. On the other hand, in the second case, is considered that the cognitive radios in the network exchange information among them. Concerning the centralized spectrum sensing system, the case in which the cognitive radios send only one decision bit (hard decision) to the fusion center and the case in which the detector send the statistic test (soft decision) are considered. Regarding the spectrum sensing system with distributed detection, the work analysis three different scenarios. In the first one, where the cognitive radios explore an a priori knowledge of the primary user signal and the channel between the primary user and the cognitive radio. In the second one, the cognitive radios use an a priori knowledge of only the primary user signal. And, in the las scenario, there is no a priori knowledge about the primary user signal. Besides the distributed detection problem, the thesis also presents a chapter dedicated to the estimation problem, which is directed related to the detection problem. This last issue is approached using adaptive algorithms derived from the classic adaptive filtering theory.

Due to the increasing use of more and more powerful smart devices demands on the scarce radio spectrum are becoming more intense. One way to cope with increasing demands on radio spectrum is to apply innovative and flexible authorization schemes like spectrum sharing. Under the spectrum sharing paradigm, multiple users and/or systems are authorized to utilize the same spectrum band in a defined sharing agreement. A technology that is generally recommended for the implementation of spectrum sharing is called cognitive radio (CR). In this thesis, we design and implement a cognitive engine (CE) that intelligently and dynamically allocates spectrum resources to users. We first consider a scenario where a network has an exclusive access to a spectrum band and the CE accepts or rejects the arrival user requests based on an algorithm that takes into account a user’s priority and its bandwidth demand. We then consider a spectrum sharing scenario where along with the exclusive utilization to its own spectrum band a network also can opportunistically utilize another network’s spectrum band. For this scenario, we design and implement a CE that performs two main tasks: 1) Accepts or rejects arrival user requests based on a priority based algorithm; and 2) it intelligently load balances the user traffic between the two available network resources, while taking into account the primary user activity in the shared spectrum band. We provide a load balancing algorithm and evaluate its performance under different primary and secondary user traffic scenarios. We show that the proposed load balancing algorithm increases average throughput of the network and it also reduces the average number of users rejected by the network
Yhä tehokkaampien älykkäiden langattomien päätelaitteiden nopea lisääntyminen johtaa niukan radiospektrin yhä kiihtyvään käyttöön. Eräs menetelmä radiospektrin lisääntyvän kysynnän tyydyttämiseen on hyödyntää innovatiivista ja joustavaa resurssin käytönjakoa kuten spektrin jakamista. Spektrinjakamismalli mahdollistaa useiden käyttäjien ja/tai järjestelmien yhtäaikaisen käytön samalla taajuuskaistalla hyödyntämällä sovittua käytäntöä resurssien jakamisesta. Radiospektrin jakaminen on tänä päivänä yleisesti suositeltu toteuttamaan hyödyntämällä kognitiivista radioteknologiaa. Tässä työssä suunnittellaan ja toteutetaan kognitiivinen päätöksentekokone, joka jakaa radiospektriresursseja käyttäjille älykkäästi ja dynaamisesti. Kognitiivista päätöksentekokonetta radioresurssien jakamisessa hyödynnetään kahdessa skenaariossa. Ensimmäisessä skenaariossa radioverkolla on yksinomainen pääsy taajuuskaistalle, jonka käyttöä kognitiivinen päätöksentekokone säätelee joko hyväksymällä tai hylkäämällä verkkoon liittyviä käyttäjiä. Kognitiivinen päätöksentekokoneen päätökset perustuu algoritmiin, joka ottaa huomioon käyttäjien määritetyn tärkeyden ja käyttäjän vaatiman kaistanleveyden. Seuraavassa skenaariossa radioverkko voi oman yksinomaisen taajuuskaistan lisäksi hyödyntää opportunisesti toisen radioverkon taajuuskaistaa silloin, kun siellä ei ole liikennettä. Tätä skenaariota varten suunnitteltiin kognitiivinen päätöksentekokone, jolla on kaksi päätehtävää: 1) hyväksyä tai hylätä verkkoon liittyviä käyttäjiä edellämainitun tärkeysperusteisen algoritmin avulla; ja 2) jakaa käyttäjien liikennettä kahden tarjolla olevan verkon välillä samalla ottaen huomioon opportunistisen resurssin pääkäyttäjien liikenteen jaetulla taajuuskaistalla. Tässä työssä esitellään toteutettu kuormantasausalgoritmi, jonka suorituskykyä tarkastellaan erilaisissa pääkäyttäjien ja toissijaisien käyttäjien liikenneskenaarioissa. Simulaatiotulokset osoittavat, että esitellyn kuormanjakoalgoritmin hyödyntäminen kognitiivisessa päätöksentekokoneessa parantaa verkon keskimääräistä siirtonopeutta, sekä vähentää keskimääräistä käyttäjien hylkäysastetta verkossa. Algoritmimme parantaa opportunistisen taajuuskaistan käyttöastetta. Algoritmimme ottaa myös huomioon käyttäjille asetetut prioriteetit ja parantaa korkeampi prioriteettisten käyttäjien asemaa verkossa. Tämä tulee ilmi muun muassa korkeampi prioriteettisten käyttäjien pienemmässä hylkäysasteessa

Future wireless communications systems are expected to be extremely dynamic, smart and capable to interact with the surrounding radio environment. To implement such advanced devices, cognitive radio (CR) is a promising paradigm, focusing on strategies for acquiring information and learning. The first task of a cognitive systems is spectrum sensing, that has been mainly studied in the context of opportunistic spectrum access, in which cognitive nodes must implement signal detection techniques to identify unused bands for transmission. In the present work, we study different spectrum sensing algorithms, focusing on their statistical description and evaluation of the detection performance. Moving from traditional sensing approaches we consider the presence of practical impairments, and analyze algorithm design. Far from the ambition of cover the broad spectrum of spectrum sensing, we aim at providing contributions to the main classes of sensing techniques. In particular, in the context of energy detection we studied the practical design of the test, considering the case in which the noise power is estimated at the receiver. This analysis allows to deepen the phenomenon of the SNR wall, providing the conditions for its existence and showing that presence of the SNR wall is determined by the accuracy of the noise power estimation process. In the context of the eigenvalue based detectors, that can be adopted by multiple sensors systems, we studied the practical situation in presence of unbalances in the noise power at the receivers. Then, we shift the focus from single band detectors to wideband sensing, proposing a new approach based on information theoretic criteria. This technique is blind and, requiring no threshold setting, can be adopted even if the statistical distribution of the observed data in not known exactly. In the last part of the thesis we analyze some simple cooperative localization techniques based on weighted centroid strategies.

Future wireless communications systems are expected to be extremely dynamic, smart and capable to interact with the surrounding radio environment. To implement such advanced devices, cognitive radio (CR) is a promising paradigm, focusing on strategies for acquiring information and learning. The first task of a cognitive systems is spectrum sensing, that has been mainly studied in the context of opportunistic spectrum access, in which cognitive nodes must implement signal detection techniques to identify unused bands for transmission. In the present work, we study different spectrum sensing algorithms, focusing on their statistical description and evaluation of the detection performance. Moving from traditional sensing approaches we consider the presence of practical impairments, and analyze algorithm design. Far from the ambition of cover the broad spectrum of spectrum sensing, we aim at providing contributions to the main classes of sensing techniques. In particular, in the context of energy detection we studied the practical design of the test, considering the case in which the noise power is estimated at the receiver. This analysis allows to deepen the phenomenon of the SNR wall, providing the conditions for its existence and showing that presence of the SNR wall is determined by the accuracy of the noise power estimation process. In the context of the eigenvalue based detectors, that can be adopted by multiple sensors systems, we studied the practical situation in presence of unbalances in the noise power at the receivers. Then, we shift the focus from single band detectors to wideband sensing, proposing a new approach based on information theoretic criteria. This technique is blind and, requiring no threshold setting, can be adopted even if the statistical distribution of the observed data in not known exactly. In the last part of the thesis we analyze some simple cooperative localization techniques based on weighted centroid strategies.

The increased demand on wireless applications, coupled with the current inefficiency in spectrum usage, mandate a new communication paradigm shift from fixed spectrum assignment to dynamic spectrum sharing which can be achieved using the cognitive radio technology. Cognitive radio allows unlicensed secondary nodes to form communication links over licensed spectrum bands on an opportunistic basis which increases the spectrum management efficiency. Cognitive radio networks (CRN), however, impose unique challenges due to the fluctuation in the available spectrum as well as the diverse quality of service requirements. One of the main challenges is the establishment and maintenance of routes in multi-hop CRNs. In this thesis, we critically investigate the problem of routing in multi-hop CRNs. The main objective of this research is to maximize network connectivity while limiting routing delay. We developed a general connectivity metric for single-band and multi-band CRNs based on the properties of the Laplacian matrix eigenvalues spectrum. We show through analytical and simulation results that the developed metric is more robust and has lower computational complexity than the previously proposed metrics. Furthermore, we propose a new position-based routing algorithm for large scale CRNs which significantly reduces the routing computational complexity with negligible performance degradation compared to the traditional full node search algorithm. In addition, the connectivity metric developed in this thesis is used to develop a connectivity-aware distributed routing protocol for CRNs. Finally, we use a commodity cognitive radio testbed to demonstrate the concept of CR Wi-Fi networks.

The work presented in this thesis concerns one of the key enabling techniques related to cognitive radio functionalities which is spectrum sensing as well as cooperative spectrum sensing. As cooperative spectrum sensing (CSS) approaches are commonly used for combating fading and improving detection performance, their performances using different combining rules have been analyzed. Due to the low implementation complexity, Goodness of Fit based spectrum sensing has been studied for cognitive radio applications. Motivated by its nice features of local sensing, a distributed consensus spectrum sensing for CR, has been presented, integrating a Goodness of Fit based spectrum sensing scheme.
Le travail présenté dans cette thèse concerne l'une des techniques clés dans les fonctionnalités de la radio cognitive qui est la détection du spectre ainsi que la détection coopérative du spectre. La détection coopérative est couramment utilisée pour la lutte contre l’évanouissement du canal à fin d'améliorer les performances de la détection. Les performances de la détection coopérative en utilisant différentes règles de fusion ont été analysées. En raison sa simplicité, la détection du spectre par les testes d’adéquation a été étudiée pour les applications de la radio cognitive. Motivé par la caractéristique d’être indépendant de bruit, ces testes d’adéquation ont été utilisés pour la détection locale, pour la détection coopérative distribuée.
Doctorat en Sciences de l'ingénieur et technologie
info:eu-repo/semantics/nonPublished

In this thesis new methods are presented for achieving spectrum sensing in cognitive radio wireless networks. In particular, supervised, semi-supervised and unsupervised machine learning based spectrum sensing algorithms are developed and various techniques to improve their performance are described. Spectrum sensing problem in multi-antenna cognitive radio networks is considered and a novel eigenvalue based feature is proposed which has the capability to enhance the performance of support vector machines algorithms for signal classification. Furthermore, spectrum sensing under multiple primary users condition is studied and a new re-formulation of the sensing task as a multiple class signal detection problem where each class embeds one or more states is presented. Moreover, the error correcting output codes based multi-class support vector machines algorithms is proposed and investigated for solving the multiple class signal detection problem using two different coding strategies. In addition, the performance of parametric classifiers for spectrum sensing under slow fading channel is studied. To address the attendant performance degradation problem, a Kalman filter based channel estimation technique is proposed for tracking the temporally correlated slow fading channel and updating the decision boundary of the classifiers in real time. Simulation studies are included to assess the performance of the proposed schemes. Finally, techniques for improving the quality of the learning features and improving the detection accuracy of sensing algorithms are studied and a novel beamforming based pre-processing technique is presented for feature realization in multi-antenna cognitive radio systems. Furthermore, using the beamformer derived features, new algorithms are developed for multiple hypothesis testing facilitating joint spatio-temporal spectrum sensing. The key performance metrics of the classifiers are evaluated to demonstrate the superiority of the proposed methods in comparison with previously proposed alternatives.

L’aéronautique civile commerciale poursuit l’objectif du déplacement de biens ou de personnes, par les airs, en maintenant un niveau optimal de sécurité. Depuis plus de trente ans, en dépit de cadres normatifs de plus en plus stricts et d’automatismes de plus en plus performants, le rapport entre performance visée et risque encouru ne progresse plus.Le facteur humain constitue un levier d’action majeur pour franchir ce plancher de verre. Dans le cadre contraint de l’aéronautique, la conception d’outils visant à assister la cognition des pilotes est ainsi une direction d’avenir. L’anticipation a été identifiée comme un processus central dans la gestion des ressources cognitives. Dans une démarche de cognitique, nous avons ainsi entrepris la conception d’un outil d’aide à l’anticipation en impliquant des pilotes à chaque étape des développements.D’une analyse de l’activité sur la base d’enregistrements en cockpit et d’entretiens, nous avons construit un modèle de l’activité réelle des pilotes lors de la descente et l’approche sur l’aéroport de Rio de Janeiro. L’étude bibliographique mit en lumière des points critiques relevant de l’anticipation et nécessitant une expérimentation préliminaire. Les résultats expérimentaux conciliés à nos hypothèses de compréhension de l’anticipation achevèrent le cahier des charges du cœur fonctionnel de notre outil d’aide à l’anticipation. Un algorithme de planification dynamique exploitant notre modèle de l’activité fut conçu et implémenté au sein d’ASAP (Anticipation Support for Aeronautical Planning) le démonstrateur de concept industriel de Thales Avionics. 36 pilotes civils commerciaux participèrent enfin à son évaluation en simulateur
Civil aviation pursues the objective of moving people or goods through the air with an optimal level of safety. For more than thirty years, despite a stricter and stricter regulatory framework and highly reliable automation, the ratio between performance and acceptable risk is not improving anymore.Human factors are a major action lever to break this glass floor. In the constrained context of aviation, designing tools aiming at assisting pilots’ cognition is thus a promising direction. Anticipation has been identified central in the process of cognitive resources management. In a human factors engineering approach, we undertook the design of an anticipation support tool involving pilots at each step of the development.From an activity analysis performed on the basis of in-cockpit recordings and interviews we constructed a model of the actual pilots’ activity during the descent and approach phases on Rio de Janeiro airport. The state of the art highlighted the key elements related to anticipation which could take benefit of a preliminary experiment. Experimental results brought together with our hypotheses about how anticipation works completed the requirements of the functional core of our anticipation support tool. A dynamic planning algorithm was then designed and implemented within ASAP (Anticipation Support for Aeronautical Planning), Thales Avionics’ proof of concept. 36 commercial pilots took part to its evaluation in a simulated environment

Cognitive Radio (CR) is considered to be a novel approach to improve the underutilization of precious radio resources by exploiting the unused licensed spectrum in dynamically changing environments. Designing efficient resource allocation algorithms for dynamic spectrum sharing and for power allocation in OFDM-CR networks is still a challenging problem. In this thesis, we specifically deal with these two problems. Dynamic spectrum sharing for the unlicensed secondary users (SU)s with device coordination could minimize the wastage of the spectrum. But this is a feasible approach only if the network considers the fairness criterion. We study the dynamic spectrum sharing problem for device coordinated cognitive radio networks with respect to fairness. We propose a simple modified proportional fair algorithm for a dynamic spectrum sharing scenario with two constraints, time and utility. Utility is measured by the amount of data processed and time is measured as the duration of a slot. This algorithm could result in variable or fixed length time slots. We will discuss the several controls possible on the algorithm and the possible extension of this algorithm for multicarrier OFDM based CR systems. Traditional water-filling algorithm is inefficient for OFDM-CR networks due to the interaction with primary users (PU)s. We consider reliability/availability of subcarriers or primary user activity for power allocation. We model this aspect mathematically with a risk-return model by defining a general rate loss function. We then propose optimal and suboptimal algorithms to allocate power under a fixed power budget for such a system with linear rate loss. These algorithms as we will see allocate more power to more reliable subcarriers in a water-filling fashion with different water levels. We compare the performance of these algorithms for our model with respect to water-filling solutions. Simulations show that suboptimal schemes perform closer to optimal scheme although they could be implemented with same complexity as water-filling algorithm. We discuss the linearity of loss function and guidelines to choose its coefficients by obtaining upper bounds on them. Finally we extend this model for interference-limited OFDM-CR systems.

Thesis: Ph. D. in Cognitive Science, Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences, September, 2020
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 213-224).
The future of machine learning should have a knowledge representation that supports, at a minimum, several features: Expressivity, interpretability, the potential for reuse by both humans and machines, while also enabling sample-efficient generalization. Here we argue that programs-i.e., source code-are a knowledge representation which can contribute to the project of capturing these elements of intelligence. This research direction however requires new program synthesis algorithms which can induce programs solving a range of AI tasks. This program induction challenge confronts two primary obstacles: the space of all programs is infinite, so we need a strong inductive bias or prior to steer us toward the correct programs; and even if we have that prior, effectively searching through the vast combinatorial space of all programs is generally intractable. We introduce algorithms that learn to induce programs, with the goal of addressing these two primary obstacles. Focusing on case studies in vision, computational linguistics, and learning-to-learn, we develop an algorithmic toolkit for learning inductive biases over programs as well as learning to search for programs, drawing on probabilistic, neural, and symbolic methods. Together this toolkit suggests ways in which program induction can contribute to AI, and how we can use learning to improve program synthesis technologies.
by Kevin Ellis.
Ph. D. in Cognitive Science
Ph.D.inCognitiveScience Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Brain and Cognitive Sciences, 2015.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 167-176).
This thesis develops probabilistic programming as a productive metaphor for understanding cognition, both with respect to mental representations and the manipulation of such representations. In the first half of the thesis, I demonstrate the representational power of probabilistic programs in the domains of concept learning and social reasoning. I provide examples of richly structured concepts, defined in terms of systems of relations, subparts, and recursive embeddings, that are naturally expressed as programs and show initial experimental evidence that they match human generalization patterns. I then proceed to models of reasoning about reasoning, a domain where the expressive power of probabilistic programs is necessary to formalize our intuitive domain understanding due to the fact that, unlike previous formalisms, probabilistic programs allow conditioning to be represented in a model, not just applied to a model. I illustrate this insight with programs that model nested reasoning in game theory, artificial intelligence, and linguistics. In the second half, I develop three inference algorithms with the dual intent of showing how to efficiently compute the marginal distributions defined by probabilistic programs, and providing building blocks for process-level accounts of human cognition. First, I describe a Dynamic Programming algorithm for computing the marginal distribution of discrete probabilistic programs by compiling to systems of equations and show that it can make inference in models of "reasoning about reasoning" tractable by merging and reusing subcomputations. Second, I introduce the setting of amortized inference and show how learning inverse models lets us leverage samples generated by other inference algorithms to compile probabilistic models into fast recognition functions. Third, I develop a generic approach to coarse-to-fine inference in probabilistic programs and provide evidence that it can speed up inference in models with large state spaces that have appropriate hierarchical structure. Finally, I substantiate the claim that probabilistic programming is a productive metaphor by outlining new research questions that have been opened up by this line of investigation.
by Andreas Stuhlmüller.
Ph. D.

As the demand for wireless communication services grows quickly, spectrum scarcity has been on the rise sharply. In this context, cognitive radio (CR) is being viewed as a new intelligent technology to solve the deficiency of fixed spectrum assignment policy in wireless communications. Spectrum sensing is one of the most fundamental technologies to realise dynamic spectrum access in cognitive radio networks. It requires high accuracy as well as low complexity. In this thesis, a novel adaptive threshold setting algorithm is proposed to optimise the trade-off between detection and false alarm probability in spectrum sensing while satisfying sensing targets set by the IEEE 802.22 standard. The adaptive threshold setting algorithm is further applied to minimise the error decision probability with varying primary users' spectrum utilisations. A closed-form expression for the error decision probability, satisfied SNR value, number of samples and primary users' spectrum utilisation ratio are derived in both fixed and the proposed adaptive threshold setting algorithms. By implementing both Welch and wavelet based energy detectors, the adaptive threshold setting algorithm demonstrates a more reliable and robust sensing result for both primary users (PUs) and secondary users (SUs) in comparison with the conventional fixed one. Furthermore, the wavelet de-noising method is applied to improve the sensing performance when there is insu cient number of samples. Finally, a novel database assisted spectrum sensing algorithm is proposed for a secondary access of the TV White Space (TVWS) spectrum. The proposed database assisted sensing algorithm is based on the developed database assisted approach for detecting incumbents like Digital Terrestrial Television (DTT) and Programme Making and Special Events (PMSE), but assisted by spectrum sensing to further improve the protection to primary users. Monte-Carlo simulations show a higher SUs' spectrum efficiency can be obtained for the proposed database assisted sensing algorithm than the existing stand-alone database assisted or sensing models.

Wireless communications have experienced tremendous growth in the last decade, which has placed significant demand for RF spectrum, leading to spectrum "crunch." Driven by numerous studies that revealed the significant under-utilization of many licensed channels in the VHF and UHF bands, a new paradigm for spectrum sharing has emerged in the past decade. In this paradigm, wireless devices (a.k.a. secondary users) are allowed to operate opportunistically in certain licensed bands without interfering with the licensed users (a.k.a. primary users). The realization of this new communication paradigm necessitates the design of a new generation of smart, adaptable, and programmable radios, called cognitive radios. Enabling opportunistic operation requires addressing various challenges including device coordination, resource allocation, channel monitoring, and various security issues. Specifically, secondary users are particularly vulnerable to node compromise and malicious jamming attacks. In this dissertation, we first develop several rendezvous algorithms for establishing unicast as well as multicast communication links in opportunistic spectrum access networks. The developed rendezvous algorithms are shown to be robust to node compromise attacks. Second, we study the anti-jamming rendezvous problem in the presence of an insider attack. We develop a combinatorial game-theoretic framework to capture the interactions between the rendezvousing nodes and the insider jammer. Third, to account for the dynamism of primary users, we propose novel stochastic resource allocation schemes under channel-quality uncertainty. The proposed schemes support channel bonding and aggregation and account for adjacent channel interference by introducing guard-bands. Fourth, to prevent interference with primary users, we design an optimal spectrum-sensing algorithm that achieves high detection accuracy and low false-alarm rate. Finally, we present an application of using cognitive radios for jamming mitigation in satellite communications. Extensive simulations are conducted to demonstrate the effectiveness and robustness of the proposed algorithms.

Dans le contexte de la Radio Cognitive, la reconnaissance des systèmes radio relève d’une grande importance (e. G. , Wifi, Wimax, 3GPP/LTE, DVB-T). Nous proposons d’étudier ce problème en se focalisant sur les systèmes basés sur la modulation OFDM largement utilisés actuellement. Nous avons noté que les standards existants diffèrent dans la valeur de l’espacement inter-porteuses utilisée. Par conséquent, nous avons développé des algorithmes qui se basent sur l’estimation de ce paramètre pour réaliser la reconnaissance aveugle du système intercepté. Les approches issues de la littérature et basées essentiellement sur l’utilisation du préfixe cyclique se retrouvent inefficaces dans le cas d’un très court préfixe ou d’un canal de propagation fortement sélectif en fréquence. Ce travail de thèse propose alors des solutions alternatives pour estimer les paramètres d’un signal OFDM en faisant appel à différentes approches : i) le kurtosis normalisé, ii) le principe du maximum de vraisemblance, iii) le filtrage adapté et iv) la cyclostationnarité d’ordre deux. Nous avons démontré la grande efficacité de ces algorithmes avec des conditions de fonctionnement très sévères ( un préfixe cyclique court, un canal de propagation très sélectif en fréquence, non sychronisation temporelle et/ou fréquentielle)
In the context of cognitive radio it is a crucial task to distinguish blindly various wireless systems (e. G. , Wifi, Wimax, 3GPP/LTE, DVB-T) from each others. We focus on the OFDM based systems which differ from their subcarrier spacing used in OFDM modulation. One can thus carry out recognition algorithms based on the value of the subcarrier spacing. Standard approaches developed in the literature rely on the detection of the cyclic prefix which enables to exhibit the value of the used subcarrier spacing. Nevertheless, these approaches fail when either the cyclic prefix duration is small or the channel impulse response is almost as large as the cyclic prefix. Therefore, this thesis proposes new algorithms to estimate the parameters of OFDM modulated signal (especially the subcarrier spacing) relying on i) the normalized kurtosis, ii) the maximum-likelihood principle, iii) the matched filter, and iv) the second-order cyclostationary property. We have shown the strong robustness of proposed algorithms to short cyclic prefix, multipath channel, time offset, and frequency offset

For the first time in decades computers are evolving into a fundamentally new class of machine. Transistors are still getting smaller, more economical, and more power-efficient, but operating frequencies leveled off in the mid-2000's. Today, improving performance requires placing a larger number of slower processing cores on each of many chips. Software written for such machines must scale out over many cores rather than scaling up with a faster single core. Biological computation is an extreme manifestation of such a many-slow-core architecture and therefore offers a potential source of ideas for leveraging new hardware. This dissertation addresses several problems in the intersection of emerging computer architectures and biological computation, termed Cognitive Computing: What mechanisms are necessary to maintain stable representations in a large distributed learning system? How should complex biologically-inspired algorithms be tested? How do visual sensing limitations like occlusion influence performance of classification algorithms? Neurons have a limited dynamic output range, but must process real-world signals over a wide dynamic range without saturating or succumbing to endogenous noise. Many existing neural network models leverage spatial competition to address this issue, but require hand-tuning of several parameters for a specific, fixed distribution of inputs. Integrating spatial competition with a stabilizing learning process produces a neural network model capable of autonomously adapting to a non-stationary distribution of inputs. Human-engineered complex systems typically include a number of architectural features to curtail complexity and simplify testing. Biological systems do not obey these constraints. Biologically-inspired algorithms are thus dramatically more difficult to engineer. Augmenting standard tools from the software engineering community with features targeted towards biologically-inspired systems is an effective mitigation. Natural visual environments contain objects that are occluded by other objects. Such occlusions are under-represented in the standard benchmark datasets for testing classification algorithms. This bias masks the negative effect of occlusion on performance. Correcting the bias with a new dataset demonstrates that occlusion is a dominant variable in classification performance. Modifying a state-of-the-art algorithm with mechanisms for occlusion resistance doubles classification performance in high-occlusion cases without penalty for unoccluded objects.

认知无线电在处理频谱稀缺的问题上是一个非常有前途的解决方案。拥有多天线认知无线电的用戶通过发射波束成形技术可以和授权用在同一时刻同一频带共存，这样大大地增强了频谱效率。在实际系统中，最理想的情况是这些拥有多天线认知无线电用戶能够分布式地优化他们的发射波束形成向量以此达到系统的最优化。由于授权用戶受到的干扰是来自于所有认知无线电用戶的，为了实现分布式算法这些干扰必须被合理地规划以至于达到最优。也就是说，每个认知无线电用戶需要知道对授权用戶产生干扰的最佳约束上限。
从优化的角度处理这种解耦问题，最常用的方法是原始分解法和对偶分解法。然而这两种方法都需要用戶之间有大量的消息传递，这对于频谱效率来说是有害的。在对偶分解法中，指向授权用戶的耦合干扰被一协调者估测(通常是授权用戶本身)。协调者需要在每次迭代中更新和广播参数给认知无线电用戶。对于原始分解法，算法同样需要一协调者进行收集认知无线电用戶的目标函数信息以此计算每个用戶的最优干扰约束上限。协调者同样需要更新和广播大量消息给认知无线电用戶。这种大量的信息计算和传递在分布式系统中是不理想的，问题在认知无线电网络显得格外严重。因为授权用戶不希望担任这样的协调者除非他的计算参与降到最低。
在此论文中，我们提出了几种新型的基于认知无线电网络的分布式算法。目的是最小化授权用戶和认知无线电用戶的消息传递。通过研究半定规划中的最优分割法，我们指出不影响最优性条件下授权用戶和认知无线电用戶的大量消息传递是可以避免的。我们又提出了在多输入多数出认知无线电网络中一种基于对偶分解的鲁捧干扰控制。在此论文中提出的低消息传递算法大大地提高了多用戶多输入多数认知无线电网络的实用性。
Cognitive radio (CR) is a promising solution to alleviate spectrum scarcity. In CR networks where mobile stations are equipped with multiple antennas, secondary users (SUs) can transmit at the same time as the primary users (PUs) by carefully controlling the interference through transmit beamforming, thus significantly enhancing the spectrum efficiency. In practical systems, it is desirable to have multiple SUs optimize their transmit beamforming vectors in a decentralized manner, and yet achieve an optimal system performance. In CR networks, the interference received by the PU is attributed to the transmission of all SUs. To facilitate distributed beamforming, the aggregate-interference constraint imposed by the PU must be decoupled, so that each individual SU knows the "fair share" of interference that is allowed to generate to the PU.
A commonly used technique for decoupling coupled constraintsin optimization problems is optimization decomposition, including dual and primal decompositions. Both the dual and primal decomposition methods require frequent message passing among users, which potentially offsets the spectrum benefit brought by cognitive radio techniques. Specifically, with dual decomposition, the aggregate interference generated to the PU must be measured by a coordinator,which is, naturally, the PU. The coordinator then updates and broadcasts the Lagrangian multiplier to all SUs. Likewise, the primal decomposition needs a coordinator, which can again be the PU, to gather the subgradient of the objective functions of each SUs for given interference partition. The coordinator then updates and broadcasts the permissible interference to all SUs. Whereas the large overhead incurred message computation and passing is undesirable in distributed systems, the problem is more acute in CR networks, because a typical PU would not be willing to take the coordinating role unless its involvement is minimized.
In this thesis, we propose several novel distributed optimization algorithms for CR networks with minimum message passing between the primary and secondary systems. By exploiting the theory of optimal partition (OP) for semi-definite programming (SDP), we show that most message passings between the primary and secondary systems can be eliminated without compromising the optimality of the solution. We also derive a robust interference control scheme based on the duality theory for MIMO CR network. The low message-passing distributed algorithms presented in this thesis greatly enhance the practicality of multiuser MIMO CR networks.
Detailed summary in vernacular field only.
Detailed summary in vernacular field only.
Detailed summary in vernacular field only.
Yao, Leiyi.
Thesis (Ph.D.) Chinese University of Hong Kong, 2014.
Includes bibliographical references (leaves 114-123).
Abstracts also in Chinese.

碩士
國立成功大學
電腦與通信工程研究所
98
We consider the secondary users have to carry out a distributed network optimization through the exchange of information between each other without infrastructure support in a cognitive radio network. Therefore, we propose the cross-layer system model including physical layer, data link layer and network layer. The physical layer involves the control of transmission power and data link layer includes the channel assignment and time scheduling while the network layer involves the selection of the routing path. For this system model, we design an inner loop and an outer loop algorithm to solve the nonconvex problem and the mixed-integer programming problem. In the inner loop, we use the gradient method to achieve the functionality of the distributed system. In the outer loop, we implement two kinds of method to allocate the time slots and channel to solve the network maximum throughput problem heuristically. According our modeling, the proposed algorithm is an effective resource allocator implemented distributed wireless cognitive radio network.

碩士
國立清華大學
通訊工程研究所
102
In cognitive radio networks (CRNs), primary users (PUs) have the absolute priority to access the license channels. To efficiently utilize the spectrum, secondary users (SUs) can dynamically access the unused channels by channel hopping (CH) schemes. Most existing CH schemes focus on the symmetric model that assumes all SUs have the same available channel set. However, the asymmetric model, where SUs may have different available channels sets, is more critical in the real CRN environment. In this thesis, we propose a universal channel hopping algorithm called Triple-Double Matrix (TDM) that can guarantee rendezvous within shorter period than the previous works under the asymmetric model without any constraints on the available channels of each SU. According to our simulation results, TDM has better maximum conditional time-to-rendezvous (MCTTR) than previous works.