http://sig.umd.edu/publications/Wang_ComNet_201010.pdf

Game theory for cognitive radio networks: An overview

In this survey paper, we characterize the learning problem in cognitive radios (CRs) and state the importance of artificial intelligence in achieving real cognitive communications systems. We review various learning problems that have been studied in the context of CRs classifying them under two main categories: Decision-making and feature classification. Decision-making is responsible for determining policies and decision rules for CRs while feature classification permits identifying and classifying different observation models. The learning algorithms encountered are categorized as either supervised or unsupervised algorithms. We describe in detail several challenging learning issues that arise in cognitive radio networks (CRNs), in particular in non-Markovian environments and decentralized networks, and present possible solution methods to address them. We discuss similarities and differences among the presented algorithms and identify the conditions under which each of the techniques may be applied.

A Survey on Machine-Learning Techniques in Cognitive Radios

The problem of distributed learning and channel access is considered in a cognitive network with multiple secondary users. The availability statistics of the channels are initially unknown to the secondary users and are estimated using sensing decisions. There is no explicit information exchange or prior agreement among the secondary users and sensing and access decisions are undertaken by them in a completely distributed manner. We propose policies for distributed learning and access which achieve order-optimal cognitive system throughput (number of successful secondary transmissions) under self play, i.e., when implemented at all the secondary users. Equivalently, our policies minimize the sum regret in distributed learning and access, which is the loss in secondary throughput due to learning and distributed access. For the scenario when the number of secondary users is known to the policy, we prove that the total regret is logarithmic in the number of transmission slots. This policy achieves order-optimal regret based on a logarithmic lower bound for regret under any uniformly-good learning and access policy. We then consider the case when the number of secondary users is fixed but unknown, and is estimated at each user through feedback. We propose a policy whose sum regret grows only slightly faster than logarithmic in the number of transmission slots.

/pdf/distributed-algorithms-for-learning-and-cognitive-medium-310prq4317.pdf

Distributed Algorithms for Learning and Cognitive Medium Access with Logarithmic Regret

We investigate the problem of achieving global optimization for distributed channel selections in cognitive radio networks (CRNs), using game theoretic solutions. To cope with the lack of centralized control and local influences, we propose two special cases of local interaction game to study this problem. The first is local altruistic game, in which each user considers the payoffs of itself as well as its neighbors rather than considering itself only. The second is local congestion game, in which each user minimizes the number of competing neighbors. It is shown that with the proposed games, global optimization is achieved with local information. Specifically, the local altruistic game maximizes the network throughput and the local congestion game minimizes the network collision level. Also, the concurrent spatial adaptive play (C-SAP), which is an extension of the existing spatial adaptive play (SAP), is proposed to achieve the global optimum both autonomously as well as rapidly.

Opportunistic Spectrum Access in Cognitive Radio Networks: Global Optimization Using Local Interaction Games

The problem of distributed learning and channel access is considered in a cognitive network with multiple secondary users. The availability statistics of the channels are initially unknown to the secondary users and are estimated using sensing decisions. There is no explicit information exchange or prior agreement among the secondary users. We propose policies for distributed learning and access which achieve order-optimal cognitive system throughput (number of successful secondary transmissions) under self play, i.e., when implemented at all the secondary users. Equivalently, our policies minimize the regret in distributed learning and access. We first consider the scenario when the number of secondary users is known to the policy, and prove that the total regret is logarithmic in the number of transmission slots. Our distributed learning and access policy achieves order-optimal regret by comparing to an asymptotic lower bound for regret under any uniformly-good learning and access policy. We then consider the case when the number of secondary users is fixed but unknown, and is estimated through feedback. We propose a policy in this scenario whose asymptotic sum regret which grows slightly faster than logarithmic in the number of transmission slots.

/pdf/distributed-algorithms-for-learning-and-cognitive-medium-2x73r41ero.pdf

We consider dynamic spectrum access among cognitive radios from an adaptive, game theoretic learning perspective. Spectrum-agile cognitive radios compete for channels temporarily vacated by licensed primary users in order to satisfy their own demands while minimizing interference. For both slowly varying primary user activity and slowly varying statistics of "fast" primary user activity, we apply an adaptive regret based learning procedure which tracks the set of correlated equilibria of the game, treated as a distributed stochastic approximation. This procedure is shown to perform very well compared with other similar adaptive algorithms. We also estimate channel contention for a simple CSMA channel sharing scheme.

Decentralized dynamic spectrum access for cognitive radios: cooperative design of a non-cooperative game

We present decentralized adaptive filtering algorithms for sensor activation control in an unattended ground sensor network (UGSN) comprised of ZigBee-enabled nodes. Nodes monitor their environment in a low-power ldquosleeprdquo mode, until an intruder is detected, then must decide whether to enter a full-power monitoring and transmission mode if their estimated average performance for activation outweighs their energy cost. The tradeoff is formulated in terms of the energy required to transmit data using the ZigBee protocol, probability of successful transmission, and the expected marginal increase in global utility resulting from a report, all of which depend on the activity of nearby sensor nodes. Since activation control is decentralized, and utilities are codependent, the adaptive filtering/stochastic approximation algorithms that we propose for sensor activation are based on game theoretic principles. We show that if each sensor operates according to this algorithm, the entire network is capable of actively tracking the correlated equilibrium set of the underlying game, which varies with target motion, node failures, or intentional parameter adjustments. We analyze the convergence and tracking properties of the adaptive filtering algorithms using differential inclusions.

Decentralized Adaptive Filtering Algorithms for Sensor Activation in an Unattended Ground Sensor Network

We describe a decentralized learning-based activation algorithm for a ZigBee-enabled unattended ground sensor network. Sensor nodes learn to monitor their environment in a low-power "sleep" mode, until an intruder is detected, then enter a full-power mode only if the benefit for doing so outweighs an energy cost. Our formulation accounts for the energy required to transmit and the probability of successful transmission in a crowded ZigBee network. Since these depend on the activity of other nodes, we propose a decentralized adaptive algorithm for sensor activation based on game theoretic principles. We show that the algorithm tracks the time-varying set of correlated equilibria of the problem, and illustrate performance through simulation. The algorithm is described as a stochastic approximation, with attendant differential inclusion analysis.

Decentralized Activation in a ZigBee-enabled Unattended Ground Sensor Network: A Correlated Equilibrium Game Theoretic Analysis

We model a decentralized network of decision makers charged with optimally deploying hardkill and/or softkill weapons (countermeasures (CMs)) to defend a task group from antiship missiles. Each platform (decision maker) observes missile threats using a combination of shipboard sensor measurement data and data from other ships, but acts independently when controlling its own weapon systems to cooperatively pursue the defensive objectives of the task group in a decentralized fashion. The main results are a formulation of the missile deflection problem as a stochastic shortest path Markovian game with constraints, a characterization of the Nash equilibrium solution, and a decentralized algorithm for computing the Nash equilibrium when the stochastic game is collaborative. Numerical examples are given to demonstrate the equilibrium policies and illustrate their sensitivity to the missile dynamics.

M. Maskery

Papers

Decentralized dynamic spectrum access for cognitive radios: cooperative design of a non-cooperative game

Decentralized Adaptive Filtering Algorithms for Sensor Activation in an Unattended Ground Sensor Network

Decentralized Activation in a ZigBee-enabled Unattended Ground Sensor Network: A Correlated Equilibrium Game Theoretic Analysis

Decentralized algorithms for netcentric force protection against antiship missiles

Game Theoretic Learning and Pricing for Dynamic Spectrum Access in Cognitive Radio