Physical Layer Security Research Based On Bayesian Coalitional Game Theory In Wireless Communication

Cooperative communication improves the physical layer security performance of data transmission among wireless network nodes in the presence of multiple eavesdroppers. However, the node's behavior in the coalition is either cooperative or not due to the environment the node located. Wireless network nodes do not have an exact and perfect knowledge of the cooperative behavior of the potential cooperators in coalition formation.In order to form the stable coalitional structure under behavior uncertainty of nodes, we propose the Bayesian Coalitional Games using partition model based on possible worlds(i.e., environments), which nodes enter into contracts(i.e., agreements) with one another to deal with the uncertain payoffs. We propose the Bayesian coalition formation algorithm to obtain the Bayesian core which is the Nash-stable and fair coalitional structure. A belief update mechanism to the nodes with cooperative behavior uncertainty is used to calculate the probability of each possible world. We also present a discrete-time Markov chain model to investigate the stability of the coalition structure. The coalition structure obtained in coalition game with incomplete information is the same as the one observed in coalition game with complete information. Our proposed Bayesian coalition formation algorithm can effectively estimate the behavior of the node, dynamically adjust the coalitional structure to maximize the security capacity of the nodes in the coalition.Cooperative relays in the coalition forwarding the received signals to the destination improve the security performance even with multiple eavesdroppers. However, the selfish relay has the incentive to refuse or only forward the received data with partially forwarding power, which decreases the security performance. Moreover, the selfish behavior is uncertain. We present a Bayesian coalition formation algorithm to investigate the coalition formation with uncertain selfish behavior of the potential cooperators. The selfish behavior is evaluated by the selfish degree which related to the payoff of the coalition. The simulation results indicate that there exists a selfish degree threshold.