B-Coop: A Novel Cooperation Enforcement Scheme for Wireless Networks

Abstract

This paper proposes a cooperative retransmission scheme B-Coop implemented at the MAC layer in wireless networks. Cooperative retransmissions can be viewed as a way to achieve virtual space-time diversity without any modification to the wireless physical layer. We model nodes in the wireless network as selfish and rational users, and apply principles of game theory to enforce cooperative retransmissions in a distributed manner. We also consider that nodes cannot always observe other nodes' actions perfectly, in order to make the decision whether to cooperate with others. Hence our scheme is based on imperfect private monitoring and it shows the existence of an efficient belief-based sequential equilibrium under certain conditions. B-Coop enforces cooperation among nodes by creating a fear of punishment and it achieves scalability by reducing an N-player game into a 2-player game, which is novel according to existing literature. Well-defined utility functions are formed to improve the throughput and energy expenditure of the nodes in the network. Through extensive simulations, we show that B-Coop achieves increased throughput with less retransmissions and less energy expenditure. We verified using simulations that our proposed B-Coop scheme achieves better throughput and energy expenditure values.