Put others before itself: A multi-leader one-follower anti-jamming stackelberg game against tracking jammer

This paper mainly investigates the coordinated anti-jamming channel access problems in multiuser scenarios where there exists a tracking jammer who senses the spectrum and traces the channel with maximal receiving power. To cope with the challenges brought by the tracking jammer, a multi-leader one-follower anti-jamming Stackelberg (MOAS) game is formulated, which is able to model the complex interactions between users and the tracking jammer. In the proposed game, users act as leaders, chose their channel access strategies and transmit firstly. The tracking jammer acts as the follower, whose objective is to find the optimal jamming strategy at each time slot. Besides, the existence of Stackelberg equilibriums (SEs) is proved, which means users reach Nash Equilibriums (NEs) for each jamming strategy while the jammer finds its best response jamming strategy for the current network access case. An active attraction based anti-jamming channel access (3ACA) algorithm is designed to reach SEs, where jammed users keep their channel access strategies unchanged to create access chances for other users. To enhance the fairness of the system, users will adjust their strategies and re-learn after certain time slots to provide access chances for those users who sacrifice themselves to attract the tracking jammer.