Authors: George T. Amariucai, Shuangqing Wei
Addresses: Department of Electrical and Computer Engineering, Louisiana State University, Baton Rouge, LA, USA. ' Department of Electrical and Computer Engineering, Louisiana State University, Baton Rouge, LA, USA
Abstract: In this article, we adopt outage probability (λ-capacity) in fast-fading channels as a pay-off function in a zero-sum game between a legitimate transceiver pair and an uncorrelated Gaussian jammer. The transmitter aims at minimising the outage probability, while the jammer attempts to maximise the outage probability. We consider both peak (over each codeword) and average (over all codewords) power constraints. For peak power constraints, a transmission rate is either supported by the system, or if too large, causes the whole transmission to fail. By imposing average power constraints, large rates can be supported at the cost of positive probability of codeword error. Maxmin (where the jammer is assumed to play first) and minimax (where the transmitter is assumed to play first) power control strategies are developed under each power constraint, which show that no Nash equilibrium of pure strategies exists under the average power constraint when the follower has perfect knowledge about the randomisation outcome of the first player.
Keywords: communications systems; lambda capacity; fast-fading channels; wireless channels; jamming games; maximin; minimax; outage probability; power constraints; transmission rate; codewords.
International Journal of Autonomous and Adaptive Communications Systems, 2008 Vol.1 No.4, pp.411 - 424
Published online: 28 Nov 2008 *Full-text access for editors Access for subscribers Purchase this article Comment on this article