Minimax causal transmission of Gaussian stochastic processes over channels subject to correlated jamming

This paper has studied the problem of transmitting a Gaussian stochastic process {θt, t ∈[0, tf]} over a continuous-time channel which is subjected to an unknown jamming noise under a given power constraint. The noise process has been allowed to be correlated with the message process, and also a noiseless feedback has been allowed between the receiver and the transmitter. Under a quadratic distortion measure, we have obtained a saddle-point solution with probabilistic encoder structure, linear receiver structure, and a least favorable jamming policy given by a Wiener intergal.