ROBUST PRECODER DESIGN FOR MIMO PACKET RETRANSMISSIONS OVER IMPERFECTLY KNOWN FLAT-FADING CHANNELS

This work investigates the design of robust precoders for ARQ packet retransmissions in Multi-Input Multi-Output (MIMO) systems under channel uncertainty. In practice, it is common for transmitters to only have imperfect Channel State Information (CSI). More specifically the channel uncertainty is characterized by the Frobenius norm between the actual and estimated channel. Given a known bound on the channel knowledge error, our objective is to sequentially design the robust precoder for each (re)-transmission that minimizes the Mean Square Error (MSE) for the worst case channel within CSI error bound. This min-max problem is solved by Lagrangian optimization. The robust precoders are proposed in this paper with simulation results demonstrating the worst case performance gain.