AN EM APPROACH FOR JOINT CHANNEL ESTIMATION AND CHANNEL DECODING IN SYSTEMS EMPLOYING PHYSICAL-LAYER NETWORK CODING

This paper applies the expectation-maximization (EM) algorithm to address the problem of joint channel estimation and channel decoding in Physical-layer Network Coding (PNC) systems. The use of PNC can significantly improve the throughput of a relay network. The throughput advantage, however, is predicated on the availability of accurate channel estimates. For channel-coded PNC systems, a major challenge is that the maximum a posteriori probability (MAP) channel estimation is nontrivial due to 1) the overlapping of signals from multiple users received at the relay; and 2) the correlations among data symbols introduced by channel coding. In this paper, we show that an EM algorithm implemented on a factor graph framework is well suited to tackle this problem. Through iterative message passing, the channel estimation component and the channel decoding component in the factor graph interact to improve each other's results progressively. Simulation results indicate that just one EM iteration of our algorithm can significantly improve the channel estimation accuracy as well as the BER performance of channel-coded PNC systems.