Vector Approximate Message Passing Based Iterative Receiver for OTFS System

Abstract

In this paper, we develop an iterative message passing receiver based on factor graph for orthogonal time frequency space (OTFS) systems over doubly selective channels. By factorizing the joint probability of the data symbols and the received signals, we construct a loop-free factor graph with vector-valued nodes. Then, vector approximate message passing (VAMP) is employed to derive message updating expressions of variables on factor graph. To improve the accuracy of symbol detection, we vectorize the scalar variance terms of VAMP. Based on the singular value decomposition of block circulant channel matrix of OTFS system, we introduce the average approximation to circumvent the matrix inversion operation and then develop a low-complexity VAMP-based algorithm. Simulation results demonstrate the effectiveness of the proposed algorithms, which outperform the existing minimum mean square error algorithm and the message passing algorithm in terms of bit error rate (BER) performance.