Joint Source Channel Symbol Level En/Decoding Based on Space Computational Intelligence Policy Trellis

Abstract

Joint source channel symbol-level en/decoding based on space computational intelligence (CI) policy trellis is presented. Through constructing a joint decoding plane trellis, it can achieve better decoding performance than the bit-level decoding algorithm. However, the plane trellis is complicated, which results in a high decoding complexity for decoding symbol - level Turbo codes. To solve this problem, we construct a space CI policy trellis and its optimization design of weight value by using constraint condition system of equation in this paper and design a low-complexity joint decoding with variable length symbol-A Posteriori Probability (VLS-APP) algorithm. Simulation results show that the proposed approach reduces the decoding complexity compared to the plane trellis, and the gain is about 1.3 dB, and it provides substantial error protection for variable-length encoded image data, which can be applied to the joint symbol-level en/decoding of streaming media in resource-constrained space communication.