中文摘要：

提出一种基于空间策略网格图的符号级联合信源信道编译码快速算法。构建联合译码的平面网格图，可以实现比比特级译码更好的译码性能。然而平面网格图结构复杂，使得符号级联合信源信道译码过程复杂度高。提出一种新的高效空间策略网格，通过满足限制条件方程对可变长符号的权重值进行优化，结合可变长符号后验概率译码，形成一个整体快速编译码算法。两项仿真实验的结果表明，相比平面网格图，该方法减少译码复杂度近50％，并使系统在相同功率信噪比条件下传输的比特错误率更低。这种方法可以很好地保护可变长编码后的信息数据，适用于资源受限的环境中的符号级联合信源信道编译码。

英文摘要：

Joint source channel symbol-level en/decoding fast algorithm based on space policy trellis is presented in this paper. Constructing a joint decoding plane trellis achieves better decoding performance than the bit-level decoding method. However, the complicated construction of plane trellis causes a high degree of complexity of the joint source channel symbol-level decoding. In this work, we construct a new efficient space policy trellis, optimize design of weight value of variablelength symbols by meeting the constraint condition system of equation, and then form an integrated fast algorithms with VLS-APP decoding. The results of two kinds of simulations indicate that our fast algorithm leads to near 50% reduction of the decoding complexity and lowers bit error rate in the same SNR of channels, compared to the decoding method based on plane trellis. So this method provides reliable error-protection mechanism for variable-length encoded information data, and it can be applied in joint source channel symbol-level err/decoding under resource-constrained situation like the space communication.