长期演进系统中一种低复杂度球形译码算法

摘要

The sphere decoding algorithm has the optimal Bit Error Ratio ( BER) performance that approximates to Maxmun Liklihood ( ML) in Long Term Evolution ( LTE) system. Concerning the computational complexity and required hardware resources of this algorithm increase significantly for detection of 16-QAM and 64-QAM modulated signal streams, an improved sphere decoding algorithm, which changed symbol search strategy, was proposed. A given symbol search scheme at different detection layer, and combined with a new definition for sphere radius of dynamic modifications was adopted in this algorithm. Both of the traditional and improved algorithms were simulated on the condition of Rayleigh fading channel. The simulation results show that the improved algorithm has a small BER degradation, and it also effectively reduces both computational complexity and required hardware resources compared to the traditional sphere decoding algorithm.%在长期演进(LTE)系统中,球形译码算法拥有接近于最大似然(ML)的误码率(BER)性能.针对在16QAM和64QAM等高阶调制情况下球形译码算法计算复杂度和所需硬件资源的急剧增加,提出了一种调整符号搜索策略的改进型球形译码算法.该算法在不同的检测层采用特定的符号搜索方案,并结合一种基于信噪比的动态调整半径方法.在无线瑞利信道环境下,对各种球形译码算法进行了仿真.仿真结果表明,提出的改进型算法基本保持传统球形译码算法较低的BER性能,同时还有效地降低了计算复杂度和硬件实现复杂度.