摘要
Viterbi作为一种最大似然译码算法广泛应用在数字地面视频广播中,但由于其较高算法复杂程度,对实现高速低功耗时延小且逻辑结构简单的译码器带来了挑战。首先为了实现高速的Viterbi译码器,ACSU采用全并行结构,度量值的溢出控制采用取模归一化方法,并简化比较器。其次为了实现低功耗时延小且控制逻辑简单的Viterbi译码器,SMU采用改进的前向追溯结构,只用一组单口的RAM实现译码输出。该译码器在Xilinx Virtex6上实现并验证通过,并具有较好的译码性能。
Viterbi decoding is widely used in DVB-T communication system as a kind of maximum likeli-hood algorithm. For its high algorithm complexity, it brings many challenges to the realization of decoder with high speed,low power-consumption,little delay and simple logic structure. Firstly, in order to imple-ment high-speed Viterbi decoder, the structure is parallel in the ACS Unit, and the measurement of over-flow control is done with modulus normalization method,and meanwhile the comparator is simplified. Sec-ondly,in order to implement low power-consumption Viterbi decoder with simple control logic, SMU a-dopts the improved trace forward architecture and only a single RAM is used for decoding output. The de-coder is simulated and tested on Xilinx Virtex6 FPGA development board, and has good performance.
出处
《通信技术》
2014年第3期324-329,共6页
Communications Technology
基金
国家科技重大专项(No.2012ZX03001033)
中央高校基本科研业务费专项资金资助(No.0800219174)~~
