Optimal design of serial-parallel Viterbi decoder with multi-rate convolutional codes

In order to reduce the hardware complexity, the structural characteristics of the Viterbi decoder were studied. The branch measure unit was optimized by analyzing the characteristics of convolutional codes. Result show that the calculated branch metric value reduces from 16 to 4. The parameters of trace back unit can be configured by the backtracking algorithm with high flexibility and speed. Besides, the multi-rate punctured convolutional codes of CCSDS are acomplished by the same hardware architecture. 60% of the resources are saved compared with the normal serial-parallel structure with the same decoding speed. The throughput is as 8 times as the normal serial-parallel structure with the same hardware resources.