A high throughput limited search trellis decoder for convolutional code decoding

Due to the lack of operational parallelism and structured data storage/retrieval, limited search trellis decoding algorithms have been traditionally ruled out for applications demanding high throughput convolutional code decoding. Among various limited search algorithms, the T-algorithm performs breadth-first limited search and has good potential for parallel decoding. In this paper, we propose two techniques at the algorithm and VLSI architecture levels for the T-algorithm to improve the decoding parallelism and tackle the data storage/retrieval problem, which enables the high throughput path-parallel T-algorithm decoder VLSI implementation. This work provides a vehicle for exploiting the merits of the T-algorithm, i.e., low computational complexity that is adaptive to the channel distortion, in high throughput applications.