Efficient MAP-algorithm implementation on programmable architectures

Abstract. Maximum-A-Posteriori (MAP) decoding algorithms are important HW/SW building blocks in advanced communication systems due to their ability to provide soft-output informations which can be efficiently exploited in iterative channel decoding schemes like Turbo-Codes. Multi-standards demand flexible implementations on programmable platforms.

In this paper we analyze a quantized turbo-decoder based on a Max-Log-MAP algorithm with Extrinsic Scaling Factor (ESF). Its communication performance approximate to a Turbo-Decoder with a Log-MAP algorithm and is less sensitive to quantization effects. We present Turbo-Decoder implementations on state-of-the-art DSPs and show that only a Max-Log-MAP implementation fulfills a throughput requirement of ~2 Mbit/s. The negligible overhead for the ESF implementation strengthen the use of Max-Log-MAP with ESF implementation on programmable platforms.