Sum-Product decoding of convolutional code for wireless LAN standard

Abstract

This paper proposes decoding methods for the convolutional code for the wireless Local Area Network (LAN) standard by using sum-product algorithm (SPA). The code is non-systematic. For this code, the conventional SPA decoding cannot provide good decoding performance. This paper shows the reason. The reason is that the conventional SPA uses a Tanner graph including information bits, however, no received signals for information bits are available for the code. This paper proposes two-step SPA decoding method (Proposed Method 1: PM1). The PM1 is as follows. (1) Only parity bits are decoded by SPA. (2) With decoded parity bits, information bits are decoded by SPA. The PM1 can provide much better decoding performance in comparison to the conventional SPA. Furthermore, this paper proposes the combined use of PM1 and method of [4] (Proposed Method 2: PM2). The PM2 provides good performance near BCJR performance. At BER=10−5, PM2 performance is only 0.7[dB] inferior to BCJR performance.