An improved implementation of MAX^* operation for Turbo decoder

Abstract

An improved method to approximate max^* operation with Taylor Series for Turbo decoder is presented. Multiple expansion points are adapted to improve the BER performance. The parameter δ is introduced to determine suitable expansion points. The simulation results show that the proposed method with three expansion points when δ is set to 0.025 has almost identical performance compared with ideal Log-MAP algorithm and outperforms both PWL method and the original Maclaurin series method. The architecture of proposed method is also presented for implementation. Compared with PWL methods, the proposed scheme has reduced computational complexity and is feasible for hardware implementation. Besides, the architecture of proposed method keeps identical for different number of expansion points.