For wireless communications, forward error correction(FEC) schemes are an important tool for improving communications reliability. Turbo codes have been shown to perform near the Shannon's limit on the additive white gaussian noise(AWGN) channels. As a powerful coding technique, turbo code offers great promise for improving the reliability of communication over wireless channels.
Recently, the trend of wireless communication is changed from the conventional narrow-band voice service to the wide-band multimedia service. Therefore, it is highly required to develop the high-speed turbo decoder structure. An important problem in high-speed applications is decoding delay inherent to turbo decoding. Conventional Radix-2 MAP decoder has difficulty in applications of high-speed wireless communication mainly due to delay in interleaving/deinterleaving and iterative decoding.
To solve the problem with latency of turbo decoder, in this thesis a Radix-4 MAP decoding algorithm and many parameters were adopted such as branch metric, forward/backward state metric, etc. Simulation results over AWGN channel show that the BER performance of the Radix-4 MAP turbo decoding is almost the same as that of Radix-2 MAP turbo decoding.
Radix-4 turbo MAP decoder was compared with the conventional one in terms of decoding speed. The decoding speed of the Radix-4 MAP turbo decoder is faster by 2.4 times at least than conventional one in the case of one iteration of turbo decoding.