A new interative receiver structure for DAB system

Abstract

The digital audio broadcasting (DAB) system provides audio quality comparable to compact disk (CD) without any impairment even in severe multipath reception situation. Also the DAB system provides various data services e.g. video, images and coded traffic messages. This system is based on orthogonal frequency multiplexing (OFDM) in conjunction with punctured convolution coding. In Eureka-147 DAB system, $\pi$/4 differential quadrature phase shift keying (DQPSK) modulation and convolutional code are used. It was recognized that the turbo principle can effectively be applied to differential modulation and coding since the advent of turbo code. In this technique, soft information is exchanged between demodulator and decoder. In this thesis, we first suggest optimum iterative demodulation and decoding algorithm using Log-MAP (LMAP). Iterative process shows a lot of performance improvement compared to existing DAB. LMAP algorithm is optimum but has high complexity. To reduce complexity, we use Max-Log-MAP (MLMAP) algorithm which is simplified version of LMAP and suggest several suboptimum iterative methods. The approximation of MLMAP cause some performance degradation. To Satisfy both good performance of LMAP and low complexity of MLAMP, we propose a low complexity iterative demodulation and decoding algorithm with good performance. In the proposed algorithm, we find proper scaling factors and multiply them to extrinsic values of MLMAP demapper and MLMAP decoder. This proposed algorithm is close to the optimum algorithm with LMAP demapper and LMAP decoder in terms of performance. In addition, the complexity of the proposed algorithm is almost same to that of low complexity scheme with MLMAP demapper and MLMAP decoder.