BICM-ID And Its Application In MIMO-OFDM Systems

Bit-interleaved coded modulation with iterative decoding(BICM-ID) is an advanced technology which has the features of simple-structure, gaining with different channel states and easy-configuration. This dissertation majorly focuses on the performance of BICM-ID system with different channel coding and the problems in the combination of BICM-ID and MIMO-OFDM. The main topics include:1. EXIT charts and system simulation are used to analyze the performance of Turbo-BICM-ID system with different coding rates, coding length and interleaver on the conditions of different mapping schemes and channel states. According to our analysis, Turbo-BICM-ID system can hardly obtain the additional iterative gain by the joint iteration of decoding and demapping if the decoder has already obtained full iterative gain, and Gray mapping is the best mapping scheme on this condition. According to the analysis of convolutional codes BICM-ID and Turbo codes BICM-ID with different coding rates, four definitions are proposed to depict the performance of BICM-ID system:narrow coding gain, generalized coding gain, upper bound of narrow iterative decoding performance, upper bound of generalized iterative decoding performance. We present that for the wireless communication system which aimed to achieve the upper bound of generalized iterative decoding performance, if the upper bound of narrow iterative decoding performance is limited by some reasons, just pursuit the narrow coding gain to achieve the upper bound of narrow iterative decoding performance can not obtain more iterative gain via joint iteration of decoding and demapping, even there is a gap between upper bound of narrow iterative decoding performance and upper bound of generalized iterative decoding performance. However, the joint optimization of coding and modulation maximize the generalized coding gain and it will make the system get closer with the upper bound of generalized iterative decoding performance.2. Using sigma mapping based on superposition coding in LDPC-BICM-ID system, this mapping scheme makes the transmit signal more Gaussian-like by superimposing several different signals, and it can get iterative gain by BICM-ID since it is non-Gray. At receiver, the analytical expression for the demapping output can be derived by the way of IDMA, and based on it, we can get the optimized LDPC codes by density evolution. Finally, we get an optimized scheme of coded modulation in LDPC-BICM-ID and get a better performance than others.3. Combining BICM-ID and MIMO-OFDM, and link adaptation are also applied to MIMO-OFDM system with BICM-ID. Different from traditional adaptive modulation and coding (AMC), different mapping schemes are adopted with channel variation and obtained a good performance. By comparison with LTE, we found that MIMO-OFDM system with convolutional codes BICM-ID can get a better performance and lower complexity on the same condition. So the thinking of BICM-ID system design is further verified by this result.4. A new explicit CSI feedback scheme is put forward to really improve the spectral efficiency when link adaptation is used in the system. This scheme utilizes the sparsity of wide band wireless channel, the channel coefficient can be derived by FFT transformation with finite multipath information of the channel, then link adaptation scheme will be determined by the channel coefficient. Compared with traditional CSI feedback method via analysis and simulation, the proposed scheme provides high throughput performance while keeping the feedback overhead at a low level.