Research On Blind Equalization Algorithms For High-order Amplitude And Phase Modulation Signal

In communication system, the inter-symbol interference(ISI) plays a great impact on the reliability of the received signal. In order to reduce the error rate, it is necessary to insert the equalizer at the receive. The traditional equalizer needs to transmit training sequences repeatedly which leads to the reduction of the rate of information transmission. Blind equalization which relys on the prior information of the received signal doesn't need any training sequences to realize channel equalization. Due to the advantages of spectrum and power efficiency, the high-order amplitude and phase modulation signal has been widely used. Therefore, it is of great value to design blind equalization algorithms of the high-order amplitude and phase modulation signal.Accordance with problems such as the corresponding blind equalization algorithms for different channels, algorithms performance and implementation complexity, the main works of this thesis are as follows.1. For channels with different eigenvalue spread, a channel identification algorithm based on correlation of the equalizer input signal is proposed. Based on the tradeoff between complexity and performance of algorithms, the feasibility of SWA or MCMA for different eigenvalue spread is analyzed. Dual-mode algorithm based on constellation-matched error is proposed for further improvement on performance and the new algorithm has a faster convergence speed and lower steady state error.2. Improved algorithms for MCMA and SWA are given. Inspired by the affine projection and ?_CMA algorithms, the MCMA_APA and ?_MCMA algorithms are put forward. Modified SWA algorithm(MSWA) is proposed in order to solve the problem of phase ambiguity of SWA algorithm. The convex combination of MSWA with RLS with better performance is given which includes the advantages of MSWA and convex combination algorithms. Meanwhile, simplified MSWA_RLS convex combination algorithm with lower complexity is proposed.3. Blind equalization algorithms for short or burst data and lower complexity are put forward. Owing to the poor convergence of short data, short data-reusing algorithm based on data reliability is proposed. For 16 QAM signal, the algorithm has a faster convergence speed by about 3000 symbols and a lower residual ISI by about 5 d B, compared with data-reusing MCMA-based algorithm. With DCD algorithm applied to the blind equalization, the improved algorithm called MSWA_DCD for MSWA is put forward and the complexity decreases from(7)(8)2O N to O(7) N(8). Dual-mode MSWA_RLS_DCD algorithm based on constellation-matched error improves the equalization performance compared with MSWA_DCD_LMS algorithm without increasing the complexity.