Research On Blind Equalization For Communication System

In modern communication systems, transmission channel is greatly distorted because of exists of factors such as channel-fading and multi-path propagation emerges, which severely reduces the performance of transmission channel. When unknown signals are transmitted through the non-ideal channels, it will inevitably produce inter-symbol interference (ISI) and inter-channel interference (ICI), which is an important factor that degrades the performance of communication. In order to reduce ISI and ICI, equalize the characteristic of channels and recover transmitted signals, equalizer has to be used in the receiver. The traditional adaptive equalizers have to repeat transmit periodic training sequence to real-time tracking channel, which wastes plenty of bandwidth and decreases the transmission rate. Blind equalization is a new self-adaptive equalization technology. It doesn’t need training sequences to maintain normal work, only relies on the prior information of transmitted signals to equalize the channel characteristic, and makes sure that the output sequence can approximates the transmitted signals as accurate as possible. At the same time, it can raise the channel’s efficiency and get much better equalization ability. Therefore, it is more applicable to advanced communication systems; the super performance of blind equalization makes it becoming a very popular research topic.Bussgang algorithms is the most widely used blind equalization algorithm, Constant Modulus algorithm (CMA) and Decision Directed (DD) algorithm have the advantages of low computational complexity and strong stability, which are the special example of Bussgang type. In order to balance the contradictions of convergence rate and convergence precision, dual-mode algorithm is a classic application of blind equalization. With the advantages of low complexity, cold start capability and low steady-state residual error, CMA+DD-LMS dual-mode algorithm has been widely used. However, the algorithm still exists the phenomenon of low convergence speed, phase rotation, poor robustness, to solve these problems, a variable step size MCMA+DD-LMS dual-mode algorithm based on joint error function is proposed. The proposed algorithm can overcome the phase rotation, has fast convergence rate and make dual-mode switch more stable. The simulation result of16QAM system proves that the improved algorithm has steady switching ability, cold start ability, strong suppression ability of ISI, fast convergence rate, and higher convergence precision.Because of Baud-spaced equalizer exists the problem of non-desired local convergence point and easily lead to severe noise amplification through the channel with deep spectral nulls, a new blind equalization algorithm of T/2fractionally spaced (T/2-FSE) MMMA+MDD hybrid algorithm is proposed to improve the convergence rate and reduce mean square error. The introduction of T/2FSE can effectively improve the performance of the system; at the same time, using the multi-mode algorithm as the start algorithm have steady cold start ability and fast convergence speed; when residual error reduce to a certain level, MMMA+MDD hybrid algorithm is enabled to reducing steady-state residual error. The simulation with16QAM system shows that the improved algorithm has better convergence performance:cold start ability, strong suppression ability of ISI and real-time tracking the channel.