Signal Sparse Decomposition Based On Joint Intelligent Algorithms And MP

Because of its excellent characteristics, sparse representation has been applied to many aspects of signal processing, such as signal denoising, coding, recognition and etc. Even as matching pursuit (MP) algorithm, which has the lowest complexity among sparse decomposition algorithms, the computation of that method is still great. Therefore, sparse decomposition of signals in the practical application is too difficult to be industrialized.The paper has mainly studied on Tabu Genetic Algorithm, Artificial fish-swarm algorithm and genetic algorithm joint algorithm for one-dimensional signal. The main tasks as follows:(1) Signal sparse decomposition taking advantages of Tabu Genetic algorithm. Based on characteristics of Genetic algorithm, Tabu search algorithm and joint intelligence algorithm, the paper proposes the algorithm of Tabu Genetic algorithm in order to solve the problem of signal sparse decomposition. Experimental results show that the algorithm enhances the capabilities for searching the best possible atom in each step of MP decomposition, and it also improves the speed of decomposition sparse.(2) Signal sparse decomposition taking advantages of algorithm of improved artificial fish-swarm algorithm combined with genetic algorithm. First, the paper analyses commonly used characteristics of atoms in dictionary and makes use of the different horizon of artificial fish and the reservation of the best individual strategy artificial fish-swarm algorithm. Second, it combines genetic algorithm with artificial fish-swarm algorithm running for signal sparse decomposition. Experimental results show that the algorithm avoids the disadvantages of solo algorithm and boosts the efficiency.