Research On GA-MUSIC Algorithm Based On DOA Estimation

Direction of arrival(DOA)estimation is an important technology for array signal processing research.With the continuous development of engineering technology,software and hardware equipment are becoming more and more mature,DOA is estimated to realize many industrial and military applications.At present,the research on the theory of high-resolution and high-precision DOA estimation technology has been successfully extended to the fields of rock cave mining,underground survey,space exploration and so on.At present,DOA estimation methods widely used in military industry and other aspects are mainly divided into two categories: beamforming and subspace classification.Among them,the multiple signal classification algorithm(MUSIC)algorithm is an important branch of subspace classification.The MUSIC algorithm can be divided into two parts: eigenvalue decomposition and spectral peak search,because spectral peak search requires a lot of time The real-time performance of the traditional MUSIC algorithm is limited.In this paper,genetic algorithm(Genetic algorithm)and MUSIC algorithm are combined to improve the real-time performance of the algorithm.Genetic algorithm is a global optimization bionics algorithm with the characteristics of simplicity,global optimization and parallelism.However,genetic algorithms also have shortcomings such as immature convergence and local optimal solutions.In order to improve the performance of genetic algorithm global search and eliminate immature convergence,this paper improves the genetic algorithm.Based on genetic algorithm,this paper proposes an optimization strategy based on operator and population.Among them,the selection operator adopts the optimal solution strategy of roulette betting;the crossover operator adopts the method of probabilistic crossover or multi-point crossover with the difference of the fitness of contemporary individuals;the mutation operator follows the difference of the contemporary population The proportion of the individual gene values of the population is balanced and varied to maintain the diversity of the population and improve the local search ability of the algorithm.While increasing the diversity of the population,the initial size parameter of the population is introduced,and the method of staged adjustment is used to control,so as to meet the needs of the algorithm for the population size.Simulation tests show that the improved genetic algorithm can overcome the immature convergence and solve the problem of falling into the local optimal solution,and has better performance.The improved genetic algorithm is combined with the MUSIC algorithm to propose the GA-MUSIC algorithm.In order to further study the reliability of the GA-MUSIC algorithm in DOA estimation,under the conditions of different signal-to-noise ratios,the spectral peak search of the GA-MUSIC algorithm was simulated under the conditions of single signal source and multiple signal sources in space.The simulation results show that the GA-MUSIC algorithm has good global search capabilities,and can search for single and multiple targets.The performance comparison with PSO-MUSIC algorithm shows that GA-MUSIC algorithm can greatly improve the accuracy and real-time performance.Related experiments were carried out in the indoor muffler pool,and the real-time performance of GA-MUSIC algorithm and PSO-MUSIC algorithm for single-source and non-coherent multi-signal source DOA estimation was analyzed.The test results show that the GA-MUSIC algorithm is ensuring successful search At the same time,it can effectively reduce the calculation time of spectrum peak search,and has better real-time performance.The GA-MUSIC algorithm simulation test proposed in this paper is consistent with the experimental results,which can effectively solve the problem of DOA estimated search success rate and real-time performance,and has important reference value for practical engineering applications.