Research On Blind Equalization Optimization Of Underwater Acoustic Channel Based On Breadth Enhanced Fireworks Algorithm

The competition of marine power is the international background of the development of marine science and technology,and marine information technology is the premise and foundation of the marine power strategy.However,the time-varying and multipath effects of underwater acoustic channels in the ocean are serious.The blind equalization technology can effectively compensate the non-ideal characteristics of underwater acoustic channels,overcome intersymbol interference,reduce bit error rate and improve the quality of underwater acoustic communication.However,the traditional blind equalization technology has some big shortcomings,such as,slow convergence speed and steady-state error.This paper proposes a breadth enhanced fireworks algorithm(BEFWA)which is used to optimize constant modulus blind equalization algorithm(CMA).The main research contents include the following three aspects:1.This paper proposes a BEFWA to improve the randomness and uniformity of population individuals.At the same time,the diversity of the population is ensured and local optimality is avoided.Fireworks algorithm is a swarm intelligent optimization algorithm,which is high efficiency,and has the ability to solve the global optimal solution of complex problems.But,it also has own shortcomings such as slow convergence speed,and poor stability.In view of the defects of the traditional fireworks algorithm,this paper proposes a BEFWA.The main innovations of this algorithm are as follows:(1).The good point set method is adopted in population initialization to improve the randomness and uniformity of population individuals.(2).In the selection of the next generation of fireworks,two selection strategies are proposed by using the location information of other fireworks with good fitness and balancing the local and global search ability: breadth-first selection strategy and good-degree preference strategy.It not only improves the search efficiency,but also improves the convergence speed.(3).By means of gaussian perturbation of the selected fireworks,the diversity of the population is further increased and local optimality is avoided.2.An CMA blind equalization algorithm based on BEFWA is proposed.It not only improves the convergence speed and stability of the algorithm,but also reduces the mean square error before and after equalization,improves the equilibrium effect.CMA algorithm uses random gradient descent method to solve the weight from the minimum value of cost function to the equalizer.The initial weight vector will directly affect the convergence of the algorithm.Therefore,an CMA blind equalization algorithm based on BEFWA is proposed,which uses the cost function of the mean square error CMA blind equalization algorithm between the balanced signal and the ideal noiseless signal is taken as the fitness function of the breadth enhanced fireworks algorithm,and optimizes the initial weight vector.This algorithm makes full use of the BEFWA's strong global random search ability and fast convergence speed,reduces the possibility of CMA algorithm falling into local convergence.Thus,it accelerates the convergence speed of the CMA algorithm,and reduces the steady-state error of the algorithm,improves the equilibrium effect,and greatly improves the real-time performance of blind equalization technology.3.The proposed BEFWA and BEFWA-CMA algorithms are evaluated by Matlab simulation software.This article uses eight standard test functions to compare the BEFWA with the fireworks algorithm(FWA),artificial colony algorithm(ABC),and grey wolf optimization algorithm(GWO).The results show that the convergence speed and stability of BEFWA are better than the other three optimization algorithms,especially the convergence speed increased by 50%,It is more suitable to optimize the initial weight vector of CMA.Several indexes are improved by combining the breadth enhanced fireworks algorithm with the underwater channel blind equalization algorithm.Simulation shows that the output constellation of BEFWA-CMA is more compact and clear than that of CMA,and the new algorithm has lower bit error rate.Simulation comparison of BEFWA-CMA with FWA-CMA,ABC-CMA and GEO-CMA shows that BEFWA-CMA improves the convergence rate by more than 25%.The mean square error of BEFWA-CMA after convergence under different SNR is lower than the other three algorithms,which shows that the equalization effect is better.