Research On Target Tracking Based On Firefly Algorithm

The main battlefield of modern warfare has developed from ground to air.Due to the massive research and development of air raid weapon,air defense has become an important issue to ensure national security.Fire control system provides operational support for air defense,which can search and find targets,identify enemy and ourselves targets,determine pre-damaged targets,and at the best time control the weapon launching warhead to block or directly hit pre-damaged targets.In this process,the accuracy of tracking pre-damaged targets directly will affect the effectiveness of defensive operations.Due to the increasing maneuverability of the target,and the large amount of interference noise in the target measurement information provided by the detection device in the fire control system,in order to accurately predict the future position of the target,it is necessary to filter out the large amount of noise in the measurement information.The prediction of the future coordinates of the target can be transformed into the filtering of the current state,and the accuracy of the target tracking will directly affect the calculation accuracy and the warhead hit rate of the fire control system.Therefore,this paper mainly studies the target tracking algorithm in the fire control system,with the particle filter as the main filtering algorithm,and optimizes the particle filter by the improved firefly algorithm to propose the target tracking algorithm in different tracking background.The main content of this dissertation includes:1.In view of the problem of particle impoverishment caused by the process of resampling,and effective state estimation requires a large number of particles.Firefly algorithm optimized particle filter is proposed.According to the operation mechanism of particle filter,the improved algorithm modifies the optimization method of firefly algorithm,which avoids the high computational complexity caused by the interaction between firefly algorithm and particle filter.The mechanism of eliminating the fittest of firefly population and the attraction and movement of firefly individuals are introduced into particle filter.Meanwhile,using the latest measurement,the particle simulated firefly individual moves to the high likelihood region intelligently,thus improving the overall quality of the particles.2.Interacting multiple model particle filter algorithm needs more models and particles to track maneuvering target accurately.To slove this problem,an interacting multiple model particle filter method based on firefly algorithm for maneuvering target tracking is proposed,which introduces the improved firefly algorithm optimized particle filter into interacting multiple model tracking algorithm,and improves tracking accuracy and stability of interacting multiple model algorithm by intelligent optimizing.With a relatively small number of models and particles to achieve the required accuracy,the comprehensive performance of maneuvering target tracking algorithm is improved.3.Aiming at the problem of particle weight degradation and particle impoverishment in particle filter track-before-detect algorithm.A new track-before-detect method based on the improved firefly algorithm optimize particle filter is proposed.The method using the optimal particle guides the movement of particle swarm.By introducing the elastic mechanism of spring,a spring model is constructed between the optimal particle and any particle,which solves the problem of particle aggregation around the optimal particle,makes the particle distribution more reasonable,and makes the improved algorithm have better detection and tracking performance for infrared small and weak target.4.Aiming at the high computational complexity of joint probabilistic data association algorithm based on particle filter.An adaptive firefly algorithm optimized particle filter is proposed,which uses the adaptive firefly algorithm to optimize the particle filter and achieves the desired accuracy with fewer iterations and particles.Then combined the improved firefly algorithm with the joint probabilistic data association algorithm to ensure the tracking accuracy and reduce the computational complexity of the algorithm.5.Aiming at the difficulty in extracting the peaks of the updated PHD function in SMC-PHD filter,a new method for extracting the peaks of the updated PHD function is proposed.This method selects the particles equal to the number of targets to form firefly individuals,one firefly is a solution of multi-target states,introduces attraction and movement mechanism of the firefly algorithm,uses the optimal firefly guiding other fireflies' movement,and undergo combination and mutation operation for the optimal firefly.After a few iterations,Search the optimal solution in the whole solution space and obtain the stable multi-target states.