Research On Multi-radar Maneuvering Target Tracking Algorithm In Observation Error Covariance Uncertain Environment

Maneuvering target tracking,which belongs to target tracking domain,is a kind of important information fusion problem.It refers to the target position and velocity estimating processes by using the measurements from different sensors,such as radar or infrared sensor.As a key technology in target tracking,maneuvering target tracking technology has been widely used in battlefield monitoring,land early warning,deep space exploration,fire control,industrial control,intelligent transportation and many other fields.With the rapid development of high performance weapon in recent years,the maneuvering target environment is getting more and more complicated,and the traditional target tracking technology is challenged seriously.Traditional maneuvering target tracking technology used to estimate the target state with one-way fusion mode by single sensor,which is not conducive for accumulation and reusing of information.Therefore,the tracking effects of the traditional maneuvering target tracking algorithms are restricted,as they are not good at dealing with the problems of maneuvering mode uncertainty or unknown observation error covariance.To this end,in this paper we research the problem of tracking a strong maneuvering target with multiple sensors in unknown observation error covariance environment.The main work and results are as follows:1)To solve the tracking effect degradation problem when observation error covariance is uncertain,a new method is proposed to estimate the observation error covariance.This method is established on the feedback fusion mechanism with the properties of unbiasedness and consistency.Then a feedback fusion based observation error covariance adaptive estimation of Kalman filtering algorithm(OECAKF)is proposed.The simulation results shows that,if the observational error covariance deviates from the reality,OECAKF gets more robust and higher estimation accuracy than some traditional algorithms.2)Traditional algorithms for maneuvering target tracking used to applied in single radar tracking system.It will limit the tracking effect for lacking the fused information from other radars.To solve this problem,a multiple models fusion algorithm for maneuvering target synergistic tracking by multiple radars(MR-MMST)is proposed.Compared to some traditional algorithms,the MR-MMST shows higher tracking precision when the target maneuvers.3)Consider using early warning satellite and radar to track the ballistic missile in simulations,three different maneuvering models are established for the ballistic missile,includes the active section,free section and reentry phase model.Then we use the MR-MMST algorithm to track the ballistic missile in above scene.4)Embed the OECAE-KF in MR-MMST,then we have multiple models fusion multiple models synergistic tracking algorithm in uncertain environment(UE-MMST).The new method enables MR-MMST to deal with the uncertainty of observation error covariance when tracking maneuvering target.Simulation results show that the proposed algorithm can obtain higher estimation accuracy when the observed error covariance is deviated from the reality.