Target Tracking Fusion Filtering Based On Multiple Radar Platform Bias Registration

In recent years,target tracking technology has been widely used in guidance,navigation,control and other fields.With the development o f the target tracking technology,the methods relying on single radar has been unable to meet the requirements of target tracking accuracy.Therefore,multi-radar cooperative target tracking methods have been paid more and more attentions.Multi-radar cooperative target tracking methods obtain more reliable fusion information than the ones relying on single radar,by fusing the measurements from multiple radars.It achieves more accurate target position estimate and better fault tolerance and anti-interference ability.However,the discreteness of the components of the radar platform,the positioning errors of the radar platforms themselves,and the aging and wear that occur during use can cause system bias of radar platforms,which affects the moving target fusion tracking performance.Therefore,the target tracking fusion filtering methods based on bias registration of multiple radar platforms are studied in this paper.(1)In the non-maneuvering target tracking scenarios of multiple radar platforms with position biases or rotation biases,and the scenarios of multiple radar platforms with position biases and rotation biases,two kinds of novel non-maneuvering target tracking methods based on bias registration of multiple radar platforms are proposed respec tively.When multiple radar platforms have position biased or rotation biases with respect to the reference radar platform,firstly,each radar measurement in polar coordinates is represented as a pseudo-linear measurement in Cartesian coordinates,by usin g the unbiased conversion technique.Then the state estimates of the non-maneuvering target are obtained by utilizing Kalman filter with the measurements from the reference radar platform and the biased ones.Afterwards,based on the idea of least squares curve fitting,the position bias or rotation bias of the biased radar platform is estimated and registered.Finally,a distributed parallel fusion filtering method is used for fusion tracking of the non-maneuvering target.When multiple radar platforms h ave both position bias and rotation bias,with respect to the reference radar platform,firstly,the target state is estimated with the reference radar measurements by using UKF.The pseudo measurement function is constructed by substituting the target sta te estimate into the measurement function of biased radar.Then the unscented strong tracking filter with unbiased converted measurements is presented to estimate and register the position biases and rotation biases of the biased radars.Finally,the regis tered radar platforms and the reference radar platform are used to achieve non-maneuvering target fusion tracking by using the distributed parallel fusion filtering method.(2)In the maneuvering target tracking scenarios of multiple radar platforms with position biases or rotation biases,and the scenarios of multiple radar platforms with position biases and rotation biases,based on the idea of joint state and parameter estimation,two kinds of maneuvering target tracking methods based on bias registration are addressed.When multiple radar platforms have position bias or rotation bias,firstly,the joint state and parameter estimation method is adopted to estimate the maneuvering target state,with the unbiased converted measurements from the reference radar platform and the biased ones.Then,the pseudo-measurement information of the position bias or rotation bias is constructed.On this basis,a radar bias estimate and registration method is presented on the basis of the least squares curve fitting technique.Finally,the registered biased radars and the reference radar are utilized to track the maneuvering target in the distributed fusion framework.When multiple radar platforms have both position bias and rotation bias,firstly,the unscented strong tracking state and parameter joint estimation method is studied to estimate the sta te and the maneuver acceleration of the maneuvering target.By substituting these estimates into the biased radar measurement function,then,a pseudo-measurement equation of the radar position bias and the radar rotation bias is constructed to estimate an d register the biases of the biased radar platforms.Finally,the maneuvering target is tracked by using the unscented strong tracking state and parameter joint estimation method with the unbiased converted measurements from the registered biased radar pla tforms and the reference radar platform.