Space-time Joint Modeling And Target Tracking Method

In conventional multi-target tracking systems,the sampling interval of the sensor under consideration is assumed to be constant or previously known.However,in the TWS(Tracking while scanning)system,since the radar antenna obtains the target point information by the electromagnetic wave scanning method,there is relative motion between the antenna and the target during the antenna scanning,so the target and the target and the same target The revisit intervals between diff erent scan cycles are different and unknown.There are two main scanning methods for TWS radar.One is circumferential scanning and the other is sector scanning.For TWS radar with circumferential scanning,the exact value of the revisit interval of a target in one scan depends on the target state and process noise.This leads to significant modeling mismatches when using traditional tracking methods in TWS systems.Especially when tracking high-speed targets with low scan rate antennas,performance degradation,such as loss of track accuracy and track continuity,may occur.For the sector-scanning TWS radar,the target scanning interval will also affect the target scanning interval,except that the target state and process noise will affect the target scanning interval.For the same target,especially the target of the scan area boundary,the scanning interval of the antenna from clockwise to reverse is inevitably larger than the scanning interval of the antenna from counterclockwise to clockwise.In view of the above problems,this paper has done the following work on the time uncertainty of TWS radar:(1)For the problem of model mismatch in TWS radar tracking system,considering the nonlinear relationship between target state and target revisit interval,a tracking system with revisiting interval uncertainty is proposed by equations rather than a single equation.System model.(2)According to the scanning characteristics of the circumferential scan,based on the space-time relationship between the target state and the revisit interval,a TWS radar target tracking system model for circumferential scanning is established.An iterative method is proposed to calculate the target state and revisit interval simultaneously.This spatiotemporal joint processing is considered to be a general nonlinear function,so a UT transform(insensitive transform)is used to generate the statistical properties of the iterative transform output.Based on UT transform and iterative method,a new tracking filter is built to achieve accurate and consistent estimation in TWS.Finally,a multi-target tracker is constructed based on the proposed filter and linear multi-objective integrated probability data association(LMIPDA).(3)The scanning mechanism of TWS radar antenna fan sweep is analyzed,including uniform acceleration scanning mode and trigonometric function scanning mode.For these two scanning methods,the relationship between target revisit interval and target state and scanning direction is analyzed respectively.The system model gives a complete target state estimation algorithm and multi-target tracking method,and further analyzes the algorithm performance.Considering the particularity of the maneuvering target,the IMM(Multiple Model Estimation)algorithm is used to implement the state estimation algorithm for the maneuvering target under the fan sweep TWS radar.