Research on SR_UPF ground moving target tracking algorithm using road information

A kind of tracking algorithm using square root unscented particle filter is proposed to solve nonlinear problem caused by maneuverability of target and low measurement update rate in ground moving target tracking. Unscented transform is used to transfer state estimation and covariance matrix, where the covariance matrix has been extracted. From this, importance sampling used in particle filter can be gained. Then, the state estimation which is obtained through particle filter is projected to road according to road information if it is off road. Using one ground moving target indication sensor to track a target on road, simulation results show that the proposed algorithm can effectively solve nonlinear problem in ground moving target tracking. Compared with the conventional particle filter algorithm, the proposed algorithm not only advances the tracking precision, also reduces the needed number of particles which mean less run time will be taken. Another way, tracking performance can be further improved due to using of road information.