A time-differential measurement based algorithm for multi-sensor target tracking

A time-differential measurement based algorithm is provided for multi-sensor target tracking. This approach decouples the sensor registration and estimation by handling the original measurements. Thus, estimation can be carried out without considering sensor registration, and then, the registration errors are subsequently corrected. It is shown that this algorithm is more computation efficient than the augmented state extended Kalman filter (ASEKF). Simulation results show that this approach significantly reduces the sensor bias errors, and has better tracking performance than ASEKF in the following cases, 1) in high SNR scenario, 2) when the initial target state vector is uncertainty.