An Improved Current Statistical Model with Maneuver Detection

摘要

Current statistical model (CSM) performs well in maneuvering target tracking, however it suffers low accuracy when the target does not or weakly maneuver. To improve the performance of CSM in target tracking, we propose a new improved current statistical model with maneuver detection (ICSM-MD). The new model reduces the tracking error in weak or non-maneuvering situations by employing a specifically designed arctangent function to adaptively adjust the maximum acceleration, but it worsens the tracking performance for strong maneuvering targets. To overcome this deficiency, strong tracking filter (STF) is employed in ICSM. Moreover, to track more accurate, a maneuver detection scheme is adopted. The state estimate and covariance matrix are properly modified if a maneuver is detected, which reduces the variance of acceleration estimate. Consequently, ICSM-MD has a higher tracking precision and converges more quickly when the target acceleration jumps. Simulation results demonstrate the high performance of ICSM-MD in both strongly and weakly maneuvering target tracking.