A novel approach to maneuvering target tracking based on random motion model using random Kalman filtering

Abstract

Traditional maneuvering target tracking algorithms assume that the target motion model is one fixed or a limited number of them. For high-speed and strong maneuvering targets, when the model set cannot cover the maneuvering mode or the deviation is large, the performance of the tracker will drop rapidly. Therefore, this paper proposes a new maneuvering target tracking method – a random motion model based on Random Kalman Filtering (RKF). This algorithm uses a random model to describe the target maneuver, which is more widely used than traditional algorithms and is more stable when the target maneuver is not covered by the model set. Compared with the traditional single model, the classic related algorithm is Kalman Filtering (KF), the new method significantly improves the tracking effect when the target is maneuvering. At the same time, when the model set of the Interacting Multiple Model algorithm (IMM) does not match the real maneuvering state, the tracking error of the new method is smaller than IMM and there is no divergence trend.