Segmenting maneuver identifier for maneuvering target tracking

The segmenting track identifier (STI) is an approach for maneuvering target tracking, which partitions the track into segments, optimizes the parameters of each segment, and obtains the state estimation. This paper proposes a segmenting maneuver identifier (SMI) based on the STI and makes it adaptable to additional maneuvering types. We present the parametric function of a new model, modify the cost function of the combined segments and the break condition, and improve the segment initialization method. The simulation shows the SMI provides better performance for tracking an accelerating target while the performance remains the same as the STI for tracking a steady-state target. It is also shown that the SMI performs as well as an interacting multiple model algorithm constructed with appropriate motion models, and much better than the Kalman filter based on the traditional uniform models for tracking highly maneuvering targets.