Comparing an interacting multiple model algorithm and a multiple-process soft switching algorithm: equivalence relationship and tracking performance

In this paper, we show that a mathematical relationship exists between the multiple-process soft switching (MPSS) algorithm and an interacting multiple-model (IMM) algorithm. Assuming that each model transition probability is equally likely, the constraints imposed on the mixing and fusion gains in the IMM algorithm are equivalent to those in the MPSS algorithm. Thus, the constraints derived from a probabilistic approach can be reduced to those obtained via a deterministic approach. Unlike the MPSS algorithm with the aid of an additional constraint, the proposed IMM algorithm uses the real-time information of the innovations and their covariances to select filter gains between 0 and 1. The MPSS and the proposed IMM algorithms are applied to tracking a maneuvering target. Simulation results show that the latter has a better position tracking capability than the former when the measurement noise is small or moderate; however, both have similar performance in position tracking when the measurement noise is large. In view of the simulation results, more insights on the performance of MPSS, the proposed IMM and IMM algorithms are provided in the paper.