The JPDA (Joint Probabilistic Data Association) algorithm has been reported to be suitable for tracking multiple targets in dense environment. M{sup}3JPDA (Multiple Maneuver Model JPDA) is the extension and improvement of the JPDA for highly maneuvering target tracking. In the M{sup}3JPDA, a target motion is postulated by multiple models with their probabilities of being correct, what is called model probabilities. But, when A models are defined and L targets are tracked, M{sup}3JPDA should calculate N{sup}L model probabilities, then the computation load should be heavy. In this report, we propose a high-speed M{sup}3JPDA where the calculations for the model probability is reduced to NL. The computer simulation results have demonstrated the tracking performance and the computation load.
展开▼
