IMM算法在车辆机动目标跟踪中的应用

Design of an Interacting Multiple Model Algorithm for Vehicle Target Tracking

针对高速公路车辆的机动目标跟踪问题,采用交互多模型算法(IMM)中的2个模型分别表示车辆的匀速运动状态和匀加速运动状态,并结合目标运动模型对目标当前加速度和其方差进行蒙特卡洛仿真。仿真结果表明该算法不仅可以在匀速运动时将关键测量噪声减低,而且在机动模型中保证状态估计量比未滤波的雷达测量值精确,同时可以对运动模型进行准确的识别,从而改善路面机动目标的跟踪性能,提高车辆的安全性和可靠性。因此,交互多模型算法可以满足高速车辆机动目标跟踪的要求。

The interacting multiple model algorithm is used to solve the vehicle＇ s tracking problems in automated highway system. It has multiple models, each of which matches to a particular mode such as uniform motion with constant velocity and maneuvering motion with acceleration. Combined with the target maneuvering models, the current acceleration of the target and its covariance are simulated. Significant noise reduction is achieved during the uniform motion. And it does better in maintaining the accuracy of the state estimates than the unfiltered radar measurements during the maneuver. A rapid detection of the maneuver is also obtained. The Monte Carlo simulation results show that the IMM algorithm will improve the target＇ s tracking performance and enhance the vehicle＇ s safety and reliability.