一种简化的交互多模式算法

通过对交互多模式(IMM)算法在机动目标跟踪中的分析,发现该算法在实现过程中滤波模式的确定和模式间转移概率的确定,要求具有一定的经验知识,对不能获得准确先验知识的研究带来了一定的困难。针对这种情况,研究了一种简化的IMM算法跟踪模型,该模型通过检测目标是否发生机动,自适应调整Kalman滤波(KF)的部分参数,从而实现不受IMM算法中有限模式集合限制的机动目标跟踪。通过仿真实验,证明了该算法在机动目标跟踪性能方面与IMM算法具有相似的效果,而在计算复杂度和对先验知识的要求方面有所降低

Simplified interaction multiple model algorithm

Based on the analysis of interaction multiple model (IMM) algorithm, it is found that some prior knowledge is required to get the Kalman filter model and Markov transition probability. It's difficult for those researchers who couldn't get them correctly. A simplified IMM algorithm is developed to deal with this situation. It can adaptively adjust some KF parameters based on whether it detects the happening of maneuvering targets or not. It could track maneuvering targets effectively while overcoming the restriction of finite mode sets in IMM algorithm. Simulation results illustrate that the performance of the proposed algorithm is similar to that of IMM algorithm but the complexity and prior knowledge are not so severel as those of IMM algorithm.