摘要
针对非线性多目标模型,应用粒子滤波算法,这种方法不受模型线性和Gauss假设的约束,是一种处理非线性非高斯动态系统状态递推估计的有效算法。在粒子滤波的基础上融合扩展卡尔曼滤波算法和无迹卡尔曼滤波算法。融合后的新算法在计算提议概率密度分布时,粒子的产生充分考虑当前时刻的量测,使得粒子的分布更加接近状态的后验概率分布,再用平滑算法处理滤波的结果。仿真结果表明,算法有较好的跟踪效果。
Particle filter algorithm is proposed for nonlinear multi-target state space models and is an effective algorithm for the state recursive estimation in nonlinear and non- Gaussian dynamic systems. Under the theory framework of particle filter, a new algorithm is presented,which combines the particle filter algorithm with extend Kalman filter algorithm and unscented Kalman filter algorithm,when it calculates the proposed probability density distribution, the sampling particles can utilize the system current measures. That gets the particles distribution more approach to the station posterior distribution. Then smooth algorithm is processing the result of filtered. The simulations show that this algorithm has better tacking effect.
出处
《计算机与数字工程》
2009年第3期65-67,77,共4页
Computer & Digital Engineering
基金
湖北省重大科技专项基金项目(编号:2007DA111)资助
关键词
粒子滤波
扩展卡尔曼滤波
无迹卡尔曼滤波
机动目标跟踪
particle filter, unscented Kalman filter, extended Kalman filter, maneuvering target tracking
