针对无线传感器网络定位误差大、有色噪声影响严重等问题,提出一种基于自适应抗差Kalman滤波的无线传感器网络节点跟踪算法.先采用直线拟合削弱初始点位观测值的横向误差,然后通过构造自适应因子和等价权函数来控制未知有色噪声和观测值粗差的影响,并利用观测环境构造可信度因子干预滤波计算.在低速、高速运动模型下,分别以直线运动和正弦曲线运动两种运动模式为例进行了仿真研究,仿真结果表明,该算法平均跟踪点位精度较EKF算法和UKF算法有较大提高,并对未知噪声和观测异常具有较强的适应性和容错性,能有效应对传感器节点机动运动,具有良好的跟踪效果.%Aiming at the problems of localization error and colored noise, a tracking algorithm has been presented based on adaptively robust Kalman filtering principle. Firstly, the transverse error of observation value was reduced using linear fitting method, and then unknown colored noise and gross error were controlled by constructing selfadaptive factor and equivalent weight function, and reliability factor was constructed according to observation environment to intervene filter computing. Simulating results show that the average position error of tracking is well improved compared with extended Kalman filtering and unscented Kalman filtering in two motion model of line and sine curve under low and high speed separately, and the algorithm has the advantages of high tracking precision,adaptability and fault tolerance for unknown noise and gross error which are effective on maneuvering motion and have good tracking effect.
展开▼
