论文部分内容阅读
针对传感器网络中每个传感器节点的邻接节点状态估计值不确定度不同的问题,提出一种基于不确定度量化加权的一致性卡尔曼滤波算法(CKF).该算法通过考虑节点度数对于传感器网络估计精度的影响,结合节点度数提出了一种衡量邻接节点状态估计值不确定度的量化函数,并把量化值作为该邻接节点与当前节点的状态估计值偏差的融合权重引入一致性协议中,利用优化后的一致性协议对传感器节点先验估计值进行更新,可提高一致性卡尔曼滤波算法的估计精度;算法同时具有非一致性误差小和鲁棒性强等特点.最后在3种不同网络类型下,通过动态目标跟踪实验仿真验证了算法的有效性.
Aiming at the problem that the uncertainty of the state estimation value of adjacent nodes in each sensor node in sensor network is different, a uniform Kalman filter algorithm (CKF) based on uncertainty quantification weighting is proposed. This algorithm considers the node degree for sensor network A new kind of quantitative function is proposed to measure the uncertainty of the state estimation of adjacent nodes by combining the degree of nodes with the quantization accuracy. The quantization value is introduced into the consensus protocol as the fusion weight of the deviation of the state estimation value between the neighboring node and the current node, Using the optimized consensus protocol to update the priori estimates of sensor nodes can improve the estimation accuracy of the consistent Kalman filter algorithm.The algorithm has the characteristics of small inconsistent error and robustness.Finally, Under the network type, the effectiveness of the algorithm is verified through dynamic target tracking experiment.