摘要
针对现有的目标跟踪算法不能兼顾精度和能耗的问题,提出了一种动态最近邻协作目标跟踪算法。本算法动态构建目标跟踪簇以更好地适应目标位置的实时变化,从而获取最佳的跟踪精度。当目标进入监控区域后,多个传感器节点感知到目标则自动成簇最小二乘方法初始移动目标的最初位置。引入预测机制,根据目标的未来位置,基于最近邻协作准则选择下一时刻的簇头节点。目标跟踪簇头节点在其邻居范围内选择任务节点观测目标位置,并采用集中卡尔曼滤波完成目标的状态估计。仿真结果表明:提出的目标跟踪算法具有跟踪精度高,节点间的单跳通信距离能够有效减少能耗。
To conquer shortcoming of equivalent between accuracy and energy consumption, a dynamically clustering algorithm, called nearest neighborhood collaboration, is proposed for collaborative target tracking in wireless sensor networks. The algorithm forms target tracking cluster to adapt the real time change of moving target for optimal tracking precision. When target moves into the monitoring area, multi-sensor nodes that can sense the target work collaboratively as a tracking cluster for initial position of target based on least square. Kalman prediction mechanism is introduced for calculation of moving target location. Cluster head is selected based on collaboration nearest neighboring criterion. Cluster head chooses the tasking nodes to observe target location in its neighborhood, and fulfills state estimation with central Kalman filter. Simulation results show that, compared with existing algorithm, the proposed algorithm not only achieve superior tracking accuracy but also energy saving due to single hop communication distance.
出处
《传感器与微系统》
CSCD
北大核心
2012年第7期135-139,共5页
Transducer and Microsystem Technologies
基金
国家自然科学基金资助项目(60870010)
关键词
目标跟踪
最近邻协作
卡尔曼滤波
无线传感器网络
target tracking
nearest neighborhood collaboration（NNC）
Kalman filter
wireless sensor networks
