基于移动节点的无线传感器网络定位算法

Mobile node-based localization algorithm for wireless sensor networks

节点定位技术是无线传感器网络应用的重要支撑技术之一,为了提高定位算法的准确性,提出了一种基于移动目标节点的两步定位算法。该算法利用一个移动目标节点遍历整个网络,并周期性地广播包含自身当前位置的信息。而传感器节点的自身定位过程则可用基于无迹卡尔曼滤波(UKF)的目标跟踪方法实现。由于所用的目标状态模型和量测模型有一定的不确定性,所以先选取不共线的3个拥有RSSI测距能力的目标节点信息,利用Euclidean定位法提高滤波的初始位置精度,从而改善定位效果。通过仿真、分析和比较该目标节点在多种移动轨迹情况下的定位误差,这种两步定位法可以改善对目标节点移动轨迹的特殊要求的限制,能取得较好的定位精度,而且更适合于实际情况。

The node localization technology is a key support to wireless sensor network application, and a two-step localization algorithm is proposed based on a mobile destination node in order to improve the accuracy of position. A mobile destination node is used to traverse the network and broadcast current position information itself while the sensor nodes can get their position information with the Unscented Kalman Filter （UKF） target tracking method. As the target state model and measurement model have some uncertainty, we select three non-collinear destination nodes with RSSI ranging measurement capability, then get the improved initial position using the Euclidean filtering method, hence the final position more accurately. The final simulation, analysis and comparison of the positional errors in a variety of trajectory indicate that the improved two-step method can relax the special requirements of moving trajectory for the destination node, achieve better positioning accuracy, and be more suitable for practical situation.