The problem of tracking multiple targets simultaneously using a wireless sensor network is studied in this paper. We introduce a new algorithm, based only on the received signal power measurements, to estimate the location of multiple indistinguishable targets. For each node, a circle centered at the location of the node with a radius equal to the estimated distance between the node and the nearest target is drawn. The intersection points of all these measurement circles are calculated and clustered using a density-based clustering algorithm. The centroid of each generated cluster can be a candidate location, corresponding to a target. In order to choose the best candidate locations, we introduce a new robust criterion, which is capable of dealing with the problem of malicious nodes. Besides, the selected candidate is given to the Gauss–Newton iterative search method, which can increase the accuracy of tracking. We also propose three different approaches for reducing the effect of malicious nodes on the accuracy of tracking. Furthermore, a scheme is proposed for identifying the malicious nodes. We demonstrate the robustness and accuracy of our proposed tracking algorithm via simulation results and compare our results with the Multi-resolution search algorithm and the Expectation–Maximization algorithm.

本文研究了使用无线传感器网络同时跟踪多个目标的问题。我们仅基于接收到的信号功率测量结果引入一种新算法，以估计多个无法区分目标的位置。对于每个节点，绘制一个以节点位置为中心的圆，其半径等于该节点与最近目标之间的估计距离。使用基于密度的聚类算法计算并聚类所有这些测量圆的交点。每个生成的簇的质心可以是对应于目标的候选位置。为了选择最佳的候选位置，我们引入了一个新的健壮标准，该标准能够处理恶意节点的问题。除了，选定的候选对象将被赋予高斯-牛顿迭代搜索方法，这可以提高跟踪的准确性。我们还提出了三种不同的方法来减少恶意节点对跟踪准确性的影响。此外，提出了一种用于识别恶意节点的方案。我们通过仿真结果证明了我们提出的跟踪算法的鲁棒性和准确性，并将我们的结果与多分辨率搜索算法和Expectation-Maximization算法进行了比较。