Cooperative Tracking for WSN With DOA-Based Factor Graph Technique

This article proposes a cooperative tracking scheme in wireless sensor networks (WSNs), which utilizes a direction-of-arrival (DOA)-based factor graph (FG) approach. It focuses on a scenario that sensors served as the anchor nodes in the WSN are insufficient, such that the performance of tracking may significantly degrade, and even cause severe detection ambiguity with the conventional angle-based schemes. To address this issue, a cooperative target is invoked, which is capable of measuring the DOAs of the main target in real time, despite that its own position is assumed to be unknown. By exploiting such additional DOA measurements besides those obtained by the sensors, the proposed system has exhibited clear performance gains over the non-cooperative scheme, in terms of lower root-mean-square errors (RMSEs). Moreover, the detection ambiguity encountered by the target entering some special areas has been effectively resolved, such that the reliability of the system has been significantly enhanced. Finally, the robustness of the proposed system has been evaluated, which utilizes the predicted Cramér-Rao lower bound (P-CRLB) as the estimated variance of the observation in the extended Kalman filter (EKF).