Localizing mobile robotic nodes in indoor and GPS-denied environments is a complex problem, particularly in dynamic, unstructured scenarios where traditional cameras and LIDAR-based sensing and localization modalities may fail. Alternatively, wireless signal-based localization has been extensively studied in the literature yet primarily focuses on fingerprinting and feature-matching paradigms, requiring dedicated environment-specific offline data collection. We propose an online robot localization algorithm enabled by collaborative wireless sensor nodes to remedy these limitations. Our approach's core novelty lies in obtaining the Collaborative Direction of Arrival (CDOA) of wireless signals by exploiting the geometric features and collaboration between wireless nodes. The CDOA is combined with the Expectation Maximization (EM) and Particle Filter (PF) algorithms to calculate the Gaussian probability of the node's location with high efficiency and accuracy. The algorithm relies on RSSI-only data, making it ubiquitous to resource-constrained devices. We theoretically analyze the approach and extensively validate the proposed method's consistency, accuracy, and computational efficiency in simulations, real-world public datasets, as well as real robot demonstrations. The results validate the method's real-time computational capability and demonstrate considerably-high centimeter-level localization accuracy, outperforming relevant state-of-the-art localization approaches.

翻译：在室内及无GPS环境中定位移动机器人节点是一个复杂问题，尤其在动态、非结构化场景下，传统相机与基于LIDAR的感知及定位方法可能失效。无线信号定位虽在文献中被广泛研究，但主要聚焦于指纹匹配与特征匹配范式，需要针对特定环境的离线数据采集。为解决上述局限，我们提出一种基于协作无线传感器节点的在线机器人定位算法。该算法的核心创新在于通过利用几何特征与无线节点间的协作，获取无线信号的协作到达方向（CDOA）。将CDOA与期望最大化（EM）算法及粒子滤波（PF）算法相结合，可高效、精确地计算节点位置的高斯概率。该算法仅依赖于RSSI数据，因此适用于资源受限设备。我们从理论上分析了该方法，并在仿真、真实公共数据集以及真实机器人演示中，充分验证了所提方法的一致性、精度与计算效率。实验结果证实了该方法的实时计算能力，并展示了显著的高厘米级定位精度，优于相关最先进的定位方法。