Indoor positioning systems have become a feasible solution for the current development of multiple location-based services and applications. They often consist of deploying a certain set of beacons in the environment to create a coverage volume, wherein some receivers, such as robots, drones or smart devices, can move while estimating their own position. Their final accuracy and performance mainly depend on several factors: the workspace size and its nature, the technologies involved (Wi-Fi, ultrasound, light, RF), etc. This work evaluates a 3D ultrasonic local positioning system (3D-ULPS) based on three independent ULPSs installed at specific positions to cover almost all the workspace and position mobile ultrasonic receivers in the environment. Because the proposal deals with numerous ultrasonic emitters, it is possible to determine different time differences of arrival (TDOA) between them and the receiver. In that context, the selection of a suitable fusion method to merge all this information into a final position estimate is a key aspect of the proposal. A linear Kalman filter (LKF) and an adaptive Kalman filter (AKF) are proposed in that regard for a loosely coupled approach, where the positions obtained from each ULPS are merged together. On the other hand, as a tightly coupled method, an extended Kalman filter (EKF) is also applied to merge the raw measurements from all the ULPSs into a final position estimate. Simulations and experimental tests were carried out and validated both approaches, thus providing average errors in the centimetre range for the EKF version, in contrast to errors up to the meter range from the independent (not merged) ULPSs.

中文翻译：

---

室内超声定位多传感器融合方法评价

室内定位系统已成为当前多种基于位置的服务和应用发展的可行解决方案。它们通常包括在环境中部署一组特定的信标以创建覆盖范围，其中一些接收器，例如机器人、无人机或智能设备，可以在估计自己位置的同时移动。它们的最终精度和性能主要取决于几个因素：工作空间大小及其性质、所涉及的技术（Wi-Fi、超声波、光、RF）等。这项工作评估了基于 3D 超声波本地定位系统 (3D-ULPS) 的在安装在特定位置的三个独立 ULPS 上，以覆盖几乎所有工作空间并将移动超声波接收器定位在环境中。因为该提案涉及众多超声波发射器，可以确定它们与接收器之间的不同到达时间差 (TDOA)。在这种情况下，选择合适的融合方法将所有这些信息合并到最终位置估计中是提案的一个关键方面。线性卡尔曼滤波器 (LKF) 和自适应卡尔曼滤波器 (AKF) 在这方面被提出用于松耦合方法，其中从每个 ULPS 获得的位置合并在一起。另一方面，作为紧耦合方法，扩展卡尔曼滤波器 (EKF) 也用于将来自所有 ULPS 的原始测量值合并到最终位置估计中。进行了模拟和实验测试并验证了两种方法，从而为 EKF 版本提供厘米范围内的平均误差，