Skip to Main Content
Indoor position tracking systems are essential to support new types of applications for domotics and elderly care services. Unfortunately, while locating moving objects (e.g., people in a room) typically requires accuracy in the order of a few tens of cm, the intrinsically crowded nature of indoor environments (e.g., due to the presence of obstacles and/or multiple targets) as well as manifold sources of uncertainty may considerably degrade measurement results. In this paper, we present a local positioning system (LPS) provided with wireless connectivity. The proposed solution relies on two cascaded extended Kalman filters. The first one estimates the attitude of the platform within a global reference frame. The second one relies on the estimated attitude to return the planar position of the moving object in a room. The proposed approach is much more scalable than centralized location tracking techniques (e.g. based on external cameras only) because it does not require collecting and processing large data sets in real-time. Also, just low-rate position corrections are needed to keep uncertainty within given boundaries. Such position calibration values, measured with any type of external positioning infrastructure, can be sent to the LPS through a low-cost radio link like in a Wireless Sensor Network (WSN), at no risk of saturating the communication channel even when multiple objects are present in the room.