Navigation Algorithm Combining Building Plans With Autonomous Sensor Data

This paper presents an approach to navigation system’s position and heading correction using building floor plans. The algorithm includes three steps: (a) autonomous sensors data processing to obtain position and heading, (b) map-matching correction, and (c) navigation system errors estimation. A particle filter is used to incorporate the building plan information and a Kalman filter estimates the dead reckoning error states. This algorithm was designed for vehicle navigation systems operating inside buildings with known floor plans and can be adapted for implementation on real-time navigation systems using low-cost MEMS gyroscope and speed sensor as dead reckoning instruments. The real-world data collected from the vehicle indoor tests has shown that the proposed algorithm is able to correct significant errors in dead reckoning position and heading by applying the map constraints.