Multi-Sensor Fusion Floor Location Method Based on Extended Kalman Algorithm

This paper investigates an extended Kalman filter-based barometer and IMU fusion floor localization method, aiming to solve the problem of precise floor localization for mobile robots that need to take elevators. By establishing an index between spatial height and floor levels, the paper uses the extended Kalman algorithm to fuse filtered IMU acceleration data with barometric data optimized by the least squares method to obtain the robot's height data. Subsequently, the robot's floor data is obtained based on the floor index. Experiments show that the comprehensive height prediction error of this paper is within 0.10m, and the floor prediction accuracy reaches 98%. The research provides effective technical support for floor localization of mobile robots in elevator scenarios.