Indoor Intelligent Mobile Robot Localization Using Fuzzy Compensation and Kalman Filter to Fuse the Data of Gyroscope and Magnetometer

This paper develops an indoor intelligent service mobile robot that has multiple functions, can recognize and grip the target object, avoid obstacles, and accurately localize via relative position. The locating method of the robot uses the output values of the sensor module, which includes data from a gyroscope and a magnetometer, to correct the current rotation direction angle of the robot. An angle correction method can be divided into three parts. The first part calculates the angle values obtained from the gyroscope and the magnetometer that are installed on the robot. The second part explores the error characteristics between the sensor module and the actual rotation direction angle of the robot. The third part uses the error characteristic data to design the fuzzy rule base and the Kalman filter to eliminate errors and to get a more accurate orientation angle. These errors can be described as either regular or irregular. The former can be eliminated by fuzzy algorithm compensation, and the latter can be eliminated by the Kalman filter. The contribution of this paper is to propose an error correction method between the calculus rotation angle determined by the sensor and the actual rotation angle of the robot such that the three moving paths, i.e., specified, actual, and calculus paths, have more accurate approximation. The experimental results demonstrate that the combination of fuzzy compensation and the Kalman filter is an accurate correction method.