Research of self-localization method of mine rescue robot

Components of self-localization system of mine rescue robots was introduced, self-localization method based on dead reckoning was introduced and improved. The improved method uses limited memory Kalman filter with exponential fading factor, and fuses data of multiple sensors in a stepwise manner. The Matlab simulation results show that the improved method can stably achieve tracking accuracy of 0.5 m. Experiments on mine rescue robot entity were carried out to justify the validity of the proposed self-localization method of mine rescue robot.