Electrically Powered Wire Localization System for Mobile Robots

Among the technologies for mobile robot navigation, electrically powered wire-based navigation has the advantages of resisting environmental condition variance and readily sensing. In corresponding sensing systems, anisotropic magnetoresistive (AMR) sensors are desirable due to their excellent performances such as extreme sensitivity, wide linear range, low noise, and highly reliability. This paper proposes several positioning algorithms based on such sensing system, using signals from a single AMR sensor and from a pair of sensors. Dual-sensor algorithms can remove influence of the navigation current strength and the deflection angle between the induction magnetic field and the surface of the sensors. Algorithm parameters are evaluated offline from experimental data by least square method (LSM). A sensing system with four AMR sensors is developed and mounted on a car-like robot for testing. Static and dynamic positioning experiments are conducted to test accuracy of the sensing system. Experimental results verify the effectiveness of the sensing system and the positioning algorithms.