Abstract
Skid-steering mobile robots suffer from slip effect inevitably during their turnings, which results in imprecise kinematics model and the degradation of navigation and control performances. Hence, in this paper, we aim at developing an online estimation method to acquire the robot's instantaneous centers of rotation (ICRs), a kind of slip parameters, by means of data fusion technologies. The sensor system is composed of two incremental encoders, a compass, a Global Positioning System (GPS)unit, a camera and a data fusion unit. Based on the data gathered from these sensors, the data fusion unit is able to provide accurate global location, absolute heading and robot's ICRs in real time by applying the proposed terrain adaptive innovation-based extended Kalman filter. With the aid of terrain vision, the process noise covariance can be adjusted according to the terrain type adaptively, and therefore, the ICR estimation converges rapidly and smoothly. The real-world experiment conducted on a four-wheel mobile robot is exhibited to validate the effectiveness. Additionally, the results show that the terrain adaptive odometry has higher accuracy than the traditional ones.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
