2009 Volume 71 Issue 4 Pages 4_43-4_51
In terms of obstacle avoidance and autonomous navigation, it is a very significant issue for a mobile robot to identify its positions. An algorithm for an autonomous vehicle utilizing the real-time kinematics differential global positioning system (RTK-GPS) and an omnidirectional camera was proposed as navigation sensors in this research. The goal of the paper is to develop a method that can estimate the absolute heading angle of the robot with respect to the world coordinate system. The developed algorithm is based on a stereo vision system. It estimates the rotation directly from spherical images and without correspondences. Instead, it uses a cross correlation method because camera rotation doesn't affect the field view and scene will remain the same. Horizon lines of pair spherical images has been extracted and used as the related periodic signals. The estimated error of heading angle is less than two degrees, and by not using heavy calculation image processing techniques, the process time for heading angle estimation becomes fairly faster to calculate in real-time application of robot vehicle navigation.
