Obstacle Avoidance System in Autonomous Ground Vehicle Using Ground Plane Image Processing and Omnidirectional Drive

Abstract

Today one of the major concerns for any industry is availability of workspace and adequate intra-transport systems. This paper suggests the possible and probably cheap solution for both of above incapacities by implying sensor controlled motion of Omni-directional wheels. We have synchronized the image processing capability of Raspberry Pie processor with the sensitivity of motor driver circuit to obtain a precise and accurate motion of platform mounted with mecanum wheel that do not require any turning radius. As an image sensor we have used the simple CCD camera. The set of Image transformations are applied on the image to detect the obstacle precisely. We designed mecanum wheel with main constraint kept as shortage of space and load carrying capacity, hence we were able to report design of the wheel with specific dimensions. The image processing and simulation is done in the MATLAB environment. This choice has added the possibilities of the inclusion of advance control strategies and the use of image processing toolbox.