Line Tracking Algorithm Based on Computer Vision Leveraging Stereo Camera to Support Decision of Automated Guided Vehicle Navigation

An Automated Guided Vehicle (AGV) is a specialized vehicle capable of autonomous movement and performing specific tasks. AGVs are commonly utilized in the warehouse area to transport some materials from one location to another. Some research have been conducted on the development of AGV, but there are still some problems such as inability to detect the changing line ahead of the AGV platform and the effect of light illumination that can obscure the guide lines trajectory. To overcome these issues,this research has developed an alternative line tracking algorithm based on the contour-based search method, leveraging a stereo camera. This algorithm was tested on an AGV system consisting of ZED 2i stereo camera, Jetson Xavier NX computing unit and Jetracer platform. Experiments were done by varying the Jetracer speed from 30 to 65 cm/s to evaluate the performance of the algorithm. According to the experimental results, the Jetracer platform is able to follow the guide line path with small average distance bias ranging from 3.85 - 9.12 cm from the actual trajectory as reference data. As a result, this experiment indicates that the developed line tracking algorithm is suitable for effective operation within the low to medium speed range.