Manoeuvring of AGV Using Vision System and PID Controlled DC Motor

Tan, Yi Ming (2018) Manoeuvring of AGV Using Vision System and PID Controlled DC Motor. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text TAN YI MING - FULL THESIS.pdf Restricted to Registered users only Download (2MB)

Abstract

The projects aims to fabricate an automated guided vehicle (AGV) that is able to navigate using line following system with the aid of camera. The AGV must be able to recognize the path and navigate precisely at different speeds. Image processing algorithm is implemented for path detection and recognition. The Raspberry Pi camera will first perform real time image capturing of the path. Different image algorithm methods such as blurring module, grey scale conversion, image thresholding, contour detection and centroid detection are implemented to detect the path. The path for the AGV is detected and recognized successfully with contour drawn along the boundary of the path. The centroid allows the AGV to understand its current position on the line. The AGV uses 3-wheeled differential steered mechanism to manoeuvre. The 2 DC motors at the front of the AGV is used to drive the rubber wheels while the free rotating wheel at the back is used to keep the body in balance. The motor speed is controlled by PWM with different percentage of duty cycle. PID control algorithm is implemented to improve the manoeuvring of the AGV. The steering of the AGV is dependent on the position of Cx and Cy of the centroid. The AGV is able to detect and recognize the path with different illuminance value ranging from 10 lux to 103 lux. The optimum threshold value for the vision system to detect and recognize the path accurately is in the range of 60 to 130. The AGV executes corrective algorithm when moving out of track with tolerance of +/- 3cm. However, it is not able to navigate perfectly on the path with small corner radius. The AGV is able to reduce its speed and stop precisely when approaching the end of the path with tolerance of +/- 3cm. This project contains improved performance on the conventional affordable AGV in the market. The vision system using image processing algorithm has replaced the implementation of IR sensors in conventional line following AGV which have several disadvantages. The manoeuvring of the AGV is improved with PID control algorithm to enhance the precision and the stability of the AGV at various speed.