Automated Storage and Retrieval Systems Integration (AGV)

Lai, Eric Weng Lec (2017) Automated Storage and Retrieval Systems Integration (AGV). Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Eric Lai Weng Lec Full text.pdf Restricted to Registered users only Download (2MB)

Abstract

This dissertation proposes on a method where it allows AGV’s to localize itself through a more robust machine vision algorithm while using MQTT protocols for wireless communications between the control centre and the AGV. The AGV was tasked to travel to the desired location and using the data obtained by the measured trajectory from the control centre (Raspberry Pi) through extracted image from a low-cost webcam. The data was conveyed through WiFi using MQTT broker’s services that was powered by the Raspberry Pi. The data was published by the Raspberry Pi and the ESPresso will act as the subscriber to receive these data. Arduino Uno was connected to the ESPresso using serial communication and was responsible for controlling the AGV itself. Prior to the main function, a test of the AGV was done to determine the accuracy of the AGV using the dead reckoning test and was hypothesized that the cumulative positioning error of the AGV will increase as the distance travelled and the turns made by the AGV increases. The machine vision algorithms were tested that it has lower effects from illumination, colour, shape distortion and orientation during object tracking when introduced different methods of image filtering.