Please use this identifier to cite or link to this item: https://www.um.edu.mt/library/oar/handle/123456789/92142
Title: Path following robot
Authors: Chetcuti, Dale (2021)
Keywords: Mobile robots
Fuzzy logic
PID controllers
Reinforcement learning
Issue Date: 2021
Citation: Chetcuti, D. (2021). Path following robot (Bachelor’s dissertation).
Abstract: Robotics has seen an increase in interest for several applications such as transportation. These include public transport and self driving cars which require following a path. Path-following is a task that deals with robots following a predefined path from a starting position to an end goal. However, path-following poses different challenges. One main challenge is smooth movement. Obtaining smooth movement is a requirement in path-following, as this decreases the error which in larger scale projects can be fatal if the error is above a threshold. In this project we investigate different techniques that can be adopted by a robot to follow a predefined path. Additionally, we investigate ways of rendering smooth movement. We used a Lego Mindstorms EV3 Robot Educator with a colour sensor attached to the front in the centre. We set up the environment using a black path on a white background for maximum contrast. Using the colour sensor, the robot can identify whether it is on the black path or the white background. We investigated three algorithms that are commonly used in path-following robots, and compared their performance on four different paths. The first approach uses a Proportional Integral Derivative (PID) Controller, which is commonly used for similar projects as it achieves smooth movement. The second approach uses Q-Learning, whereby the robot attempts to learn from its actions to navigate the predefined path efficiently. Thirdly, we used Fuzzy Logic method, which makes decisions based on a predefined set of rules. To evaluate the performance of the algorithms, we attached a marker to the front of the colour sensor to trace the path followed by the robot. We computed the root mean square error (RMSE), between the predefined path and the actual path taken by the robot. We used this evaluation method for the PID controller and Q-Learning algorithm with the PID controller yielding smoother movement. The Fuzzy Logic algorithm was based on a simulation and the RMSE and mean error were higher than the other two algorithms although a direct comparison cannot be made.
Description: B.Sc. IT (Hons)(Melit.)
URI: https://www.um.edu.mt/library/oar/handle/123456789/92142
Appears in Collections:Dissertations - FacICT - 2021
Dissertations - FacICTAI - 2021

Files in This Item:
File Description SizeFormat 
21BITAI020.pdf
  Restricted Access
1.99 MBAdobe PDFView/Open Request a copy


Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.