| Author: | Fan, Bowen |
| Title: | Development of ant-inspired foraging methods for multi-agent robotic systems |
| Advisors: | Navarro-Alarcon, David (ME) |
| Degree: | M.Sc. |
| Year: | 2023 |
| Subject: | Robots -- Control systems Intelligent agents (Computer software) Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Mechanical Engineering |
| Pages: | x, 69 pages : color illustrations |
| Language: | English |
| Abstract: | This dissertation describes a novel foraging methodology that draws inspiration from the behaviour of ant colonies in navigating through unknown environments. Emulating the ants, we develop a system that allows agents to follow pheromone trails to locate food or detect potential threats, as well as to move in a linear fashion. Unlike traditional approaches that rely on complex kinetic models, our system employs fuzzy logic to control the agents’ movements based solely on the pheromone values they receive from the computer. The agents do not communicate with one another, but instead make individual decisions based on the pheromone information they receive, allowing them to construct paths from their starting point to the food source. This distribution system is not only simple to implement but also highly effective in achieving its intended purpose. In order to construct this foraging system, we have divided it into three distinct parts: agent state update rules, agent movement rules, and pheromone update rules. Each agent has two distinct targets, requiring it to have two separate states in order to match each target accordingly. To facilitate the movement of the agents, we have implemented a fuzzy logic controller to regulate their movements. As the agents move, they also deposit artificial pheromones on the map, which gradually fade over time. The agents rely on the pheromones and the fuzzy rules to make decisions about whether to turn or continue moving in a straight line. Through this process, the agents are able to construct a path from their starting point to the food source, effectively mimicking the behaviour of ant colonies in similar situations. Each agent has five sensors to avoid collision. Besides that, we design an offset pheromone method which is designed to minimize the chances of agents disturbing each other while following their respective paths. The offset of the pheromone method can build different pheromones on different paths. This can help to differentiate between the paths and minimize the chances of agents colliding with each other while following their respective pheromone trails. The combination of sensors and pheromone methods can be used to create a robust and efficient system for guiding agents towards different targets while avoiding collisions and decreasing interference between agents. The experiment we conducted demonstrates the effectiveness of our method in solving foraging problems with agents. Initially, the agents move randomly across the map, but eventually, they converge and begin constructing a path towards the food source. The use of offset pheromones plays a crucial role in this process, as it enables the agents to avoid face-to-face collisions and facilitates the overall convergence of the system. Overall, our results indicate that the method is highly effective in solving foraging problems and has significant potential for other applications as well. |
| Rights: | All rights reserved |
| Access: | restricted access |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 7468.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 20.92 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/12988