Seminar: Priority-Based Human-Swarm Interaction Applied to a Foraging Application

Dalia Said
M.Sc. Candidate
Supervisor: Dr. Andrew Vardy

Priority-Based Human-Swarm Interaction Applied to a Foraging Application

Department of Computer Science
Monday, October 31, 2016, 2:00 p.m., Room EN 2022


Abstract

A robot swarm employs a large number of robots to facilitate sophisticated problem-solving as well as improved load-balancing and time efficiency compared to single robot systems. Swarm Intelligence is based on the local interactions between agents of the swarm that enables the emergence of a desired global behaviour, thus allowing the swarm to be autonomous. While autonomy is efficient for straightforward applications, in complex problems and environments human intervention may be more efficient. Human control of a swarm remains an open problem with multiple approaches proposed, each designed for a specific type of application. This work suggests a priority-based approach inspired by well known Human-Swarm Interaction techniques. The approach aims to serve as a high-level guide for the agents of a swarm, allowing them to use Swarm Intelligence on a low level. It also allows the division of the swarm into subswarms that can be easily controlled separately.

Before experiments could be carried out to validate the proposed approach, the robots used in our experiments had to be put together, and their software needed to be designed to put to use their various components. A vision system upon which their sensing is dependent needed to be established, with defined visual markers and obstacle detection.

This thesis tests the proposed priority-based Human-Swarm Interaction system by implementing a simple foraging application, using simulated and real robots, and studying the effects of introducing such a system to a group of robots that use simple Swarm Intelligence.

Results show that the proposed approach does succeed in dividing the swarm into subswarms and increasing the efficiency of the foraging solution, however, some drawbacks manifested themselves throughout the process. We discuss these advantages and issues as well as future work.

 

Contact

Department of Computer Science

230 Elizabeth Ave

St. John's, NL A1B 3X9 CANADA

Tel: (709) 864-2530

Fax: (709) 864-2552

becomestudent@mun.ca