Department
Electrical and Computer Engineering
Major
Electrical Engineering
Research Advisor
Venayagamoorthy, Ganesh K.
Advisor's Department
Electrical and Computer Engineering
Funding Source
UMR Opportunities for Undergraduate Research Experiences (OURE) Program
Abstract
Collective Robotic Search (CRS) is useful in applications such as radioactive source detection where little to no human intervention is desired. In CRS, a group of intelligent mobile robots collectively explores a dangerous environment in order to locate and converge on a specified target. This project applies two search algorithms, Particle Swarm Optimization (PSO) and greedy search, to a CRS problem containing an environment with multiple targets and obstacles where the entire swarm of robots is needed to complete the desired task. The simulation results for both methods are compared on the following: simulated run time, number of evaluations, distance traveled by the robots, resilience against communication and robot loss, and target convergence percentage. Simulation results show that PSO performs better than greedy search in terms of convergence time, distance traveled, and adaptability, for CRS problems requiring the entire swarm of robots for collective mission completion.
Biography
Lisa Lorena Smith is a senior undergraduate student at the University of Missouri-Rolla, majoring in Electrical Engineering. Lisa is a member of the Rea/Time Power and Intelligent Systems Laboratory at UMR.
Research Category
Engineering
Presentation Type
OURE Fellows Final Oral Presentation
Document Type
Presentation
Location
Havener Center, Ozark Room
Presentation Date
11 April 2007, 9:30 am - 10:00 am
Comparison of Particle Swarm Optimization and Greedy Search for Collective Robotic Search in a Complex Environment
Havener Center, Ozark Room
Collective Robotic Search (CRS) is useful in applications such as radioactive source detection where little to no human intervention is desired. In CRS, a group of intelligent mobile robots collectively explores a dangerous environment in order to locate and converge on a specified target. This project applies two search algorithms, Particle Swarm Optimization (PSO) and greedy search, to a CRS problem containing an environment with multiple targets and obstacles where the entire swarm of robots is needed to complete the desired task. The simulation results for both methods are compared on the following: simulated run time, number of evaluations, distance traveled by the robots, resilience against communication and robot loss, and target convergence percentage. Simulation results show that PSO performs better than greedy search in terms of convergence time, distance traveled, and adaptability, for CRS problems requiring the entire swarm of robots for collective mission completion.
