Improving robot Darwinian particle swarm optimization using quantum-behaved swarm theory for robot exploration and communication / Duaa Mehiar

Duaa , Mehiar (2021) Improving robot Darwinian particle swarm optimization using quantum-behaved swarm theory for robot exploration and communication / Duaa Mehiar. PhD thesis, Universiti Malaya.

	PDF (The Candidate's Agreement) Restricted to Repository staff only Download (187Kb)
	PDF (Thesis PhD) Download (2033Kb)

Abstract

Despite the significance of the Robotic Darwinian Particle Swarm Optimization (RDPSO) algorithm on swarm-robot exploration and communication, there remain notable gaps such as premature and slow convergence, collisions between robots, and communication breaks and constraints. The quantum computing theory has several advantages that can improve the searching capabilities of PSO-based algorithms. However, there has yet an attempt to adopt quantum behaviour onto robot-based system such as the RDPSO. In this study, a new algorithm called the Quantum Robot Darwinian Particle Swarm Optimization (QRDPSO) is contributed with the hypothesis that quantum behaving particles can address the RDPSO main gaps; i.e. to improve the exploration and communication performance of a swarm robotic system. In terms of convergence time, the experiment done shows the QRDPSO algorithm is faster to reach an optimal solution than the RDPSO algorithm. The QRDPSO algorithm also shows tolerance to premature convergence compared to RDPSO. This study also contributed a distributed swarm navigation strategy that allows the QRDPSO robots to communicate directly with other robots in the swarm. Two popular communication schemas over wireless sensor network have been adopted and tested on the QRDPSO, the Multi-hop Routing Algorithm with Low Energy Adaptive Clustering Hierarchy (MR-LEACH) and the mobile ad hoc communication network (MANET). The QRDPSO algorithm with MR-LEACH consumes less power with energy consumption at 48% compared to the QRDPSO with MANET at 63%. Less power allows the MR-LEACH to increase lifetime for the nodes more than MANET while reducing interruptions between robots but not faster to reach the optimal solution than the QRDPSO algorithm with MANET. The QRDPSO with MANET needs 180 iterations, while the QRDPSO with MR-LEACH needs 202 iterations. The predecessor, RDPSO, needs 210 iterations for comparison to reach a victim. Given the importance of a swarm’s sustainability; swarm not losing robots, able to conserve energy and explore farther, the MR-LEACH schema is proposed to complement the QRDPSO communication.

Actions (For repository staff only : Login required)