碩士 / 國立中興大學 / 電機工程學系所 / 103 / This thesis proposes a novel multi-objective ant colony optimization (ACO) algorithm to optimize a fuzzy system (FS) with application to robot navigation control. The novel algorithm is called multi-objective cooperative continuous ACO with auxiliary colony optimization (MO-CC2ACO). The MO-CC2ACO consists of sub-colonies which cooperatively optimize an FS consisting of rules, where a sub-colony is only responsible for optimizing a single fuzzy rule. In addition, an auxiliary colony is created to store all the fuzzy rules in the best-so-far FSs. To evaluate the performance of each single rule in a sub-colony, a complete FS is formed with the accompanying rules selected from the best FS in the auxiliary colony. Performance of a fuzzy rule is ranked based on the multi-objective function values of its participating FS using the Pareto non-dominated sorting and the crowding distance computed in the auxiliary colony. New sub-solutions in each sub-colony are generated using the processes of ant-path selection, Gaussian sampling, and best-ant attractions. The MO-CC2ACO-designed FC is applied to multi-objective speed and location control of a mobile robot for obstacle boundary following (BF) in a navigation problem. A learning environment is created for learning the BF behavior without an exhaustive collection of input-output training pairs in advance. A fast navigation approach that consists of BF, target seeking (TS), convex point search (CPS), and behavior supervisor is proposed. The CPS is introduced to force the robot to move toward a sub-goal selected from the convex point(s) in the obstacle boundary to shorten the moving path. Performance of the MO-CC2ACO is verified through comparisons with various multi-objective population-based optimization algorithms. Simulation and experimental results verify the effectiveness and efficiency of the new robot navigation approach.
| Identifer | oai:union.ndltd.org:TW/103NCHU5441067 |
| Date | January 2015 |
| Creators | Chan-Hung Lin, 林展弘 |
| Contributors | Chia-Feng Juang, 莊家峰 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | en_US |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 56 |
Page generated in 0.0095 seconds