Novel Fuzzy Systems for Human-Autonomous Agent Teaming

Publication Type:
Thesis
Issue Date:
2021
Full metadata record
The Multi-agent Teaming (MAT) systems that have been widely applied in many fields provide a novel method for establishing models, conducting the analysis, implementing complex tasks and so on. The agents in an MAT system can be defined as intelligent agents, machine agents and human agents based on a particular task to exhibit flexible behaviours. This research investigates various fuzzy models to resolve the problems of designing MAT systems. The main contributions are as follows. 1) For multiple-agent coordination, a hierarchical fuzzy system is proposed and applied to navigation and simultaneous arrival of mobile agents. 2) Explainable fuzzy systems are proposed. We developed an interpretable fuzzy model for human agents to understand the decision rules learned by machine agents and a fuzzy rule information visualisation framework for machine agents to understand human cognitive states. 3) Finally, the distributed fuzzy system is proposed to resolve the data privacy and high-dimensional data in designing MAT systems. A novel consensus learning is developed for the distributed fuzzy system to learn antecedent and consequent components.
Please use this identifier to cite or link to this item: