Intelligent control & navigation of mobile robots

本研究的主要目标是研究和开发移动机器人的智能控制和导航体系结构，使其能够在动态和不断变化的条件下与人类和其他移动代理进行自主和协作操作。工业应用包括露天采矿、农业、林业、服务机器人、高速公路养护和建筑，这些领域都需要机器人在操作过程中应对动态和变化的环境。全自动和智能机器人导航将允许更好地利用时间和资源，从而提高许多工业应用的生产力。在动态条件下运行的机械（或移动设备）通常需要熟练的操作员，并且在大多数情况下，它们在环境发生意外变化的复杂环境中工作。本研究旨在开发与多智能体系统自主操作相关的控制和导航技术。该建议包括几个主要的研究组成部分；非结构化环境中机器人行为协调与导航的高级控制使用多个感官数据对环境动力学和移动物体进行建模；人机交互；以及机器人系统的多变量智能预测低级控制。这些问题是在机器人系统的感知和行动（控制）水平上解决的。模糊推理在低级和高级控制方面都有显著的改进。使用模糊上下文混合的行为调制或使用模糊离散事件系统的行为状态建模已经在机器人控制中显示出显著的改进。因此，在本研究中，基于模糊逻辑的推理系统被用于控制、导航和预测控制。

The main objective of this research is to investigate and develop intelligent control and navigational architectures for mobile robots to operate autonomously and cooperatively with humans and other mobile agents in applications where the operational environment is dynamic and subjected to changing conditions. Industrial application includes open-pit mining, agriculture, forestry, service robots, highway maintenance and construction, where robots are required to deal with dynamic and changing environments during operations. Fully automated and intelligent robot navigation would allow better utilization of time and resources resulting in higher productivity for many industrial applications. The machineries (or mobile equipment) operating under dynamic conditions generally require skilled operators and in most cases they work in complex environments where the surroundings are subjected to unexpected changes. This research aims at developing control and navigational techniques that are relevant for autonomous operation of multi-agent systems.The proposal includes several main research components; high-level control of behaviour coordination and navigation of robots in unstructured environments; modelling of environmental dynamics and moving objects using multiple sensory data; human robot interactions (HRI); and multi-variable and intelligent predictive low-level control of robotic systems. The issues are addressed at perceptual and action (control) levels of the robotic system.Fuzzy reasoning has shown significant improvements to both low-level and high-level control. Behaviour modulation using fuzzy context blending or modelling of a behavioural state using fuzzy discrete event systems has shown significant improvements in robotic control. Therefore, in this research, fuzzy logic based reasoning systems are used for control, navigation and predictive control.