Investigation into communication and control networks for cellular robotics applications

该项目将建立在现有的模块化机器人系统设计的基础上，重点是自主行为集成到模块化系统中，作为一个更大的研究方案的一部分，以概率和知识为基础的推理，自我意识的自我配置的自主系统。该项目的目的是提供一个分散的系统，在实验室中开发的机器人模块完全基于知识的推理和通信能力。模块将能够在任何可能的物理配置中相互通信。知识作为语义基础概念和基础概念的连接而被共享，以使整个系统能够在其任务的上下文中从其元件中学习，并基于语义推理方法产生紧急的和自适应的系统行为。将开发在模块化硬件元件上运行的软件节点系统，以及与所创建的模块化机器人相关的基类的自适应本体，其将用作模块之间的通信协议。将收集关于不同联网、推理、知识共享和基于目标的任务完成情况的数据，这些数据将用于进一步改进系统内的知识，以用于未来的任务完成。目前还没有模块化机器人系统实现一个正式的语义知识收集和共享系统，可以推理和世界范围内很少有项目实现真正的基于知识的推理机器人。该博士学位将为高度新颖的基于知识的分布式系统奠定基础，并应用于机器人和自主系统。这项研究计划的最终目标是构建基于推理和估计的“细胞”机器人实体，并通过推理将基本的机器人感知数据转换为关于机器人自身状态及其环境的复杂结论，从而实现一种形式的“常识”。配置机器人将使英国处于模块化机器人研究这一基本未开发领域的最前沿。这种基于知识的模块化系统本质上是适应性的，并且在恶劣/不可预测的环境操作（例如长期空间或深海勘探任务）中具有广泛的应用。该项目将与EPSRC目标1：“发挥经济影响并确保社会繁荣”密切配合。一个模块化的自主系统，如正在研究的这个项目将提高未来的数字技术，为开发人员提供一个平台，为一个殖民地的处理器或自主代理创建高层次的指令，而不需要他们的低层次的系统功能或操作环境的详细知识。该项目还将与目标3：“使英国工程和物理科学景观能够交付”紧密结合，因为它将提供一个开放和最先进的研究软件平台和硬件演示器，这将有利于机器人和自主系统的未来合作项目。这将使计算机科学、软件工程、天文学、海洋生物学等领域的研究人员能够使用基于知识的自主方法为模块化、自配置的机器人开发软件和硬件解决方案，并在现实环境中快速测试研究假设。目前还没有公司或合作者参与该项目，但是，即将提出的研究建议将在这项工作的基础上进一步发展基于知识的模块化自主概念，并与更多的应用和研究小组接触。

This project will build upon existing modular robot system designs, focusing on the integration of autonomic behaviours into modular systems, as a part of a larger programme of research into probabilistic and knowledge-based reasoning for self-aware self-configuring autonomous systems. The aim of the project is to give a decentralised system of robotic modules being developed in the laboratory fully knowledge-based reasoning and communication capabilities. Modules will be able to communicate with each other in any possible physical configuration. Knowledge is shared as semantic base concepts and conjunctions of base concepts to enable the whole system to learn from its elements in the context of their tasks and to produce emergent and adaptive system behaviours based on semantic reasoning methods.A system of software nodes running on modular hardware elements is to be developed, and an adaptive ontology of base classes relevant to modular robotics created that will serve as the protocol for communications between modules. Data will be gathered on the performance of different networking, reasoning, knowledge sharing and goal-based task completions, and this will be used to further improve the knowledge within the system to be used for future task completion. There are no modular robotic systems implementing a formalised semantic knowledge gathering and sharing system that can be reasoned upon and very few projects worldwide implementing true knowledge-based reasoning for robotics. This PhD will lay the groundwork for highly novel knowledge-based distributed systems with applications in robotics and autonomous systems. The ultimate goals of this research programme are to construct "cellular" robotic entities based on reasoning and estimation, and to achieve a form of "common sense" by using reasoning to convert basic robotic awareness data into complex conclusions regarding the robot's own state and its environment.The creation of such a system for autonomously sharing knowledge and reasoning on it within a self-configuring robot would put the United Kingdom at the forefront of this largely unexplored area of modular robotics research. Such a knowledge-based modular system is intrinsically adaptable and has a wide variety of applications in harsh/unpredictable environment operations such as long term space or deep sea exploration missions. This project will align closely with EPSRC Objective 1: "Delivering economic impact and ensuring social prosperity". A modular autonomous system such as that being researched in this project will enhance future digital technologies by providing a platform for developers to create high-level directives for a colony of processors or autonomous agents without them requiring detailed knowledge of the low-level system functions or environment for operation. The project will also align closely with Objective 3: "Enabling the UK engineering and physical sciences landscape to deliver" in that it will deliver an open and state-of-the-art research software platform and hardware demonstrator that will benefit future collaborative projects in Robotics and Autonomous Systems. This will allow researchers in fields such as computer science, software engineering, astronomy, marine biology and others to develop software and hardware solutions for modular, self-configuring robotics using knowledge-based autonomy methods and rapidly test research hypotheses in a real-world environment.There are currently no companies or collaborators involved with this project, but upcoming research proposals will build on this work to further develop the concept of knowledge-based modular autonomy and reach out to additional applications and research groups.