EMT/BSSE Programmable Self-Adaptation: A Bio-inspired Approach To Multi-agent Robotic Systems

EMT/BSSE 可编程自适应：多智能体机器人系统的仿生方法

• 批准号: 0829745
• 负责人: Radhika Nagpal
• 金额: $ 30万
• 依托单位: Harvard University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2008
• 资助国家: 美国
• 起止时间: 2008-09-01 至 2012-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0829745&HistoricalAwards=false
• 关键词: EMT; BSSE; Programmable; Self; Adaptation

Emerging computing technologies have made it possible to manufacturelarge-scale embedded multi-agent systems, from vast sensor networks tomodular robots and smart materials. A key challenge is understandinghow to program such systems at the individual agent level in order toachieve system-level behavior that is complex, fault-tolerant, andadapts to the environment. One source of inspiration is multicellularbiological systems (tissues, organs, and simple organisms) thatachieve complex self-adaptation in changing environments through thedistributed cooperation and sensing of vast numbers of cells. Suchsystems can provide novel bio-inspired principles for the design andprogramming of multi-agent systems.This research investigates new computational paradigms for programmingmulti-agent robotic systems to achieve complex self-adaptation inresponse to the environment. The two main thrusts are: (a) thedevelopment of a global-to-local programming methodology fordescribing complex global adaptation goals and automatically derivingprovably robust multi-agent control (b) the development of atissue-inspired modular robotic system that can demonstrateself-adaptive structures. A key source of inspiration is thedecentralized control strategies that cells use to achieveenvironment-responsive structures and functions, e.g. shape adaptationin plants and vascular networks, and locomotion in simple animals. Theaim is to harness these bio-inspired principles to create multi-agentrobotic systems that replicate the adaptability and responsiveness ofliving systems. This work has broad application, from robust controlin distributed sensor-actuator networks to the development ofself-adapting architectural structures and prosthetics. This researchsignificantly advances our understanding of how to design autonomousmulti-agent computing systems that can self-organize, self-repair, andrespond to the environment.

新兴的计算技术已经使大规模嵌入式多智能体系统的建模成为可能，从庞大的传感器网络到模块化机器人和智能材料。一个关键的挑战是理解如何在个体代理级别上对这样的系统进行编程，以实现复杂的、容错的和适应环境的系统级行为。灵感的来源之一是多细胞生物系统（组织、器官和简单的有机体），它们通过大量细胞的分布式合作和感知，在不断变化的环境中实现复杂的自我适应。这类系统可以为多智能体系统的设计和编程提供新的生物启发原则。本研究探讨了新的计算范式，用于编程多智能体机器人系统，以实现对环境的复杂自适应。这两个主要目标是：（a）发展一种全局到局部的编程方法，用于描述复杂的全局适应目标，并自动导出可证明鲁棒的多智能体控制（B）发展一种满足感的模块化机器人系统，该系统可以展示自适应结构。灵感的一个关键来源是分散控制策略，细胞使用这些策略来模拟环境响应的结构和功能，例如植物和血管网络中的形状适应，以及简单动物的运动。他们的目标是利用这些生物启发的原则，创造多代理机器人系统，复制生命系统的适应性和反应能力。这项工作具有广泛的应用，从分布式传感器-执行器网络的鲁棒控制到自适应建筑结构和假肢的发展。这项研究极大地推进了我们对如何设计能够自组织、自修复和对环境做出响应的复杂多代理计算系统的理解。