AUTONOMOUS MULTI-AGENT ROBOTIC SYSTEMS

机器人的使用已经从其主要的制造业基地扩展到服务业、医疗保健以及地球和行星探测。在这种情况下，多智能体协作解决在公共环境中操作的一组自主智能体(安装在移动平台上的传感器和/或操纵器)的部署、任务规划和控制，以实现通常需要有效的人-机器人交互的全局目标。正如最近由政府和行业资助的合作倡议所强调的那样，这种以人为中心的机器人技术有很大的增长空间，这将使加拿大总体人口和许多特定的关键行业受益。我们提出的关于自主多智能体智能系统和相关机器人应用基础的变革性研究是创新的，有许多预期的尖端和领先的贡献。我们预计将在两个协同主题上取得显著进展：(1)可变形对象的多智能体跟踪、传感和操纵，以及(2)多智能体移动团队的部署和运动协调。总而言之，这些构成了宏观、微型和微型机器人技术融合的重要一步。分主题1：我们这一分主题的总体目标是在我们过去工作(即关于主动视觉系统和超精密机器人)的基础上，开发离线和实时的方法，用于视觉跟踪、形状恢复和操纵先验未知的(可跟踪的)可变形物体。这一分主题的三个长期目标是：(A)具有可变形表面的动态目标的在线视觉跟踪技术；(B)通过主动视觉恢复可变形物体形状的方法；(C)机器人操纵规划的方法，包括针对检测到的物体形状可变性进行在线重新规划。两个基于短期应用的计算机视觉项目的例子是：(D)操纵微型可变形物体，(E)辅助机器人。其他的将来自我们在长期目标方面的工作。分主题2：我们这一分主题的全球目标是在我们过去工作(即关于多机器人和集群机器人系统)的基础上，开发多机器人团队部署和协调方法，以便在无结构、杂乱和无边无际的环境中拦截(无法追踪的)移动目标。这一分主题的三个长期目标是：(A)多机器人部署和协调的方法；(B)传感器网络部署的技术；(C)群体-机器人定位、编队和人机接口的方法。这四个基于短期应用的多机器人项目是：(D)荒野搜索和救援(SAR)，(E)城市SAR，(F)用于群体机器人的千禧机器人，以及(G)用于群体机器人的人工智能。其他的将来自我们在长期目标方面的工作。

The use of robots has been expanding from their principal base of manufacturing toward service industries, healthcare, and terrestrial and planetary exploration. In this context, multi-agent collaboration addresses the deployment, task planning, and control of a team of autonomous agents (sensors and/or manipulators mounted on mobile platforms) operating in a common environment to achieve a global goal, frequently, requiring effective human-robot interaction.    As highlighted by recent collaborative Government- and Industry-funded initiatives, there exists significant room for growth in such human-centred robotics that would benefit the Canadian population in general and numerous key industries in specific.    Our proposed transformative research on fundamentals of autonomous multi-agent intelligent systems and pertinent robotic applications is innovative, with numerous anticipated cutting-edge, first-in-the-field contributions.   We expect to make notable advancements on two synergetic themes: (1) multi-agent tracking, sensing, and manipulation of deformable-objects, and (2) deployment and motion coordination of multi-agent mobile teams. Collectively, these form an important step in the convergence of the macro-, milli-, and micro-robotics. Sub-Theme 1: Our global objective for this Sub-Theme is to build on our past work (i.e., on active-vision systems and ultra-precision robots) for the development of off-line and real-time methods for the visual tracking, shape-recovery, and manipulation of a priori unknown, (trackable) deformable objects.    The three long-term objectives of this Sub-Theme are to develop: (A) techniques for the on-line visual tracking of dynamic targets with deformable surfaces; (B) methods for deformable-object shape-recovery through active-vision; and, (C) approaches for robotic manipulation planning, including on-line re-planning in response to detected object-shape variability.    Two example short-term application-based computer-vision projects are: (D) manipulation of micro-sized deformable objects, and (E) assistive robotics. Others will derive from our work on the long-term objectives. Sub-Theme 2: Our global objective for this Sub-Theme is to build on our past work (i.e., on multi-robot and swarm-robot systems) for the development of multi-robot team deployment and coordination methods for the interception of (untrackable) mobile targets in unstructured, cluttered, and boundless environments.    The three long-term objectives of this Sub-Theme are to develop: (A) approaches for multi-robot deployment and coordination; (B) techniques for sensor-network deployment; and, (C) methods for swarm-robotics localization, formation, and human interfaces.    Four example short-term application-based multi-robot projects are: (D) wilderness search and rescue (SAR), (E) urban SAR, (F) millli-robots for swarm robotics, and (G) AI for swarm robotics. Others will derive from our work on the long-term objectives.