MRI:Acquisition of an Experimental Testbed for Research and Teaching in Mobile Collaborative Agents

这项工作支持一个关于移动、协作、基于代理的计算的大型项目，为以传感器网络为中心的各种研究活动建立试验台，并将机器人、无线传感器网络和分布式计算联系在一起。该项目由分布式机器人和计算机视觉、分布式传感器网络和分布式代理计算等较小的项目组成，使仿真能力和验证这些仿真的实验工作得以增加。机器人和视觉项目寻求实现分布式代理之间的新通信模式，实现对人和事件的情景感知识别的新技术，以及模拟多代理学习的可扩展手段。移动自组织/传感器网络项目通过使用不同传感器的目标跟踪和监视应用来检查传感器信息的数据融合和网络内处理，探索与网络中的数据注入攻击相关的安全问题(LEAP)，以及与采用不同节能技术相关的能源管理问题。此外，该研究还涉及一个关于分布式移动代理软件的项目。将受益于拟议的基础设施的研究活动包括分布式机器人项目，涉及异质代理团队分布式视觉观察者、群体代理、多代理学习、无线传感器网络的安全协议、使用移动机器人和传感器网络的分布式监控、轻量级代理体系结构、用于分布式虚拟环境的基于代理的多播、以及用于移动软件代理的动态重构和中间件。广泛影响：基础设施支持关于机器人学、传感器网络、计算机视觉和多代理技术的本科生和研究生课程，帮助培养学生对这些技术领域的学习和兴趣。结果和开发的软件将向更广泛的社区传播。14名教职员工将为加强与其他机构的合作做出贡献，以解决国土安全、医疗保健、家庭和工厂自动化、灾难恢复以及环境和基础设施监测等关键问题。

This work, supporting one large project on mobile, collaborative, agent-based computing, builds testbeds for a variety of research activities centered around sensor networks, and ties together robotics, wireless sensor networks, and distributed computing. The project, composed of smaller projects on distributed robotics and computer vision, distributed sensor networks, and distributed agent computing, enables both an increase in simulation capability and experimental work to validate these simulations. The robotics and vision project seeks to enable new modes of communication between distributed agents, new techniques for context-aware recognition of people and events, and scalable means for simulating multi-agent learning. The mobile ad-hoc/sensor network project examines data fusion and in-network processing of sensor information with an application of target tracking and surveillance using heterogeneous sensors, exploring security issues related to data injection attacks in the networks (LEAP) and energy management issues related to employing different techniques for power saving. Additionally, the research involves a project on distributed, mobile agent software. The research activities that will benefit from the proposed infrastructure include projects on Distributed robotics involving teams of heterogeneous agents Distributed visual observers, Swarm agents, Multiagent learning, Security protocols for wireless sensor networks, Distributed surveillance using mobile robots and sensor networks, Light-weight agent architectures, Agent-based multicasting for distributed virtual environments, and Dynamic reconfiguration and middleware for mobile software agents.Broader Impact: The infrastructure supports undergraduate and graduate courses on robotics, sensor networks, computer vision, and multi-agent technology, helping spawn student learning and interest in these areas of technology. The results and the software developed will be disseminated to the wider community. Fourteen faculty members will contribute to enhance collaborations with other institutions to address critical problems such as homeland security, healthcare, home and factory automation, disaster recovery, and environmental and infrastructure monitoring.