SENSORS: Approximate Dynamic Programming for Dynamic Scheduling and Control in Sensor Networks

智能优点传感器网络在从环境监测、导航到边境监视的应用中正迅速变得重要。该项目探索了使用近似动态规划概念的传感器调度和控制的新技术，以提供有限和可变带宽问题的计算可行和最优/接近最优的解决方案。用于传感器数据选择的虚拟传感器的概念也将被用于在动态通信约束下加速传感器网络的管理。其目的是提高分布式传感器网络的运行性能，增进对如何在传感器网络中进行动态随机调度和控制的认识和理解。针对大规模传感器网络，提出了一种新颖的局部和全局动态随机调度与控制策略，并通过实验室仿真和实时实验室实现进行了验证。提出的方法进行有效的数据约简和表示将导致克服带宽限制。在此方案中使用类脑结构开发的算法将提供具有保证稳定性的最优调度。对社会的影响包括高效运行的可靠和安全的传感器网络，这些传感器网络涉及国家和全球利益，应用包括边境监视、地雷探测、无人驾驶飞行器、车辆导航、森林火灾响应、严重依赖传感器网络进行控制的关键基础设施，如电网等。在此方案中开发的传感器调度算法直接适用于许多其他众所周知的问题，如仓库中的供应链管理，其中数十个移动个人数字助理(能够传输图像的传感器，文本和语音)与中央复杂的服务器交互，为顺利交付产品和维护库存提供命令和控制解决方案。研究人员将通过将研究整合到教学中来促进工程、科学和教育方面的最佳做法。电气和计算机工程专业的少数族裔学生和女生以及目前在各大学就读的其他系的学生将被招募参加这项提案的研究活动。其他更广泛的影响包括美国和澳大利亚在这一提议上的国际合作。

Intellectual merit Sensor networks are rapidly becoming important in applications from environmental monitoring, navigation to border surveillance. This project explores new techniques using concepts of approximate dynamic programming for sensor scheduling and control to provide computationally feasible and optimal/near optimal solutions to the limited and varying bandwidth problem. The concept of virtual sensors for sensor data selection iwill also be used to accelerate management of sensor networks under dynamic communication constraints. The goal is to enhance the operational performance of distributed sensor networks and advance knowledge and understanding on how to carry out dynamic stochastic scheduling and control in sensor networks. A novel local and global dynamic stochastic scheduling and control strategy for a large scale sensor network will be designed and demonstrated with laboratory simulation and real-time laboratory implementations. Methods proposed to carry out efficient data reduction and representation will result in overcoming bandwidth constraints. The algorithms developed using brain-like structures in this proposal will provide optimal scheduling with guaranteed stability.Broader impacts The benefit to the society includes efficiently operated reliable and secure sensor networks of national and global interest for applications including border surveillance, landmine detection, unmanned aerial vehicle, vehicle navigation, forest fire response, critical infrastructures heavily dependent on network of sensors for control such as the electric power grid, etc. The sensor scheduling algorithms that are developed in this proposal are directly applicable to many other well known problems such as the supply chain management in a warehouse where several tens of mobile Personal Digital Assistants (sensors capable of transmitting images, text and voice) interacting with central sophisticated servers provide command and control solutions for smooth delivery of products and maintenance of inventory. The investigators will promote best practices in engineering, science and education by integrating research in teaching. Underrepresented minority students and female students from Electrical and Computer Engineering as well as students from other departments currently enrolled at the universities will be recruited to participate in the research activities of this proposal. Other broader impacts include international collaboration, between the U.S. and Australia on this proposal.