Dynamic mobile sensor networks are a distributed collection of mobile robots, each of which has sensing, computation, wireless communication and mobility capabilities. In a traditional wireless sensor network, although mobile robots are being used as sensor nodes, their dynamics and mobility are not fully exploited to improve the quality of collaborative sensing. This project explores research challenges in learning-based robotic wireless sensor networks by exploiting techniques in multi-robot research and wireless sensor networks with mobility taken into consideration. The proposed research aims to develop distributed learning algorithms and adaptive sensing coordination strategies for monitoring properties of mobile wireless sensor networks and adjusting locations of mobile sensor nodes to achieve better quality of collaborative sensing and environmental inference. The impact of this research lies in the novelty and synergism among networking, machine learning, signal processing and control to adaptively control the network structure, data routing, and signal processing in the mobile robotic sensor network to achieve better collaborative sensing. Expected results from this research will find applications in homeland security, environmental monitoring and sampling, and autonomous mobile robots for better collaborative sensing and information processing. The proposed project will serve as an excellent vehicle to recruit graduate students, especially those from underrepresented populations, to pursue their doctoral degree, and to recruit high-school students and encourage them to select engineering/computer science for their choice of career.
