This Small Business Technology Transfer (STTR) Phase I Project investigates the streaming of video through a cluster of mobile ad hoc network (MANET) autonomous robots operating in a challenging multi-path propagation environment. This project will quantify the gains in video throughput and power consumption achievable by automatic robot re-positioning with respect to these initial positions. This research is motivated by the fact that, in multi-path and fading environments, small changes in a robot's location can lead to significant gains in the received signal strength. Automatic iterative optimization of the robot positions could therefore lead to robot constellations that achieve significantly higher throughput or require significantly less power. Classical approaches for addressing multi-path degradations include robust modulation techniques, such as orthogonal frequency-division multiplexing (OFDM), and spatial diversity. This project will determine if an OFDM-based 802.11g system, can achieve significant performance improvements when combined with an automatic robot re-positioning algorithm.
Teams of robots are ideally suited for investigating environments intrinsically hostile to humans. Such environments arise in a variety of situations, including, in mines after an accident, in urban areas after a natural catastrophe, in buildings after a hazardous materials release, and in military operations. Additionally, in various video surveillance and security applications it will ultimately be cheaper to employ networks of mobile robots than to hire teams of security guards. Because humans rely disproportionately on vision for sensing their environment, the ability to stream video through MANET robot is essential. From a technology perspective, automatic cooperative re-positioning of the clusters nodes is a new dimension to explore with respect to optimizing a network's performance. The results of this project will assist in determining whether re-positioning algorithms should be incorporated in the design of future robot clusters, and if so, how such algorithms should operate, and how dynamic bandwidth allocation schemes can exploit the gains they enable.
