UC Santa Barbara

In recent years, there has been considerable interest in wireless sensor networks and networked robotic systems. In order to achieve the full potential of such systems, integrative approaches that design the communication, navigation and sensing aspects of the systems simultaneously are needed. However, most of the existing work in the control and robotic communities uses over-simplified disk models or path-loss-only models to characterize the communication in the network, while most of the work in networking

and communication communities does not fully explore the benefits of motion.

This dissertation thus focuses on co-optimizing these three aspects simultaneously in realistic communication environments that experience path loss, shadowing and multi-path fading. We show how to integrate the probabilistic channel prediction framework, which allows the robots to predict the channel quality at unvisited locations, into the co-optimization design. In particular, we consider four different scenarios: 1) robotic router

formation, 2) communication and motion energy co-optimization along a pre-defined trajectory, 3) communication and motion energy co-optimization with trajectory planning, and 4) clustering and path planning strategies for robotic data collection. Our theoretical, simulation and experimental results show that the proposed framework considerably outperforms the cases where the communication, motion and sensing aspects of the system are optimized separately, indicating the necessity of co-optimization. They further

show the significant benefits of using realistic channel models, as compared to the case of using over-simplified disk models.