Message passing in wireless networked sensing and control systems must be reliable and in real-time. Such design challenges require new models and distributed protocols for messaging that can handle interference locally and accurately, given the basic problem of computing probabilistic path delays is NP-hard. Focusing on single-hop transmission scheduling and multi-hop spatio-temporal data flow control, this project makes novel contributions by proposing two major research tasks: 1) investigation of control-theoretic approaches to online model instantiation, based on the physical-ratio-K (PRK) interference model, and addressing the challenges of large interference range as well as anisotropic, asymmetric wireless communication; 2) developing a lightweight approach to computing probabilistic path delays followed by a multi-timescale adaptation framework for real-time messaging. In particular, the PRK interference model integrates protocol model's locality with physical model's high-fidelity, thus bridging the gap between the suitability for distributed implementation and the enabled scheduling performance. By controlling network operations at the same timescale as the corresponding dynamics, the proposed multi-timescale adaptation framework ensures long-term optimality while simultaneously addressing short-term dynamics. Additionally, this project includes an integrated, multi-level, multi-component education plan. The education activities will raise public awareness, improve student retention and participation of underrepresented groups in computing.
