Executive Summary : | The development of effective solutions for time-sensitive global challenges requires sensing of critical environmental processes (EP), such as detecting toxic gases, forest fires, and monitoring water, air, and soil quality. A network of multiple sensing agents can be used to accurately detect these processes, but a sufficient agent density and sensing interval are required to cover the entire space and time duration. The variation in EPs over space and time may lead to large redundancy in the sensed data. Incorporating knowledge of the environmental variable's spatial and temporal variation can reduce the required density and sensing frequency, avoiding data redundancy and decreasing operating costs. The optimal density and sensing frequency of agents required to cover a region, given the spatio-temporal profile, is crucial. Estimation accuracy depends on previous sensing events and the optimal trajectory for the agents. This proposal aims to study a network consisting of moving sensing agents to sense an EP with a known spatial and temporal correlation profile, utilizing this information to estimate the value of the environmental variable for all locations and time instants. The first objective is to develop a tractable mathematical framework that incorporates key features of the system, including the spatio-temporal profile of an environmental process, agent deployment, trajectories, and temporal sensing process. The study will also study the mean estimation error and the impact of various trajectories on sensing performance. |