Near Earth Sensing with Multi-Robot Systems

Also on zoom-https://umd.zoom.us/j/97114322433?pwd=TWw0OG8yV3ZTc1d2V0RlYXB6RkNWQT09 Consider domains such as agronomy and environmental science in which access to high-resolution spat-temporal data is critical in understanding the physical world. One of the main bottlenecks in such settings is collecting the data in the first place. In this talk, I will present an overview of my group's research on planning and coordination algorithms for data collection with heterogeneous robot teams. I will focus on three types of algorithmic problems that are grounded in environmental applications but address fundamental challenges in realizing multi-robot systems. I will start with the question of where should the robots gather data from to learn a spatiotemporal field. Then, I will discuss how heterogeneity in a multi-robot system can be exploited to enable long-term autonomy. Finally, I will discuss our recent efforts aimed at bridging the gap between theory and practice by studying issues of resilience and risk-aware planning in the presence of uncertainty and adversarial attacks/failures.