CPS: Synergy: Distributed Sensing, Learning and Control in Dynamic Environments

The objective of this project is to improve the performance of autonomous systems in dynamic environments by integrating perception, planning paradigms, learning, and databases. For the next generation of autonomous systems to be truly effective in terms of tangible performance improvements (e.g., long-term operations, complex and rapidly changing environments), a new level of intelligence must be attained. This project improves the state of robotic systems by enhancing their ability to coordinate activities (such as searching a disaster zone), recognize objects or people, account for uncertainty, and – most important – learn, so the system’s performance is continuously improving. To do this, the project takes an interdisciplinary approach to developing techniques in core areas and at the interface of perception, planning, learning, and databases to achieve robustness. This poster shows the recent results achieved during the last reporting period in the areas of recognition in the presence of non-overlapping cameras, sparse representation for re-identification, distributed estimation and network consistent re-identification, exploration with localization guarantees and queries over uncertain trajectories and uncertainty in stream data management.