Jay Farrell
Current research directions, philosophy, research activities:
Prof. Farrell’s research is focused on developing advanced navigation, control, and planning methods to enable new capabilities for autonomous vehicles. His preference is for every article and student thesis to contain both rigorous theoretical analysis and experimental demonstration of performance. He has research interests in the following directions:
- Aided inertial navigation for highway applications - Application include lane-level positioning to support precision lane mapping, vehicle assist and automation applications, transit vehicle automation, and highway maintenance vehicle guidance. Aiding from diverse sources is of interest: GNSS, vision, LIDAR, RADAR.
- Self-Organizing Approximation Based Control - Precision tracking control for nonlinear systems with significant modeling error is a challenging task. Our approach uses on-line function approximation to decrease the effects of the model error. Due to the fact that the nonlinearities are unknown, self-organization of the approximator basis elements has both performance and computational benefits. Our applications have included land, air, and underwater vehicles as well as critical care ventilators and robotic systems.
- Online Planning and System Performance Optimization - Currently, our research in this area is studying distributed self-configuration of camera networks for security applications. The objective is for a set of cameras communicating over a wireless network to configure their camera pan, tilt, and zoom settings to both track the locations of people and acquire high-resolution imagery at times of convenience. Our approach currently uses methods from game theory and consensus. Previous topics under this line of research include tracing of chemical plumes to their sources and declaration of the source location.
In addition to sponsored research in these areas, Prof. Farrell advises various undergraduate and graduate student projects in related areas.