- This event has passed.
GRASP Special Seminar: Pedro Roque, KTH Royal Institute, “Predictive and Vision-based Control for Coordinating Multi-Agent Systems”
June 1, 2022 @ 11:00 am - 12:00 pm
*This is a HYBRID Event with in-person attendance in Levine 512 and virtual attendance via Zoom.
ABSTRACT
As aerial and space systems become ubiquitous, the challenges associated with coordinating large constellations
of autonomous agents become even more important, as these systems are both safety-critical and expensive. In this talk, we
will go through a few applications of Model Predictive Control (MPC) for load transportation and safe trajectory tracking of
time-varying trajectories. Then, we will expand on how MPC can be used with vision feedback from generalized onboard cameras
to achieve an efficient distributed formation control scheme. In particular, we will explore the geometric properties of these
imaging sensors to develop a framework (and Python library) for Visual Coordination of multi-agent systems. Results depicting
the usage of this framework in both simulation and experimental scenarios shed light on the main advantages of the proposed
method, as well as possible research directions to improve its performance.
Presenter
Pedro Roque
Pedro Roque is a Ph.D. Candidate with the Division of Decision and Control Systems at KTH Royal Institute of Technology, Stockholm, Sweden, supervised by Prof. Dimos Dimarogonas, Prof. Mikael Johansson and Prof. Jana Tumová. His research interests lie at the intersection of Model Predictive Control for multi-agent systems, Robotics in weightless environments, and Computer Vision. He obtained a Double Degree M.Sc. in Systems, Control and Robotics from IST Lisbon and KTH Royal Institute of Technology. He collaborated with the MIT Space Systems Lab to develop distributed formation control schemes for the MIT SPHERES and NASA Astrobee free-flyer platforms, including a total of 5 test sessions in the International Space Station. He is a contributor to the NASA Astrobee platform, and is part of the Development team of the PX4 Autopilot firmware for unmanned aerial vehicles.