PhD Candidate
University of California, Berkeley
Areas of Interest
- Artificial Intelligence
- Control, Intelligent Systems, and Robotics
- Cyber-Physical Systems and Design Automation
- Human-Computer Interaction
Poster
Predictive and Collaborative Multi-Robot Coordination with Model Predictive Control and Duality Theory
Abstract
Robotic systems can hugely benefit from communication technologies. Advances in vehicular communication technologies such as Vehicle-to-Vehicle (V2V), Vehicle-to-Infrastructure (V2I) and Vehicle-to-Cloud (V2C) connectivity facilitate cooperative driving. Connected and Automated Vehicles (CAVs) are able to collaboratively plan and execute driving maneuvers by sharing their perceptual knowledge and future plans. By using the theory of model predictive control (MPC) and strong duality, I’ve developed provably safe algorithms for collaborative navigation of multiple heterogeneous autonomous mobile robots with different (arbitrary) polytopic shapes and different dynamical models in tight environments. The coordination strategies are classified in two categories of centralized and distributed. I’ve developed optimization-based centralized coordination strategies for formation, reconfiguration and autonomous navigation of CAVs, travelling on public roads. Using the proposed approach, CAVs are able to form single or multi-lane platoons of various geometrical configurations. They are able to reshape and adjust their configurations according to changes in the environment. In addition, I’ve designed the distributed coordination algorithm that exploits the problem structure to decompose the large optimization problem into smaller local subproblems solved in parallel. Using this approach, the robots can cooperate (while communicating their intentions to the neighbors) and compute collision-free paths in a distributed way to navigate in tight environments in real time.
Bio
Roya Firoozi is a Ph.D. candidate in Mechanical Engineering at University of California, Berkeley, advised by Professor Francesco Borrelli. Her research lies at the intersection of optimization, predictive control and machine learning. Her work on predictive and collaborative robotics incorporates communication and connectivity technologies in autonomous systems to enable efficient and safe planning, control, estimation and fault diagnostic algorithms for single and multiple agents in complex environments. She received her B.S. in Mechanical Engineering from University of California, Berkeley in 2014. She has received UC Berkeley Chancellor’s Award and Graduate Remote Instruction Innovation Fellowship Award. Outside of research, she is active in outreach and mentoring programs, acting as STEM*FYI ambassador and GWE mentor at UC Berkeley.
