Robust, Adaptive, and Uncertainty-Aware Autonomy

Autonomous systems are moving rapidly from laboratories into safety-critical domains such as self-driving cars, surgical robotics, energy systems, and other cyber–physical platforms. However, real-world deployment remains hindered by the uncertainty arising due to, e.g., environmental variability, unmodeled disturbances, and adversarial interference. In this talk, I will describe a principled approach that integrates nonlinear control, hybrid dynamical systems, and optimal transport to address two core questions: (1) how to design model-free online-learning and optimization loops with global stability and robustness guarantees, and (2) how to steer uncertainty itself to meet terminal specifications on demand. I will close by outlining a vision for robust uncertainty-aware autonomy, with applications to safe adaptive robotic navigation in cluttered environments, and online learning-enabled autonomous systems under constraints.

Mahmoud Abdelgalil is a Postdoctoral Scholar in the electrical and computer engineering department at the University of California San Diego. He graduated in 2018 with a bachelor's degree in aeronautical and astronautical engineering from the University of Science and Technology at Zewail City in Egypt. He received his MSc and PhD in mechanical and aerospace engineering from the University of California Irvine in 2020 and 2023, respectively. Dr. Abdelgalil was the recipient of the Holmes fellowship and the Henry Samueli Fellowship at the University of California, Irvine, and the awardee of the 2023 ACC Best Student Paper.