Accelerated and Memory Efficient Edge Reinforcement Learning
Accelerated and Memory Efficient Edge Reinforcement Learning
Project Overview
Machine learning-based robotics applications are increasing in prevalence and importance. Whether structured as foundation-models, pixels-to-actions policies, or as learned hyperparameters or models for classical approaches, these algorithms will need to run on the edge to support the next generation of field robots. We are therefore working to reduce the challenging memory requirements and long runtimes of such algorithms, enabling their use at the edge.
Publications
-
(2026). MPC-Injection: Biasing Off-Policy Locomotion RL Toward Controller-Induced Behavior Basins. In CoRL 2026.
-
(2026). Solvers at the Edge: Embedded Optimization for Safe and Intelligent Robotics. Columbia University Masters of Science in Computer Science Thesis.
-
(2026). Hardware-Algorithm Co-Design for Real-Time Linear Model Predictive Control: FPGA Implementation and Deployment on a Resource-Constrained Quadrotor. École Polytechnique Fédérale de Lausanne Masters Thesis.
-
(2024). Differentially Encoded Observation Spaces for Perceptive Reinforcement Learning. In ICRA 2024.
-
(2022). Closing the Sim-to-Real Gap for Ultra-Low-Cost, Resource-Constrained, Quadruped Robot Platforms. In Workshop on Closing the Sim2Real Gap at RSS 2022.
-
(2020). Reinforcement Learning to Enable Robust Robotic Model Predictive Control. Harvard University Bachelors of Arts Thesis.
-
(2019). Application of Approximate Matrix Multiplication to Neural Networks and Distributed SLAM. In HPEC 2019.
