Here is a list of projects in RoPAL.
Learning and Planning
Integrating learning components into planning allows us to leverage rapidly advancing machine learning tools, particularly the recent developments in generative AI. This project focuses on creating new paradigms that combine learning and planning, aiming to produce trustworthy, robust, and safe robotic AI.
Motion Planning
Motion planning algorithms guide robots from a start location to a goal location. Sampling-based motion planning algorithms work well in many scenarios, but they face challenges such as sampling in narrow passages and handling complex constraints. This project focuses on developing new motion planning algorithms and/or tools to gain deeper insights of the configuration space.
Task and Motion Planning
Task and motion planning (TAMP) integrates high-level decision-making (task planning) with low-level physical actions (motion planning). TAMP has a wide range of applications, including manufacturing, domestic use, and logistics. Previously, we combined sampling-based planning with machine learning to characterize configuration space connectivity and construct infeasibility proofs. These new tools provide stronger completeness guarantees and open new directions for solving the TAMP problem. This project focuses on applying extensions of infeasibility proofs to TAMP.
Teleoperation
While full autonomy is not quite in the picture, teleoperation or shared autonomy with robots is a great way to perform tasks in hazardous and unknown environments. Removing radioactive waste is one such example. We have a funded traineeship program accross campuses. Check for more information here
AR/VR with Robotics
Robotic systems integrate sensors, algorithms, and hardware, producing complex data that’s often challenging to interpret and use effectively. AR/VR offers a unique opportunity to bridge this gap, especially with recent advances in both AR technology and robotics. This project seeks to develop an AR/VR robotics platform for education and research. There are two main focuses: 1) real-time data visualization of robotic systems and their surroundings; 2) immersive and interactive teleoperation of robots. This platform aims to improve both robotics education and research.