EmPRISE Robotics Laboratory

Contact-Rich Dynamic Planning for Safe Manipulation and Control

Working on whole-arm contact-rich planning for robotic arms, focusing on ways to combine reaction-based control (such as Model Predictive Control) with offline contact-focused planning for a robotic arm equipped with external haptic force sensors. I am specifically focusing on creating trajectories that ensure the robot arm has enough flexibility of movement at contact points for the controller to move it away upon obstacle movement or collision with minimal disruption to the overall goal. This ability to react is crucial to ensure safety around dynamic and unpredictable obstacles such as humans. Currently writing paper as first author.

My current approach relies on maximizing manipulability around contact points using RRT* and signed distance functions (allowing for some taxel penetration). I am simulating using PyBullet.

This experience has been absolutely crucial with regards to understanding and experiencing the full cycle of research PhD students and professional researchers go through from first author perspective. I worked on my own research project with complete independence other than weekly advisor meetings, performing literature review, narrowing down key exploration topics and strategies, building code infrastructure, deriving and testing various original algorithms, testing my methods, adjusting my focus, and writing reports on my findings.

Contact-Aware Multimodal MPC for Whole-Arm Dexterous Manipulation

I led hardware experiments for a paper under review, for which I was in charge of training a latent dynamics model on robot manipulation tasks using various sensor modalities, focusing on robotic assistance and pHRI scenarios. Such modalities include whole-arm tactile sensors, an RGB-D camera, and joint states and velocities over time. By capturing and predicting complex contact-rich interactions, the latent model improves the performance of model predictive control for manipulation tasks with whole-arm contact. My task included determining which scenarios would be best for demonstrating the method, setting up and running experiments, and writing the experimentation and results section of the paper.