Our group has two broad research goals:
(1) To develop a deeper understanding of the neural control and biomechanics in the human body using robotics techniques.
(2) To develop the design and control methodologies (including human-inspired) that enable robots to operate robustly in unstructured environments
We have on-going projects in the below areas. Please see the publications page for papers on these projects.
The goal of this project is to understand and improve upon current algorithms for robotic grasping and manipulation by identifying the strategies that humans use for physical interaction. We use a variety of human-robot interaction paradigms to collect the data, which includes a variety of information such as robot hand posture, contact points between the robot and the object, and the volume of the object enclosed.
This data will be analyzed using a variety of machine learning techniques (joint work with Prof. Weng-Keen Wong) to improve a robot's grasp prediction capability.
When a robot's conventional operation mode is ineffective, the robot's operation fails. For example, when car's wheel slips in snow, its mobility is compromised. Similarly, when contact fails between a manipulator and an object, the robot loses grip. Such events are called locomotion or manipulation errors.
We use a combination of dynamics and kinematics analysis and AI techniques like dynamic programming and reinforcement learning to enable the robot to discover new movement capabilities. This project builds on my doctoral work at CMU that explored the structure of locomotion errors and devised new ways to approach the mobile-robot recovery problem. Joint work with Prof. Matthew Taylor.
Tendon-transfer surgeries are commonly performed for a variety of conditions such as stroke, palsies, trauma, and congenital defects. It involves re-routing a tendon from a dysfunctional muscle to a still-functioning musle to restore lost function. However, a fundamental aspect of the current surgical procedure, namely the suture that attaches the tendon(s) to the muscles, results in poor hand function in physical interaction tasks such as grasping, because of the fixed 1:1 coupling the suture produces between the muscles and tendons.
Our group is investigating the use of passive engineering mechanisms, such as hierarchical pulleys, to attach the muscles and tendons. It is expected that this will significantly improve post-surgery grasping capability. Joint work with Prof. Chris Allan, MD.