Research Experience

My research work is in developing algorithms for mobile manipulator robots that can help older people and disabled people to do their essential daily tasks. Interested in transforming robots into cobots to collaborate with us on everyday activities.

Robot reaches human body parts and perform itchscrating task

• Incorporated collaborative reinforcement learning algorithms into assistive tasks to improve human-robot interaction.

• Both the Human and Robot are active agents, learn to cooperate and complete the task

Real-world Implementation

• Implemented the algorithm on a real robot and tested it on a human subject.

• Implemented human pose detection in the occluded region using trained deep-learning models.

Object Handover to active human (sim2real)

The robot learns the handover policies on pybullet simulation environment

Real world transfer of robot policies

Human behaviour generation using physics model

Human on bed using public slp dataset

Human on chair sitting in quasi-static pose

Robotic Caregivers course project

• Developed a robot that can perform a variety of tasks such as reaching, grasping, and placing objects in a cluttered environment.

• Implemented a aruco based planning algorithm on the real robot to do the tasks.

• The real time autonomous robot demo was performed on the class

Robotic Intelligent Planning course project

• Human-robot interaction is a key part of humanity’s transition to an automated society

• Developed a realistic human-robot simula- tion environment to analyze dynamic human behavior in the control loop.

• mplemented and evaluated various super- vised and unsupervised learning algorithms to complete the given task successfully (PPO, SAC, A2C, behavior cloning, MARWIL).