Ajinkya Jain

I obtained my Ph.D. in robotics from the University of Texas at Austin under the guidance of Prof. Scott Niekum and Prof. Ashish Deshpande . My research aims at combining AI and Machine Learning with robot motion planning, dynamics, and controls for performing dexterous robot manipulation. During my Ph.D., I worked on to develop algorithms for learning object interaction models from images and human demonstrations and using them to perform robot manipulation tasks robustly even under uncertainty. Before joining UT Austin, I obtained my Bachelors and Masters in Mechanical Engineering from IIT Kanpur. Currently, I am working as a roboticist at Intrinsic.ai .

Email  /  GitHub  /  Google Scholar  /  LinkedIn  /  Resume  /  CV

profile photo

Research

I'm interested in robot learning, dexterous robot manipulation, motion planning under uncertainty, model learning for planning and control, reinforcement learning, and optimal control.

project image

Distributional Depth-Based Estimation of Object Articulation Models


Ajinkya Jain, Stephen Giguere, Rudolf Lioutikov, Scott Niekum
5th Conference on Robot Learning (CoRL), 2021
paper / webpage / code / poster /

A novel method that efficiently learns distributions over articulation model parameters directly from depth images without the need to know articulation model categories a priori

project image

ScrewNet: Category-Independent Articulation Model Estimation From Depth Images Using Screw Theory


Ajinkya Jain, Rudolf Lioutikov, Caleb Chuck, Scott Niekum
IEEE International Conference on Robotics and Automation (ICRA), 2021
paper / video / webpage / code / poster /

A novel method for estimating articulation models for objects directly from raw depth images without knowing their articulation type a priori using screw theory

project image

Learning Hybrid Object Kinematics for Efficient Hierarchical Planning Under Uncertainty


Ajinkya Jain, Scott Niekum
IEEE/ RSJ International Conference on Intelligent Robots and Systems (IROS), 2020
paper / video / slides /

A method for learning planning-compatible hybrid kinematics models for articulated objectsfrom human demonstrations

project image

Efficient Hierarchical Robot Motion Planning Under Uncertainty and Hybrid Dynamics


Ajinkya Jain, Scott Niekum
Conference on Robot Learning (CoRL), 2018
paper / video / code /

A POMDP-based Motion planner that generates efficient motion plans leveraging object-object interactions to perform manipulation tasks under uncertainty with high success rates

project image

Belief Space Planning under Approximate Hybrid Dynamics


Ajinkya Jain, Scott Niekum
Workshop on POMDPs in Robotics, Robotics: Science and Systems (RSS), 2017
paper /

Extending Belief-space LQR to hybrid systems for robot motion planning under uncertainty

project image

A piezoelectric model based multi-objective optimization of robot gripper design


Rituparna Datta, Ajinkya Jain, Bishakh Bhattacharya
Structural and Multidisciplinary Optimization, Springer, 2015
paper /

Multi-objective design optimization of a piezoelectric actuator driven gripper

project image

Unified Minimalistic Modelling of Piezoelectric Stack Actuators for Engineering Applications


Ajinkya Jain,
Advances in Intelligent Systems and Computing, Springer, 2014
paper /

A simple, computationally-cheap, yet effective model for piezoelectric stack actuators as a replacement of black-box models used in engineering design optimization problems.





Design and source code from Jon Barron's website. Jekyll version from Leonid Keselman's website