Ravi Prakash

Hello there!, I am an Assistant Professor at the Robert Bosch Centre for Cyber- Physical Systems in the Indian Institute of Science Bengaluru. Before this, I was a Postdoctoral Researcher in the Learning and Autonomous Control group at the department of Cognitive Robotics, TU Delft. I earned my Ph.D. in Control & Automation , from the Indian Institute of Technology Kanpur. My research has contributed to leveraging human prior knowledge and dynamical analysis to design purposeful robot learning and control strategies for intelligent robots. I am a recipient of the DAAD Postdoc networking fellowship for AI and Robotics, with funded research visits to the German Aerospace Center (DLR), Munich. My current research interests include learning complex manipulation policies from human demonstration/corrections, bimanual robot manipulation, task generalization in a novel environment, and human-friendly safe compliant control.

I have founded and direct the Human-interactive Robotics Lab at the Indian Institute of Science. Details about my research group can be found there.

 |   |   |   |   | 

Interactive Imitation Learning in Robotics: A Survey
Carlos Celemin, Rodrigo Pérez-Dattari, Eugenio Chisari, Giovanni Franzese, Leandro de Souza Rosa, Ravi Prakash, Zlatan Ajanović, Marta Ferraz, Abhinav Valada, and Jens Kober
Foundations and Trends® in Robotics: Vol. 10: No. 1-2, pp 1-197.
[NOW Publishers] [IEEE Xplore] [arxiv]

Single Network Adaptive Critic Optimal Control of an Uncertain Robot Manipulator With Robotic Applications
Ravi Prakash, Laxmidhar Behera
IEEE Transactions on Control Systems Technology
[Conditionally Accepted]

Neural Optimal Control for Constrained Visual Servoing via Learning from Demonstration
Ravi Prakash, Laxmidhar Behera
IEEE Transactions on Automation Science and Engineering
[In Press]

Model Free Adaptive Neural Controller for Standalone Photovoltaic Distributed Generation Systems With Disturbances
Meher Korukonda, Ravi Prakash, Suvendu Samanta, Laxmidhar Behera
IEEE Transactions on Sustainable Energy, vol. 13, no. 2, pp. 653-667, April 2022
[IEEE Xplore]

Dual-loop Optimal Control of a Robot Manipulator and its Application in Warehouse Automation
Ravi Prakash, Laxmidhar Behera, Santhakumar Mohan, Jagannathan Sarangapani
IEEE Transactions on Automation Science and Engineering, vol. 19, no. 1, pp. 262-279, Jan. 2022
[IEEE Xplore] [PDF]

Dynamic Trajectory Generation and a Robust Controller to Intercept a Moving Ball in a Game Setting
Ravi Prakash, Laxmidhar Behera, Santhakumar Mohan, Jagannathan Sarangapani
IEEE Transactions on Control Systems Technology, vol. 28, no. 4, pp. 1418-1432, July 2020
[IEEE Xplore] [PDF]

Closed Form HJB Solution for Path Planning of a Robot Manipulator with Warehousing Applications
Ravi Prakash, Jayanta Kumar Mohanta Laxmidhar Behera
2022 IEEE 18th International Conference on Automation Science and Engineering (CASE), Mexico City, Mexico, 2022, pp. 2049-2055
[IEEE Xplore] [DOI]

End-To-End Real-Time Visual Perception Framework for Construction Automation
Mohit Vohra, Ashish Kumar, Ravi Prakash, Laxmidhar Behera
2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC), Melbourne, Australia, 2021, pp. 3485-3490
[IEEE Xplore] [DOI]

Edge and Corner Detection in Unorganized Point Clouds for Robotic Pick and Place Applications
Mohit Vohra, Ravi Prakash, Laxmidhar Behera
2021 18th International Conference on Informatics in Control, Automation and Robotics - ICINCO, 2021, pages 245-253
[scitepress] [DOI] [arxiv]

Towards Deep Learning Assisted Autonomous UAVs for Manipulation Tasks in GPS-Denied Environments
Ashish Kumar, Mohit Vohra, Ravi Prakash, Laxmidhar Behera
2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Las Vegas, NV, USA, 2020, pp. 1613-1620
[IEEE Xplore] [DOI]

Adaptive Backstepping Neural Tracking Control of an Uncertain Robot Manipulator with Dynamic Disturbances
Ravi Prakash, Kurusetti Vinay Gupta, Laxmidhar Behera
2020 IEEE Symposium Series on Computational Intelligence (SSCI), Canberra, ACT, Australia, 2020, pp. 1936-1943
[IEEE Xplore] [DOI]

Discrete-Time Lyapunov based Kinematic Control of Robot Manipulator using Actor-Critic Framework
Ankur Kamboj, Ravi Prakash, Jayanta Kumar Mohanta Laxmidhar Behera
2020 International Joint Conference on Neural Networks (IJCNN), Glasgow, UK, 2020, pp. 1-7
[IEEE Xplore] [DOI]

Combined Online and Offline Inverse Dynamics Learning for a Robot Manipulator
Amrut Sekhar Panda, Ravi Prakash, Laxmidhar Behera Ashish Dutta
2020 International Joint Conference on Neural Networks (IJCNN), Glasgow, UK, 2020, pp. 1-7
[IEEE Xplore] [DOI]

Adaptive Critic Based Optimal Kinematic Control for a Robot Manipulator
Aiswarya Menon, Ravi Prakash, Laxmidhar Behera
2019 International Conference on Robotics and Automation (ICRA), Montreal, QC, Canada, 2019, pp. 1316-1322
[IEEE Xplore] [DOI]

Learning Optimal Parameterized Policy for High Level Strategies in a Game Setting
Ravi Prakash, Mohit Vohra, Laxmidhar Behera
2019 28th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), New Delhi, India, 2019, pp. 1-6
[IEEE Xplore] [DOI]

Real-time Grasp Pose Estimation for Novel Objects in Densely Cluttered Environment
Mohit Vohra, Ravi Prakash, Laxmidhar Behera
2019 28th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), New Delhi, India, 2019, pp. 1-6
[IEEE Xplore] [DOI]

Reward based Single Network Adaptive Critic for Optimal Kinematic Control of a Redundant Robot Manipulator
Jenin Jose, Ravi Prakash, Laxmidhar Behera
2018 IEEE Symposium Series on Computational Intelligence (SSCI), Bangalore, India, 2018, pp. 1555-1562
[IEEE Xplore] [DOI]

Learning Object Manipulation from Demonstration through Vision for the 7-DOF Barrett WAM
Niladri Das, Ravi Prakash, Laxmidhar Behera
2016 IEEE First International Conference on Control, Measurement and Instrumentation (CMI), Kolkata, India, 2016, pp. 391-396
[IEEE Xplore] [DOI]

European Research Council (ERC) Project: TERI
Project Title: Teaching Robots Interactively (TERI). PI: Jens Kober, TU Delft. Working as a Scientific Researcher in this project during my Postdoc at TU Delft. My goal is to realize a scientific contribution in enabling robots to learn and generalize how to perform manipulation tasks from few human demonstrations, based on novel interactive machine learning techniques. Developed interactive imitation learning method to generalise the robot demonstrations in novel situations using vector field transformation. Applications : Cleaning novel surfaces, Dressing clothes, Pick and Place.

Indo German Science and Technology Centre (IGSTC) Project: TransLearn
Project Title: Robot Skill transfer for Real World Deployment in Manufacturing Industries and Warehouses. Joint Collaboration between Karlsruhe Institute of Technology (PI: Torsten Kroger), KUKA Deutschland GmbH (PI: Rainer Bischoff), TATA Consultancy Services (PI: Swagat Kumar) and Indian Institute of Technology Kanpur (PI: Laxmidhar Behera). Worked as a Scientific Researcher in this project towards the partial fulfilment of my Ph.D. degree. Developed of a imitation learning framework to transfer human skills to a robot in real world. Integrated a intelligent optimal adaptive control to account for uncertainities in the robot model. Practical Demonstrations on household chores like organizing a dining table and serving drinks. Representative Publication under review in IEEE T-CST.

DST UK-India Education and Research Initiative (DST-UKIERI) Project
Project Title: Learning Robotic Motor Skills, Visual Control and Perception for Warehouse Automation. Joint Collaboration between University of Edinburgh (PI: Sethu Vijayakumar) and IIT Kanpur (PI: Laxmidhar Behera). Worked as a Scientific Researcher in this project towards the partial fulfilment of my Ph.D. degree. Visited School of Informatics, University of Edinburgh to work on this collaboration with exchange of ideas and dissemination of results. Designed a skill learning framework from human demonstrations with coordinated optimal planning and control schemes for real-world robotic applications in dynamic environments. Applications shown in Warehouse Automation with Item Sorting on a moving conveyor belt. Representative Publications in CASE 2022, IEEE T-ASE.

TATA Consultancy Services (TCS) Project: TSaL
Project Title: Teaching Skills to a Robot. PI: Laxmidhar Behera, IIT Kanpur. Worked as a Scientific Researcher in this project. Designed a skill learning framework from constrained optimal visual servo control via learning from demonstration. Applications shown in retail environment for robotic order picking. Representative Publication in IEEE T-ASE.

Department of Science and Technology (DST) FIST Project
Project Title: Tennis Playing Robot. PI: Laxmidhar Behera, IIT Kanpur. Formulated a Markov Decision Process model of Table tennis. Proposed a Pavlovian learning model for representingand inferring low-dimensional strategic state features from high dimensional sensory observations at the top level. Dynamic Motor Primitives was used to model hitting motions demonstrated by the expert. Proposed a piecewise-linear canonical system with lyapunov based adaptive gradient descent to learn the shape parameters. Robust controller for mitigating environmental uncertainities. Finally, we have developed a complete system (including a vision system for tracking the ball) using a real 4 degree-of-freedom (DOF) Barrett WAM robotic arm and show that the proposed overall framework is able to respond to an incoming ball with high accuracy. Representative Publication in Ro-Man 2019, SSCI 2020 and IEEE T-CST

IROS 2020: Open Cloud Robot Table Organization Challenge, Online
Team :ISCon_IITK, Lead: Laxmidhar Behera, IIT Kanpur. In this cloud based robotics competition the focus was on the task of table organization, which is the essential capability for service robots and requires breakthrough technology to make it mature. I was responsible for motion control of the robot.

Mohammed Bin Jayed International Robotics Challenge, Abu Dhabi
Team : IIT Kanpur, Lead: Prof. L. Behera, IIT Kanpur. The Challenge comprised of a team of UAVs and a UGV collaborating to autonomously locate, pick, transport and assemble different types of brick shaped objects to build pre-defined structures, in an outdoor environment. It is motivated by construction automation and autonomous robot based 3D printing of large structures. Developed Visually Guided UGV and AUV for Autonomous Mobile Manipulation in Dynamic and Unstructured GPS Denied Environments for Challenge 2. Specifically, Designed the motion planning and control of the robots along with ROS integration of the overall system.

Amazon Robotics Challenge 2017, Nagoya, Japan
Team : IITK-TCS, Lead: Laxmidhar Behera, IIT Kanpur. The Challenge comprised of designing and developing a fully autonomous system which can attempt simplified versions of the general task of object manipulation stowing picking in the warehouses. Each task in the challenge was comprised of known and novel items (provided 45 minutes prior to the challenge) distributed equally. In this competition, I worked as a team member responsible for robot motion control. We defined a cartesian approach based on the MoveIt! pick and place pipeline that took the target grasp candidate and computed a combination of linear segments to app- roach, contact grasp the target object, lift it after grasping and retreating with it. The TRAC-IK library is used for inverse kinematics, configured to enforce minimal configuration changes, and then collision checking is done with MoveIt! using the PointCloud information from the camera. We won 3rd Prize in Picking and 5th Prize in Combined Finals among a total of 40 International Teams

Amazon Picking Challenge 2016, Leipzig, Germany
Team : IITK-TCS, Lead: Laxmidhar Behera, IIT Kanpur. Theme of challenge remains same as of Amazon robotics challenge, 2017 with all items known a priory and no clutter. As a team member, I was responsible for motion control of mobile manipulator. We won 5th Prize in Stowing among a total of 40 International Teams.