Graduate Student in Robotics
Projects
Project | 01
Hybrid model-based hierarchical reinforcement learning for contact-rich manipulation tasks
(Master thesis)
In this work, contact-rich manipulation tasks are approached from the perspective of a hybrid dynamical system. A novel hierarchical reinforcement learning is developed: model-based option critic which extensively utilizes the structure of the hybrid dynamical model of the contact-rich tasks. The proposed method outperforms the state of the art method in terms of ability to adapt to uncertainty or changes in the contact-rich tasks. [Read more]
Project | 02
Factories of the Future: Collaborative mobile manipulator
(KTH Stockholm)
The aim of the project was development collaborative mobile manipulator. The system consisted of mobile robot Clearpath's Ridgeback with a Rethink's Baxter as payload. I was primarily involved in developing 3D map of the environment using multiple Intel Realsense RGBD cameras, LIDAR and Odometery. Additionally, I also ensure mobility & autonomous navigation for manipulation tasks. [Read More]
Project | 03
Sensorless full body active compliant 6-DOF parallel manipulator
(IIT Jodhpur)
This project was in collaboration with DFKI GmbH, Germany. The aim of the project was to develop a novel compliant parallel mechanism that can in used human robot interaction and haptic applications. [Read More]
Hybrid reactionless manipulation and visual servoing of a satellite mounted robot for autonomous on-orbit services
(IIT Jodhpur)
This was a minor project at IIT Jodhpur. The topic comprised of of two parts. First was dynamic parameter estimation of a space robot platform and secondly reactionless manipulation using visual servoing and model predictive control [Read More]
Sign Language Recognition and Translation using Custom Data Glove
(B.E Thesis Project)
The objective of this final year project was design a wearable glove device that takes in Indian sign language and interprets them into the desired characters and expression. The focus was not only identification of static gestures but dynamic gestures as well. [Read more]
Position/Force Control of Delta Manipulator with Voice Coil Actuator
(Academic Project, BIT Mesra)
The aim of the project was development of a direct drive 3-DOF parallel manipulator and its control. A novel passive compliance in the mechanism and a unique method of detecting external interference was also implemented. The proposed method worked satisfactorily for the purpose of safe manipulation in Human Robot Interaction (HRI). [Read More]
Team Pratyunmis (ROBOCON)
As a part of official robotics team of BIT Mesra - Team Pratyunmis, I participated in ABU Robocon 2015 & 2016, an International Robotics Competition.
ROBOCON 2016 - "Clean energy recharging the world"
As a part of electronics sub team, designed the control circuit for both Hybrid Robot and Eco Robot and developed an algorithm for autonomous manoeuvring capability of the eco robot using encoder, line scan
camera and IMU.
ROBOCON 2015 - "Robominton"
As a part of Digital Image Processing group, successfully implemented autonomous positioning of the robots by tracking the shuttlecock trajectory. The methodology involved parabolic segmentation of shuttlecock trajectory and future estimation using Kalman Filter.
Tracking and monitoring of underground pipeline using Pioneer 3dx Mobile robot
(Intern project, IIT Kanpur)
The goal of the project was to navigate an autonomous ground vehicle to detect and follow underground pipelines. The robot had to localize itself in outdoor environment and follow a presumed trajectory of the underground pipeline. Thereafter a kinematic trajectory controller was implemented to follow the robot along the desired trajectory. [Read More]
Pulmonary Cancer Cell Detection in Lung CT Images
(Intern project, IIT Kharagpur)
This was the research project on biomedical image processing during Summer 2015. The research theme included biomedical image processing particularly the diagnosis of cancer cells in lung CT(computer tomography) images.
I worked on two major challenges -
-
Developing efficient real time CAD system for pulmonary nodule detection in CT images.
-
Multi-feature analysis in classification of Pulmonary Nodules into Malignant and Benign
The general methodology applied for the to approach the above challenges was pre-processing, segmentation, feature extraction and classification. [Read more]