Photo of Sudhir Pratap Yadav

Sudhir Pratap Yadav

Phone: (+91)-637-698-4160
Email:

I am on the path of discovering who am I by creating robots. I hypothesize, we (Humans) are what can't be automated. Therefore, my efforts are towards creating intelligent autonomous robots so we can be free of mundane and focus on that which is truly important (and fulfilling).

Education

Indian Institute of Technology (IIT) Jodhpur May 2014 - May 2018

B.Tech. in Electrical Engineering

Academic Distinction Award (2016-17)


Indian Institute of Technology (IIT) Jodhpur July 2023 - Ongoing

Ph.D. - Robotics and Mobility System

CGPA 10/10

Technical Skills

Software (Languages, Frameworks and Tools)

Python, C++, MATLAB, JAVA, Pytorch, ROS, Isaac Sim, Pybullet, OpenCV, Tensorflow, Gazebo, GPU Server (A100)


Hardware (Robots)

UR5, Kinova, Turtle Bot arm, TAL Brabo Industrial Robot, Husky, Open Manipulator, Quadcopter, Humanoid robot, Quadraped robot, P3DX robot, several hand-made robots


Hardware (Embedded)

Jetson Nano, Intel realsense camera, Kinect camera, Motion Capture System, IMU, GPS, Lidar, Dynamixel Motors, Arduino, RaspberryPi, UP square board

Experience

Robotics Intern September 2024 - March 2025

Black Coffee Robotics

  • Working on fine-tuning large AI models for robotic tasks. Designed and implemented custom MoveIt! Pro behaviours and AR projection system for clients. Developed farm robot simulation in Isaac Sim.

Junior Research Fellow January 2023 - July 2023

IIT Jodhpur

  • Developed unity simulation to collect visual and LIDAR data for desert environment. Developed custom 3D-LIDAR sensor in unity.

Research Engineer September 2021 - January 2023

Vision based Autonomous Systems Lab - TIH iHub Drishti

  • Developed technology for vision-based navigation in GPS-denied environments, demonstrated on Quadcopter. Developed algorithms and frameworks for learning multiple vision-based manipulation tasks on robotic arms using Reinforcement Learning.

Robotics Software Engineering Intern May 2017 - July 2017

TCS Innovation Lab

  • Developed ROS interface for 5-DOF TAL BRABO Industrial robot arm. Created a pipeline for making the arm follow a green-colored ball to demonstrate proper working of the interface.
  • Documentation Code

Robotics Research and Development Intern May 2016 - June 2016

Robotics Research Centre - IIIT Hyderabad

  • Developed kinematic model of 26-DOF humanoid. Implemented static balance by calculating ZMP,COM and using force sensors. Implemented passive perturbation rejection using compliance of motors.
Publications

Learning vision-based robotic manipulation tasks sequentially in offline reinforcement learning settings 2024

Robotica: 1-16

Yadav SP, Nagar R, Shah SV.

Pre-print Code arXiv
Projects

3D Human Skeleton Tracking using RGBD August 2024

Side Project

  • Developed a real-time human skeleton tracking system using a 2D deep learning-based detector and depth data to triangulate 3D coordinates using an Intel RealSense D455 camera. Developed visualization software for real-time rendering.

Multi-agent ecosystem Simulation in Unity April 2024

Side Project

  • Implemented a virtual simulation of a natural ecosystem highlighting the interactions between various animals and their environment. Created a realistic and dynamic simulation that demonstrates the behaviours of different organisms in response to environmental changes and interactions with each other. Also, demonstrate planning and other capabilities to navigate the environment.

Visual Servoing on Turtlebot December 2023

Side Project

  • Developed visual servoing algorithm for turtlebot (mobile robot). A camera is attached to the robot, and a visual target is defined. The robot will try to track and follow the target.

3D Scene Reconstruction of a Desert Environment November 2023

Side Project

  • Compared various NERF approaches to reconstruct 3D desert environment. Simulated as well as real dataset was used to compare various techniques.

VR Tours December 2022

Side Project

  • Created a virtual reality tour of the Robotics Lab and a restaurant.
  • VR Tour Code

Pick-Place Robot August 2022

Side Project

  • Developed a vision-based robot system to pick and place objects. Intel RGBD camera was used to detect the object location and then motion is planned accordingly for the robot to perform the pick-place operation.

Neural Radiance Field - NERF July 2022

Side Project

  • Created NERFs for various locations using images. NERFs are 3D models created using Neural Networks. Images are collected and then a neural network is trained to create 3D models.

National Park Quiz - WebUI July 2021

Side Project

  • Developed a WebUI for learning about the national parks of India.

Vision based control of UR5 robotic arm using Deep RL May 2017 - May 2018

Robotics Lab IIT Jodhpur — Dr. Suril V. Shah

  • Developed and Trained RL agent for moving UR5 robotic arm. Aim was to train arm for Visual servoing a specific object of certain colour. ROS was used to control UR5 robot while RL agent was trained using TORCH(LUA) on simulator GAZEEBO.
  • Documentation Code

Quadruped Navigation using RL Dec 2016 - May 2017

Robotics Lab IIT Jodhpur — Dr. Suril V. Shah

  • Main focus of this project was to make quadruped learn to walk on terrain without any previous knowledge. The Kinematic model of quadruped and RL agent were developed in C++.
  • Documentation Code

Playing 2048 game using RL July 2016 - Dec 2016

Robotics Lab IIT Jodhpur — Dr. Suril V. Shah

  • Developed copy of 2048 game in C++. Used Reinforcement Learning to learn and play the game.
  • Code

Simulating Autonomous Mini-Helicopter July 2016 - Dec 2016

Robotics Lab IIT Jodhpur — Dr. Suril V. Shah

  • Design and development of Autonomous Helicopter to carry 2.6 kg Nuclear Radiation Sensor in Emergency Scenario in desert area. This problem was given by Defence Lab Jodhpur. We developed a very accurate simulation of helicopter in matlab. PID control was applied on this simulation for simple control.

Quadcopter trajectory tracking June 2016 - July 2016

Robotics Lab IIT Jodhpur — Dr. Suril V. Shah

  • Implemented PID control for trajectory tracking using VICON motion capture system and ROS.
  • Video

Control of P3DX and Mapping June 2016 - Dec 2016

Robotics Lab IIT Jodhpur — Dr. Suril V. Shah

  • Controlled P3DX using ROS with PID controller. Used Kinect for depth sensing and implemented Occupancy Grid mapping in MATLAB.
  • Documentation Code Video

Gait Planning Of Quadruped Jan 2016 - May 2016

Robotics Lab IIT Jodhpur — Dr. Suril V. Shah

  • Developed a kinematic simulation of quadruped in MATLAB with crawl gait. Crawl gait was also implemented on real quadruped robot.
  • Code

Unblock-me puzzle 2015

Robotics Lab IIT Jodhpur

  • Image Processing based game solver. Developed a GUI in JAVA, to automatically solve a game named UNBLOCK ME. User has to show the mobile to laptop camera, software will detect game state and solve it.
  • Code
Side Fun Projects

IITJ Nerfs 2022

  • Developed a kinematic simulation of quadruped in MATLAB for walking using crawl gait. Crawl gait was also implemented on real quadruped robot.
  • Coming Soon

Playing 2048 game using R 2017

  • Playing 2048 game using RL (2017)) Game created in C++ and python and solved using TD-learning
  • Code

Unblock-me puzzle 2014

  • Vision based detection of game state and automatically solving the puzzle.
  • Code
Contributing back to Community

Teaching Assistant 2023

  • Teaching Assistant in the Experimental Robotics Course, teaching students how to apply various robotics concepts on real hardware.

ROS Workshop 2022

  • Conducted a small workshop on Introduction to ROS (Robot Operating System)
  • Content

Mentor on Coursera 2017

  • Given opportunity to be mentor of the course on Robot Estimation and Learning
  • Certificate

Head (2015), Mentor (2016) Robotics Club 2022

  • Mentored students and took few lecture series so to strengthen robotics culture in IITJ

Volunteer Parivartan Group 2014-2017

  • Group was focused on social service. My group was given opportunity to teach school students who could not afford tution.