Navneet Paul

Curriculum Vitae

navneetpaul@student.utwente.nl | @nav74neet | Google Scholar

Masters student(MSc) in Electrical Engineering with specilization in Robotics and Mechatronics, at Universiteit Twente. Research interest relate to computer vision, fusion of machine learning techniques (deep reinforcement learning and inverse reinforcement learning) & control theory in the domain of robot manipulations, legged-robot locomotion strategies, robot teleoperation, etc.

Research Interests:

Robotics (manipulations, motion planing, etc) | Machine learning (generative adversarial & deep reinforcement learning methods) | Human-robot collaboration.

Research Experience:

  • Project Assistant, Robert Bosch Center for Cyber Physical Systems (RBCCPS), IISc. [August, 2018 - July, 2019]
  • Project Assistant, Advanced Flight Simulation (AFS) Laboratory & Computational Intelligence (CInt) Laboratory, Aerospace Engineering Division, IISc. [August, 2017 - July, 2018]
  • Research Intern, Surface Interaction & Manufacturing (SIAM) Laboratory, Mechanical Engineering Division, IISc. [May, 2016 - July, 2016]

Professional Experience:

  • Graduate Engineer Trainee, ABB Robotics and Motion Division, Bangalore. [April, 2017 - August, 2017]
  • Industrial Intern, GAIL India Limited, Uttar Pradesh [December, 2014 - January, 2015]

Education:

  • MSc Electrical Engineering (specialization: Robotics and Mechatronics), Universiteit Twente, Enschede, Netherlands
    • Duration: September, 2019 - Present

  • B.Tech Mechanical Engineering, VIT University, Vellore.
    • CGPA: 8.3/10
    • Duration: July, 2013 - May, 2017

Projects:

  • Learning manipulation tasks using generative adversarial learning methods.

    • Using inverse reinforcement learning approache like GAIL & deep reinforcement learning tehniques to leverage human demonstrations for training industrial robots to learn complex manipulation tasks such as peg in hole insertions. The initial phase is development of simulation framework using ROS and Gazebo. In simulation, the human demonstrations are carried out via haptics 3D mouse to render force feedback.

  • Deep Deterministic Policy Gradient for Bipedal walking robot.

    • A bipedal walking robot was developed in Gazebo simulation environment and Reinforcement learning algorithm, Deep Deterministic Policy Gradient based on actor-critic learning framework was implemented for generating a stable planar bipedal walking patter.

  • Q-Learning for autonomous navigation of UAVs in indoor environments.

    • Q-Learning with a simple PID tuned control was adopted for the autonomous navigation of an ArDrone in a 5x5 grid space in simulation environment.

  • Anthropomorphic Robot Arm (ARA).

    • A 3D printed humanoid arm(with 5 DoF) is attached to the ABB IRB 1600 ID industrial robot(6 DoF), actuated using servomotors controlled using an Arduino Uno microcontroller which is interfaced to the ABB IRC5 controller via a custom build step-down voltage regulator circuit board.

Publications:

  • Kumar Arun, Navneet Paul and S. N. Omkar. "Bipedal Walking Robot using Deep Deterministic Policy Gradient." arXiv:1807.05924 (2018). [arxiv][Google Scholar]

Technical Skills:

  • Programming: Python, C, Matlab, TeX
  • Frameworks: ROS, OpenAI gym
  • Libraries: Tensorflow
  • Simulation: Gazebo, Moveit, Mujoco (basics), ABB RobotStudio, Ansys Workbench
  • Microcontroller: Arduino
  • Designing and others: SolidWorks, Catia, Blender, MS-Office

Reference(s):