Sambaran Ghosal

Sambaran ghosal

bookmark on deepenrich
location of Sambaran GhosalSan Diego, California, United States
Followers of Sambaran Ghosal763 followers
  • Timeline

  • About me

    Autonomy Engineer I @BrainCorp, MS ECE (Intelligent Systems, Robotics and Controls)

  • Education

    • Indian institute of technology, kharagpur

      2015 - 2020
      Integrated btech and mtech mechanical engineering 8.82/10

      Coursework: -Robots and Computer Controlled Machines -Mechanisms and Robot Kinematics-Dynamics of Machines -Systems and Controls -Automatic Control -Neuro Fuzzy Control -Robust Control -Soft Computing-Finite Element Methods-Solid Mechanics-Applied Elasticity-Kinematics of Machines -Partial Differential Equations -Laplace Transform

    • University of california, san diego - jacobs school of engineering

      2021 - 2023
      Master's degree electrical and computer engineering 4.0/4.0

      Coursework :-Statistical Learning I(ECE271A)-Programming for Data Analysis(ECE143)-Sensing and Estimation in Robotics(ECE276A)-Linear Algebra(ECE269)-Planning and Learning in Robotics(ECE276B)-Machine Learning for 3D geometry(CSE291)-Intro to Visual Learning(ECE285)

  • Experience

    • Indian institute of technology, kharagpur

      Dec 2016 - Jun 2020

      -Worked extensively in the Centre for Robotics, IIT Kharagpur under the guidance of Professor D.K Pratihar. -Worked on the development of a Hybrid Mobile Manipulator capable of uninterruped manipulation. -Worked on the mechanical design and modelling of the proposed Hybrid Mobile Manipulator in Solidworks.-Worked in the simulation of the proposed model in ADAMS Tracked Vehicle Toolkit and obtained the motor requirements for carrying out desired movement and manipulation tasks. -Simultaneously worked on the design of a Six Legged Hexapod Robot for Research and Education purposes. -Developed the 3D model of the robot in Solidworks and obtaining the URDF model of the robot for implementation in ROS-Gazebo framework. Show less

      • Student Researcher

        Aug 2019 - Jun 2020
      • Teaching Assistant

        Sept 2019 - May 2020
      • Student Researcher

        Dec 2016 - Jul 2019
    • Institute for systems and robotics (isr-lisboa)

      May 2019 - Aug 2019
      Summer research intern

      -Worked as a research intern under the guidance of Professor Rodrigo Ventura in the field of Robotics and Controls.-Developed a Model Predictive Controller(MPC) in ACADO-C++ interface. -Implemented the MPC controller on NASA Astrobee robot using ROS-Gazebo framework. -Obtained comparison in performance of the MPC controller implemented and the default PID controller.

    • Uc san diego

      Apr 2021 - Apr 2023

      -Proposed an image based reinforcement learning method to achieve manipulation tasks. - Constructed simulations for manipulation tasks like reach, push, peg in a hole, open and close drawer, pegging a nail with a hammer using MUJOCO simulator and connect to Reinforcement Learning interface. - Showed that multiple camera paradigm is able to solve the tasks better and faster than single camera paradigm. - Implemented the learned policies in a real UFACTORY XARM7 manipulator and compared the success rate of using multiple cameras and single camera for each of the above mentioned tasks Show less

      • Teaching Assistant

        Sept 2021 - Apr 2023
      • Student

        Mar 2021 - Mar 2023
      • Image Based Reinforcement Learning to achieve Robotic Manipulation

        Apr 2021 - Mar 2022
    • Existential robotics lab, ucsd

      Jun 2022 - Sept 2023
      Associate research fellow

      -Proposed and implemented a Hamiltonian based Neural ODE architecture to learn the dynamics model of Jackal ground vehicle using point cloud observations.-Modelled Jackal as a differential drive robot in SAPIEN(https://sapien.ucsd.edu/) simulator and created a sample environment to collect point cloud observations from the simulator.-Trained a Hamiltonian based neural ode architecture to model the dynamics of the simulated jackal and developed an energy based controller with desired Hamiltonian function shaping to achieve pose stabilization as well as tracking of the robot. -Demonstrated the advantages of the learned model controller over a nominal controller for pose stabilization as well as tracking a circular and lemniscate trajectory.-Arxiv : https://arxiv.org/pdf/2309.09163.pdf Show less

    • Brain corp

      May 2023 - now

      Member of Perception Team in the Autonomy R&D Department at Brain Corp R&D Robotics Software Engineer (VSLAM, Camera Calibration)

      • Autonomy Engineer I

        Nov 2023 - now
      • Research Software Engineer Intern

        May 2023 - Nov 2023
  • Licenses & Certifications

  • Honors & Awards

    • Awarded to Sambaran Ghosal
      .1% Merit Certificate for highest marks in Chemistry and Physics in the Higher Secondary School Examination, 2015 Central Board of Secondary Examination Jul 2015
    • Awarded to Sambaran Ghosal
      Kishore Vaigyanik Protsahan Yojana (KVPY) qualifier(2014) - Mar 2014