Vaibhav Nandkumar Kadam

Vaibhav Nandkumar Kadam

Autonomy Systems Intern

Danfoss Autonomy

Biography

Hello! I’m Vaibhav Kadam, graduating my Master’s in Robotics from Worcester Polytechnic Institute in May 2024. I’m pursuing my research in path planning & sensor fusion under the mentorship of Prof. Jing Xiao.

In my role as an Autonomy Systems Co-op at Danfoss Autonomy, I’m developing decision-making architectures for Off-Highway vehicles, engaged with integrating the autonomy stack on Polaris Ranger where I have extensive hands-on experience working in V-Model Automotive software development and integrating industrial sensor stack. Along with that I’m working on 3D SLAM solution for localization in GNSS denied outdoor and indoor environments.

Before this, I served as a Senior Robotics Engineer at Peppermint Robots in India from 2021 to 2022. There, I led the robotics development for autonomous industrial floor cleaning and material movement robots. I have contributed to obstacle avoidance for floor cleaning robots currently deployed for gangways for warehouses in India.

During my time at IIT Bombay (2018-2021), I built an exciting project Spherical Robot tailored to be small-size handheld robot for surveillance. Spherical Robot being wobbly in nature, led the investigations for system modelling and nonlinear control development. I hold 2 patents for small-factor design of spherical robot which is currently adapted by industry for product development. Press Release in leading newspaper in India link

I’m all set to transition back to industry starting May 2024 for full-time roles. Seeking opportunities to further enrich my expertise and make impactful contributions to the robotics industry. Please feel free to connect with me at vkadam@wpi.edu to explore potential opportunities.

Interests
  • Multi-sensor fusion
  • Motion Planning
  • SLAM
  • Reinforcement Learning
Education
  • M. S. in Robotics Engineering, 2024

    Worcester Polytechnic Institute

  • B. E. in Electronics Engineering, 2018

    Ramrao Adik Institute of Technology, Mumbai University

Experience

 
 
 
 
 
Danfoss Autonomy
Autonomy Systems Intern
May 2023 – May 2024 Minneapolis, USA
  • Developed decision making architecture for decision making in Danfoss Autonomous Control Library (ACL) ROS2 using behavior trees.
  • Migrate & Integrate ACL packages from ROS1 to ROS2 adhering V-Model automotive software development.
  • Leverage simulations & deploy on Syslogic (Nvidia Orin AGX) for off-highway autonomous vehicles.
  • Effort for systems integration and testing for ACL Packages on Polaris Ranger from vehicle interface to highway level navigation packages.
  • Currently working on 3D SLAM solution for outdoor and indoor localization in GNSS denied environments.
 
 
 
 
 
Adaptive and Intelligent Robotics Lab, WPI
Research Assistant
November 2022 – December 2023 Worcester, USA
  • Worked on DigSafe robot - Autonomous Cable Detection & Marking Robot funded by ROSE-HUB WPI & Eversource.
  • Surveying and Marking power cable laid underground is labourous task, my work was focused on developing sensor fusion algorithm.
  • I was engaged in development of multi-sensor fusion of Vivax Metrotech cable detector sensor to predict trajectory to be followed.
  • Implemented local planning based on trajectory prediction and tracking given global path generated from the Utility GIS Data. For more check project Digsafe
 
 
 
 
 
Peppermint Robots
Senior Robotics Engineer
July 2021 – July 2022 Pune, India
  • Contribute to planning & controls of industrial floor cleaning & material movement AGV autonomy stack, focused obstacle avoidance local planner using Motion primitives. Address safety edge cases and collision prediction.
  • Maintain indoor navigation stack for floor cleaning robots (SD45, SD100)
  • Developed user interface for deployment for floor cleaning robots, used boustrophedon decomposition for coverage path planning given safe and no-go zones.
 
 
 
 
 
Indian Institute of Technology Bombay
Project Research Assistant
July 2018 – June 2021 Mumbai, India

Advised by Prof. Leena Vachhani, Prof. Abhishek Gupta.

  • Worked as Project staff for development of Spherical Robot for project Search and Reconnnaissance Using Spherical Robot.
  • Designed & Developed small-form factor spherical robot prototype and experimented with Nonlinear control techniques to reduce wobble.
  • Performed various trials and colloborative feedback with security agencies for product development.
  • Effort for technology transfer to Industry for robust product development.
  • Granted 2 Patents for Novel Design of Spherical robot.

Projects

*
Digsafe Autonomous Cable Detection & Marking Robot.
Utility companies require conducting surveys to locate the distribution of power cables, a task that can be laborious. Inorder to solve this probelm, we've designed the DigSafe robot, capable of autonomously detecting cables, predicting their trajectory, and marking their locations on the ground. Our contribution involves novel algorithm of multi-sensor fusion and accurately predicting the path of underground power cables. Additionally, we've developed a tracking controller to follow the generated trajectory, validating it against global plan derived from GIS data when available, or autonomously exploring the environment to locate and mark cables as needed.
Digsafe Autonomous Cable Detection & Marking Robot.
Saliency Object Detection using Attention blocks in U2Net Architecture
We explore the application of the U2Net-attention framework to enhance this performance of Salient Object Detection (SOD). Specifically, we analyze standard and modified attention blocks in conjunction with various U2Net architectures, such as attention-inside, attention-outside, and a combination of both.
Saliency Object Detection using Attention blocks in U2Net Architecture
Multi-Agent Path Finding for Robots in Large-Scale Warehouses
We implement Multi-agent Path Finding algorithms like Space-time Astar, Conflict Based Search and variants for warehouse mobile robots to find collision free path to the target position. We provide comparative analysis of the studied algorithms, with increasing environmental complexity and increasing number of agents
Multi-Agent Path Finding for Robots in Large-Scale Warehouses
Robust Trajectory Tracking for Quadrotor UAVs using Sliding Mode Control
For a generated quintic trajectory we derive a sliding mode control law to track the reference velocity in ROS Gazebo environment.
Robust Trajectory Tracking for Quadrotor UAVs using Sliding Mode Control
Valet Parking using Hybrid Astar
Valet Parking of a differential drive, Car-Ackermann Steering and Truck Trailer we implement Hybrid Astar considering holonomic and non-holonomic constraints with trajectory rollouts to park the vehicle between two other vehicles and maneuvering through obstacles.
Valet Parking using Hybrid Astar
Search and Reconnassiance using Spherical Robot.
Spherical robot is palm-sized robot with in built Fish-eye camera based perception. We investigate Nonlinear Feedback linearization control and develop a robust and ergonomic spherical robot to control its wobble while manevuering due to its highly nonlinear nature.
Search and Reconnassiance using Spherical Robot.
Mobile Base with Spherical Robots as wheels
The project investigated the use of spherical robots as wheels. A omnidirectional mobile base was developed each spheros at 120 degrees apart connected via a link. A kinematic model for commanding the velocities of these actuation is developed. Various experiments were performed with feedback from Vicon Motion Capture systems.
Mobile Base with Spherical Robots as wheels
Ball balancing Robot
Generally differential drive mobile robots are non holonomic. The motion constraints pose hinderance in manuevering in indoor cluttered environments.The work focuses on holonomic Ball-Balancing Robot (BBR). BBR system is inherently unstable, and can be dynamically stable. BBR does not have any constraint and need to turn in cluttered places.
Ball balancing Robot

Contact