Varun Lakshmanan

M.Eng Robotics @ University of Maryland

About Me

Hi, I’m Varun!

I’m a robotics engineer passionate about building intelligent systems that solve real-world challenges. My journey began with a degree in mechanical engineering from Panimalar Institute of Technology, Chennai, and led me to complete my Master’s in Robotics at the University of Maryland, College Park.

I’ve developed strong expertise across autonomy, motion planning, perception, robot learning, natural language processing (NLP), and human–robot interaction. I also have hands-on experience with UAV systems and the PX4 autopilot stack, having worked on projects involving autonomous aerial navigation and offboard control.

Whether it's simulating robots in complex environments or integrating vision and control systems, I enjoy turning complex ideas into reliable, working solutions. I'm always excited to collaborate on forward-thinking robotics projects — let’s connect and create something impactful!

📄 Download My CV

Education

Master of Engineering (M.Eng) in Robotics

University of Maryland

College Park, Maryland, United States

August 2023 - May 2025

Bachelor of Engineering (B.E) in Mechanical Engineering

Panimalar Institute of Technology

Chennai, Tamil Nadu, India

August 2018 - July 2022

Projects

FireDroneX – Autonomous UAV Fire Detection

FireDroneX is an autonomous UAV system that detects and localizes fire using computer vision and monocular depth estimation. Built with ROS 2 and PX4, it integrates YOLO for real-time fire detection and uses offboard control with PID-based trajectory planning for smooth, responsive aerial navigation in simulated emergency scenarios.

Perception Based Robot Navigation

This project involves developing a perception system for a TurtleBot3 robot to navigate in a dynamic environment. The perception system uses OpenCV to detect and track objects in the environment and provides the robot with the necessary information to navigate safely.

Multi-Robot Navigation in Dynamic Environments Using Centralized and Decentralized Monte Carlo Tree Search

Monte Carlo Tree Search (MCTS) for multi-robot navigation in a dynamic environment. The simulation features four TurtleBot3 Waffle robots in a Gazebo world that includes moving obstacles, making the task more challenging.

Performance Analysis of Dueling Double DQN on TurtleBot3 for Navigation in Dynamic Environments

This project evaluates and compares the Dueling Double Deep Q-Network (Dueling Double DQN) against Vanilla Deep Q-Network (DQN) for navigation in a dynamic Gazebo simulation environment.

Adaptive Text-to-Command Translation for Robot Navigation

Autonomous Mobile Robot Navigation by integrating Natural Language Processing (NLP) for command-based navigation in a simulated environment.

A* Algorithm Implementation for a Rigid Robot

Implementation of A* Algorithm for a TurtleBot3 to navigate successfully through the custom made environment in Gazebo.

Alpha - Mobile Pick and Place Manipulator

This project involves designing a mobile pick-and-place manipulator in SolidWorks and simulating it in Gazebo using Teleop, Open-loop controllers, and Inverse Kinematics.

Gesture-Based Control in Assistive Technology

Gesture-based control for assistive technology using Google's Mediapipe Library to process hand gestures from a webcam. The recognized gestures are mapped to control DexHand in ROS2 simulation and a TurtleBot for intuitive robotic interactions.

Performance Analysis of Variants RRT Path Planning Algorithm

This Project involves comparison of performance of different variants of RRT Path Planning algorithms in 2D and implementing them to a TurtleBot3 in gazebo simulation environment by using a pid controller to check their performance on 3D simulation.

Motion Planning of PANDA Robot Using MoveIt

This project showcases a ROS 2 simulation of a Panda robot with essential configurations for a complete pick-and-place workflow. It moves from a default position, picks an object, places it in a designated location, and returns to the default pose—all within a MoveIt-based environment. This setup is ideal for anyone looking to explore motion planning, grasping, and advanced manipulation in ROS 2.