top of page
  • GitHub
  • LinkedIn

Projects

Here is a concise overview of key projects where I have made significant contributions across various domains of robotics.

My skills include:
Software Tools: ROS2, ROS1, Issac ROS, PyBullet, Gym, Mujoco, COLMAP, GStreamer, Pytorch, Tensorflow, Ardupilot, QGC, Gazebo, KiCad, Labview, Docker, Keras, CasADi, CVXPY, Numpy, SciPy, SciKit, OpenCV, Git
Programming languages: Python, C/C++, MATLAB

Safe Navigation in Perception-Limited Environments, 2024

rrtpic.jpg

Contributed in the development of rovers using PX4 and ROS2 ecosystem that used a RealSense camera to implement ESDF generation using NVBlox on IsaacRos and could implement the visibility RRT* algorithm.
Github

Adaptive MPC-CBF, 2024-25

mpccbfpic.jpg

Implemented safe control on a dynamic unicycle robot with class-K function and weights being tunable according to obstacle distance. Extending the work to include dynamic obstacles using a Maximum Entropy Inverse Reinforcement Learning Model.
Github

GraspSplats, 2024

Screenshot (5)_edited.jpg

Implemented the paper in PyBullet Environment-GraspSplats that uses feature embedded Gaussian Splats with MobileSAMv2 and MaskCLIP to generate grasp poses with sampling based Grasp Pose Detection.
Report

Perception for self driving cars, 2023

pointnetdemo.jpg

Implemented PointNet for point cloud classification and segmentation, and also recreated similar results using Monocon architecture on a custom RGB image dataset.
Code

ADAS for self driving cars using modern controls, 2023

Screenshot from 2025-02-03 17-40-38.png

Used CVXPY and CasADi to implement LQR-QP based ACC and Trajectory Tracking systems on point mass and bicycle models of a car.

Botlab, 2023

Screenshot 2025-02-02 105203.png

Developed an autonomous mobile robot that maps the environment using Breshenham's Algorithm and performd SLAM. Used A* and frontier exploration to navigate and perform pick and place tasks in a grid.
Report

Armlab, 2023

armlabdemo.jpg

Developed autonomy for a 5-DOF robotic arm by implementing computer vision techniques, kinematics, and path planning, enabling object manipulation.
Report

High Speed Autonomous Logistics Robot (HAL), 2020-23

haldemo.jpg

A full stack architecture was developed for logistics robots including autonomous navigation and mapping using YDLidar. Computer vision on a Kinect sensor and embedded subsystems design was done on ROS, with necessary instrumentation and actuators.

bottom of page