Projects and Competitions

Projects

SCARA Robotic Arm

I developed an autonomous SCARA robot capable of picking parcel packages from a tote and placing them on a conveyor in a warehouse environment. The entire operation is hardcoded in Python, and the robot uses the YOLOv8 object detection algorithm to accurately detect parcels.

This project provided me with hands-on experience in applying inverse kinematics for a 2DOF robotic arm and integrating various hardware components, such as cameras and robotic arms, with software to perform specific tasks effectively.

For more information check: Project Repo

Competitions

RoboCup ARM Challenge 2024 - Finalist (Team: Log Robotics)

I developed the complete perception and control stack for the RoboCup ARM Challenge 2024. The objective was to pick and sort objects such as bottles, cans, and pouches based on their color and label. I utilized MATLAB tools to build the project’s code base, incorporating a custom fine-tuned YOLOv4 object detector and object pose estimation through Principal Component Analysis (PCA) for the perception stack.

My submission was selected as one of the top 6 finalists in the 2024 competition.

For more information check: Submission Video

F1TENTH Sim Racing - Phase 1 Qualified (Team: Log Robotics)

I developed a wall-following algorithm for the 1st F1Tenth Sim Racing League. The algorithm was implemented using ROS2 and Python.

My docker image of the algorithm can be found here: My docker repo Link

Qualification round performance: Link | Demo Video: Link

Autonomous Drones

Screenshot from 2024-08-22 21-05-34

I have experience working on the computer vision component of an autonomous drone, including some drone control using the DroneKit library. My work primarily involves simulating YOLOv8 algorithms in a Gazebo environment, with a focus on basic object tracking and detection tasks.