• Developed ROS Wrapper for Occupancy Grid Mapping (OGM) algorithm, enabling efficient 2D and 3D space mappings to track static and dynamic objects within global reference frames.

• Implemented DBSCAN Clustering to construct convex and concave hulls for environmental modeling, improving dynamic obstacle detection and supporting multipath mitigation within perception pipeline.

• Built a custom annotation tool to label objects for classification, enhancing training data quality for neural networks in radar perception stack.

• Developed a synchronization package for camera and point cloud data from rosbags, optimizing data alignment for enhanced tracking reliability.

• Validated Multi-Sensor Tracker Performance accuracy by comparing with ground truth data, ensuring high-precision performance metrics for perception stack.

• Designed and performed autonomous navigation systems, leveraging ROS for real-time 2D/3D SLAM and sensor fusion (LiDAR, IMU, GPS) to enable effective disaster area mapping and localization.

• Devised and incorporated advanced computer vision algorithms for hazardous sign, fire, and human detection using Convolutional Neural Networks (CNNs).

• Enhanced precision and efficiency of a 6-DoF robotic arm using Pololu motor drivers and linear actuators to perform intricate manipulation tasks in disaster scenarios.

• Engineered seamless communication between GUIs and HMI, enabling robust teleoperation and monitoring.

• Authored and presented five conference papers, showcasing robotics and SAR technology innovations, and represented projects in prestigious international competitions such as RoboCup, RoboCon, and the World Robot Summit (WRS).