“Every sensor tells a story. I just help it speak.”
- 🧠 Passionate about
AI, robotics, autonomous driving and systems that think and move - 🔭 Currently exploring
Robotics, Reinforcement learning models, VLAMs - 🤝 Open to collaborations on
innovative tech projects
Multi-LiDAR Calibration | Autonomous Driving | Point Cloud Registration | Point Transformer V3 + Hierarchical Registration
Tech: Python, PyTorch, Docker, CUDA
Performs online registration of point clouds from multiple LiDARs with different viewpoints. Estimates the extrinsic transformation matrix for accurate sensor fusion in autonomous driving perception tasks.
Semantic Segmentation | Aerial Imagery | HD Maps | DeepLabV3+ | Bavarian Open Source Aerial Images
Tech: Python, PyTorch, OpenCV, CUDA
Performs semantic segmentation of lane markings from aerial images using DeepLabV3+ with Xception and ResNet101 backbones. Supports HD map creation for autonomous driving using Bavarian open data imagery.
Unity Simulation | Traffic Signal Compliance | Perception | Path Planning
Tech: ROS Noetic, C++, Python, Unity
Simulates autonomous driving in a Unity environment where the vehicle obeys traffic lights and navigates the shortest path. Integrates ROS Noetic for velocity commands, path planning, and sensor fusion. Includes manual override and real-time monitoring windows.
- 📫 Email: upendra.arun@tum.de