Project Overview
AMZ Racing is ETH Zurich's premier Formula Student team, competing in the world's most challenging autonomous racing competitions. As a key contributor to the autonomous systems team, I developed cutting-edge perception and control algorithms that enable our race car to navigate complex tracks at speeds exceeding 100 km/h.
My work focused on creating robust SLAM (Simultaneous Localization and Mapping) systems and trajectory optimization algorithms that could operate reliably in the extreme conditions of competitive racing. This project pushed the boundaries of what's possible in autonomous vehicle technology, requiring innovations that go far beyond typical road vehicle applications.
Technical Systems
SLAM Algorithms
Real-time mapping and localization for dynamic racing environments
Trajectory Optimization
High-speed path planning with dynamic obstacle avoidance
Perception Systems
Multi-sensor fusion for cone detection and track boundary identification
Control Systems
Precision vehicle dynamics control for racing performance
Racing Innovations
Formula Student autonomous racing presents unique challenges that don't exist in regular autonomous driving:
- High-Speed Processing: Developed algorithms capable of real-time decision making at racing speeds up to 120 km/h
- Precision Navigation: Achieved centimeter-level accuracy in cone detection and track mapping
- Minimal Latency: Optimized perception-to-control pipeline for sub-50ms response times
- Performance Optimization: Implemented racing line optimization for maximum speed through corners
Technical Challenges
Dynamic Environment Mapping: Unlike road driving, racing tracks are defined by temporary cone layouts that change for each event. I developed adaptive SLAM algorithms that could quickly build accurate maps of unknown tracks while maintaining real-time performance during high-speed runs.
Extreme Performance Requirements: Racing demands split-second decisions at the limits of vehicle dynamics. I created control systems that could handle the car at the edge of tire grip while maintaining stability and safety margins required for autonomous operation.
Sensor Robustness: Racing environments present challenging conditions including vibration, dust, and varying lighting. I implemented robust sensor fusion techniques that maintained performance even when individual sensors were compromised or provided noisy data.
Competition Results
The autonomous systems I developed contributed to AMZ Racing's success in multiple Formula Student competitions across Europe. Our team consistently ranked among the top performers in autonomous driving events, with our perception and control systems enabling reliable completion of complex racing scenarios.
The innovations developed for racing have direct applications to road vehicle autonomy, particularly in emergency scenarios where vehicles must operate at the limits of their performance envelope. This work has contributed to advancing the state-of-the-art in autonomous vehicle technology.
Team Collaboration
Working with AMZ Racing provided invaluable experience in collaborative engineering on a world-class team. I worked closely with mechanical engineers, electrical engineers, and other software developers to integrate autonomous systems with the overall vehicle design. This interdisciplinary collaboration was essential for achieving the performance and reliability required for competitive racing.