driveblocks develops technologies for fully autonomous commercial vehicles on its Mapless Autonomy Platform. The modular system includes software components for raw data processing, real-time detection of a vehicle’s environment and sensor data fusion, i.e. merging data from multiple sensor modalities into a consistent environmental model. The driveblocks solution is significantly safer than other systems because it uses perception, i.e. sensor-based recognition of the environment, rather than maps, which can become outdated from one day to the next. It is based on Transformer Neural Networks, the same network architecture that has led to recent breakthroughs in Large Language Models (LLMs).