Augmented Vision for Vehicle Safety Training Data

The patent creates virtual copies of real-world traffic scenarios through simulation. Instead of using actual recorded traffic data, the system generates synthetic training data by simulating vehicle environments, sensor behaviors, and traffic situations. This copying approach maintains the realism needed for effective training while eliminating the time and cost constraints of collecting and processing real-world data.

The patent introduces a sophisticated simulation environment as an intermediary between the simplicity of augmented reality methods and the realism of real-world data. This simulation layer acts as a mediator that incorporates detailed vehicle dynamics models, accurate sensor behavior models, and realistic environmental conditions, thereby generating training data that is both efficiently produced and sufficiently realistic for effective system tuning.

The patent combines multiple sensor types (cameras, RADAR, LIDAR, GPS, inertial sensors) into a unified trajectory determination system. By merging these sensors and their data processing functions into an integrated system with centralized coordinate transformation and trajectory calculation, the patent achieves high measurement precision while managing system complexity through unified architecture rather than separate independent systems.