Autonomous Route Planning Using AI and Driver Preference History

The patent segments the route determination process into two distinct models: a first model that calculates optimal routes based on traffic information and road conditions for efficiency, and a second model that determines preferred routes based on user driving history and habits for adaptability. This segmentation allows both objectives to be pursued independently and then integrated.

The patent merges the outputs of the first model (optimal route) and the second model (preferred route) to determine the final driving route. By combining both models' results, the system achieves both driving efficiency and user preference adaptation simultaneously, resolving the technical contradiction between productivity and adaptability.

The patent implements self-service through automated collection and analysis of user driving history data. The system automatically tracks user driving behaviors, preferences, and patterns over time, using this data to train the second model without requiring manual input from users. This self-service approach improves recognition accuracy while keeping the interface simple for users.

The patent applies preliminary action by pre-training both the first model and second model before actual route determination. The models are trained in advance using historical data and user driving patterns, so that when route determination is needed, the system can quickly and accurately provide optimized routes without complex real-time calculations, thus managing complexity while maintaining high accuracy.