Vehicle Path Control Using Camera-Map Reliability Weighting

The patent combines multiple data sources (camera imaging data, GPS location data, digital map data, and yaw rate sensor data) into a unified road curvature determination system. By merging these different sensing modalities, the system achieves both adaptability to operate without lane markings and improved measurement precision through data fusion that compensates for individual sensor limitations.

The patent introduces digital map data as an intermediary reference framework that mediates between raw sensor measurements and final curvature calculations. The map data provides a stable geometric reference that helps correct drift in GPS and camera-based measurements, enabling accurate curvature determination even when direct visual lane markings are unavailable.

The patent designs a processing system that performs multiple functions using a unified architecture: it processes imaging data for lane detection, processes GPS data for location tracking, integrates digital map data for geometric reference, and combines yaw rate sensor data for motion compensation. This multi-functional approach improves measurement precision while managing device complexity through integrated processing rather than separate dedicated systems.

The patent dynamically adjusts the weighting and contribution of each sensor data source based on operating conditions. The system changes parameters such as the relative importance of camera versus GPS data depending on whether lane markings are visible, thereby improving precision adaptively while managing complexity through parameter tuning rather than structural redesign.