Radar Sensor Fusion for Accurate Object Width Measurement

The patent segments the detection process by distinguishing between sensor target objects (individual reflection points detected by radar) and fusion target objects (grouped objects representing actual physical objects). By segmenting and separately managing multiple sensor targets that belong to the same physical object, the system maintains comprehensive detection coverage while enabling precise width measurement through proper selection and grouping of sensor targets.

The patent merges multiple sensor target objects into a single fusion target object when they belong to the same three-dimensional object. This merging process combines detection data from multiple radar sensors and multiple reflection points into a unified representation, allowing the system to achieve both comprehensive detection coverage and accurate width measurement by treating the grouped targets as a single measurement unit.

The patent applies local quality by treating different sensor target objects differently based on their characteristics. Instead of uniformly using all sensor targets for width calculation, the system selectively chooses appropriate sensor targets (e.g., those from the same radar sensor, or with specific spatial relationships) for width measurement while using others for different purposes. This differentiated approach ensures reliable width information without unnecessarily complicating the calculation process.

The patent introduces a fusion target object as an intermediary between sensor target objects and final target object information. This intermediary layer processes and validates the relationship between multiple sensor targets and the actual physical object, ensuring that only reliable sensor target data is used for width calculation. The fusion target object acts as a mediator that filters and organizes data, improving reliability without significantly increasing overall system complexity.