Moving Device Position Estimation Using Shared Object Relationships

The patent uses moving objects (pedestrians, vehicles, etc.) as intermediary elements to indirectly determine the moving device's position. Instead of directly observing the moving device with infrastructure sensors, the system observes moving objects that are visible to both the infrastructure sensor and the onboard sensor, then uses the relative positional relationships to infer the moving device's position. This intermediary approach resolves the contradiction by maintaining position estimation accuracy even when direct observation is blocked.

The moving device uses its own onboard sensor to actively observe and track moving objects in its vicinity. This self-service approach allows the device to contribute its own sensing capabilities to the position estimation process, rather than passively relying solely on infrastructure sensors. The onboard sensor captures images of moving objects, which are then used to establish relative positional relationships for determining the device's position, thereby improving accuracy in crowded environments without significantly increasing overall system complexity.

The system creates a virtual model or copy of the spatial relationships by capturing images of moving objects with the onboard sensor and comparing them with images from infrastructure sensors. Instead of directly processing complex coordinate transformations and positional relationships, the system works with image data copies that represent the same spatial information. This copying approach simplifies the processing by allowing direct comparison of image-based positional relationships rather than processing raw sensor data from multiple sources.