Geo-spatial Sensor Matching Using Heading Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current sensor-based vehicle navigation systems face high memory requirements and error-prone processes when identifying stationary objects on roads, as they need to store and match GPS coordinates with extensive map data, leading to inefficient and inaccurate object recognition.
Innovation Solution
The proposed method reduces memory usage by storing only known static objects' geospatial coordinates and headings, applying spatial, heading, and distance filtering criteria to match GPS messages directly with static objects, eliminating the need for traditional map data storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional map-matching is used to match GPS coordinates to static objects, then object identification accuracy can be maintained, but memory requirements become very high and the process becomes expensive
Solution Approach 1:
The patent extracts only the essential data elements needed for matching (static object locations, headings, and observable distances) from the complete map data structure. By removing unnecessary map elements and retaining only critical matching information, the system achieves accurate object identification with significantly reduced memory consumption.
Solution Approach 2:
The patent applies local quality by storing complete map data only in regions where static objects are present, while using simplified or no map data in regions without static objects. This selective data storage approach maintains matching accuracy where needed while minimizing overall memory requirements.
2Measurement precision
If complete map data is stored for matching, then matching accuracy is maintained, but the system becomes computationally expensive and complex
Solution Approach 1:
The patent extracts and stores only the critical matching parameters (object locations, headings, distances) while discarding the complex complete map data structure. This extraction reduces computational complexity and data storage requirements while preserving the essential information needed for accurate matching.
Solution Approach 2:
Instead of storing complete map data and filtering for matching information, the patent inverts the approach by storing only pre-extracted matching-critical data from the beginning. This inversion simplifies the system architecture and reduces computational burden while maintaining matching accuracy.
3Adaptability or versatility
If traditional two-stage matching is used, then comprehensive object detection is possible, but the process is error-prone due to camera seeing across road links
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing the heading information for each static object before the matching process. This pre-computation allows the system to quickly filter objects based on heading compatibility with the vehicle's orientation, reducing false matches from objects on adjacent links that the camera might mistakenly detect.
Solution Approach 2:
The patent changes the matching parameters by incorporating heading angle as a critical filtering criterion in addition to spatial distance. By adding this angular parameter to the matching process, the system can reliably distinguish between objects on the current road link versus objects on adjacent links, significantly improving matching reliability.
Data Source
AI summary
A method, system, and computer program product is provided, for example, for matching a geospatial message to one or more static objects on a link. The method may include identifying a location of an observation point for observing the one or more static objects from a vehicle on the link. The method may further include applying spatial filtering criteria to each of the one or more static objects based on an observable distance of each of the one or more static objects from the location of the observation point to filter the one or more static objects. The filtering of the one or more static objects may provide one or more filtered static objects. The method may further include calculating a heading of the vehicle and further using the heading of the vehicle for applying a heading filtering criteria to the one or more filtered static objects based on the heading of the vehicle and a pre-computed heading of each of the one or more filtered static objects to provide one or more candidate objects. Additionally, the method may include applying a distance filtering criteria to each of the one or more candidate objects to provide one or more remaining objects. Finally, the method may include matching the geospatial message to at least one of the one or more remaining objects based on the distance filtering criteria.