Vehicle Trajectory Interpretation Using Surrounding Feature Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driver assistance systems face challenges in accurately interpreting trajectories between vehicles due to measurement inaccuracies, which can lead to misinterpretations when communication occurs, especially when rotation and orientation errors are involved, making it difficult to determine if a vehicle intends to stay in its lane or change lanes.
Innovation Solution
The method involves describing trajectories relative to the sending vehicle's topology, allowing receiving vehicles to use surrounding features for matching and error compensation, using cameras and sensors to capture information about the surroundings, and transmitting this information in a global coordinate system for accurate localization and interpretation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If trajectories are communicated between vehicles using standard coordinate systems, then position information can be transmitted, but measurement inaccuracies and rotation errors cause misinterpretation of driving intentions
Solution Approach 1:
The patent introduces an intermediary coordinate transformation process that converts trajectories from the sending vehicle's local coordinate system to a global coordinate system using map data and surrounding features as reference. This intermediary transformation layer resolves the direct coordinate mismatch problem between vehicles, enabling accurate interpretation of driving intentions despite individual vehicle measurement errors.
Solution Approach 2:
The system uses surrounding features (road markings, buildings, terrain) as feedback references to verify and correct trajectory interpretations. By continuously comparing communicated trajectories against these stable environmental features, the system can detect and compensate for measurement inaccuracies and rotation errors in real-time.
2Ease of operation
If vehicles use their own local coordinate systems for trajectory description, then position data can be generated, but rotation and orientation errors make it difficult to determine lane change intentions
Solution Approach 1:
Instead of transforming trajectories from global to local coordinates (the conventional approach), the patent inverts the process by transforming from local to global coordinates. This inversion allows each vehicle to maintain its simple local coordinate generation while the central system performs the coordinate transformation using map data, thereby preserving ease of operation at the vehicle level while achieving high precision in lane change detection at the system level.
3Quantity of substance
If trajectory information is transmitted without environmental context, then communication bandwidth is reduced, but receiving vehicles cannot accurately interpret driving intentions
Solution Approach 1:
The patent extracts only the essential elements needed for trajectory interpretation: surrounding features (road markings, buildings, terrain) and map data. By separating these contextual elements from the raw trajectory coordinates, the system transmits minimal necessary information while preserving complete interpretability of driving intentions, thus reducing communication overhead without losing critical information.
Data Source
AI summary
A method for providing information about a probable driving intention of a first vehicle. The method includes ascertaining, using a locating device, first position data, which indicate the approximate position of the first vehicle, and map data for the first position data, and capturing, using cameras and/or sensors in the first vehicle, information about the surroundings in proximity to the first vehicle. The method also includes determining information about a probable trajectory of the first vehicle, and providing information about a probable driving intention of the first vehicle, based on the information about the probable trajectory of the first vehicle, the position data, the information about the surroundings and the map data. The method also includes transmitting the information about the probable driving intention of the first vehicle to a second vehicle and/or one or more vehicle-external entities.