Vehicle Sensor Fusion Using Feature-Point Position Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for using position information from another vehicle to enhance advanced safety and automatic driving systems are prone to inaccuracies due to variations in map precision and positional deviations, leading to potential errors in collision prediction and vehicle control.
Innovation Solution
An outside sensing information processing device that calculates the position of objects detected by another vehicle relative to feature points sensed by the host vehicle, correcting positional errors and integrating this information into the host vehicle's environmental map to improve precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If position information from another vehicle is used directly without correction, then the quantity of sensing information is increased, but the measurement precision deteriorates due to map variations and positional deviations
Solution Approach 1:
The patent introduces feature points as intermediary reference objects to mediate between the another vehicle's detection data and the present vehicle's map. By detecting feature points (such as road markings, signs, or structures) that are visible to both vehicles, the system creates a common reference framework that enables accurate coordinate transformation and position correction, thus maintaining high precision while utilizing data from multiple sources
Solution Approach 2:
The system implements feedback by using the present vehicle's own detection of feature points and objects to verify and correct the position information received from another vehicle. The detected positions of feature points and objects by the present vehicle serve as feedback to adjust and refine the coordinate transformation parameters, ensuring that the integrated sensing information maintains high accuracy
2Productivity
If position information from another vehicle is used to expand detection coverage, then the productivity is improved, but the reliability deteriorates due to potential position errors affecting safety system operations
Solution Approach 1:
Feature points serve as reliable intermediary references that anchor the coordinate transformation process. By detecting feature points that are stationary and easily identifiable (such as road signs, curbs, or structures), the system establishes a trustworthy reference framework that ensures the reliability of transformed position information from another vehicle, preventing erroneous data from compromising safety system operations
Solution Approach 2:
The system performs preliminary detection of feature points and calculation of coordinate transformation parameters before integrating another vehicle's sensing data. This preliminary action establishes accurate transformation relationships in advance, ensuring that when position information from another vehicle is used to expand detection coverage, the data is already corrected and reliable for safety-critical operations
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention provides an outside sensing information processing device capable of suppressing the adverse influence on warning of an advanced safety system and an automatic driving system and vehicle control, at the time of using position information of an object detected by another vehicle, by suppressing a position error at the time of reflecting the existence of the object into a map of a present vehicle, verifying position information precision of the another vehicle, and avoiding use of position information of an object provided from the another vehicle depending on a situation. At the time of transmitting position information of an object or an indication detected by an outside sensor mounted in another vehicle or a road, the positions of feature points of an object (stationary object) and an indication as configuration information of an environmental map as the reference of a position are also detected and transmitted. When the present vehicle receives those pieces of information, the position of the object or the indication on the environmental map of the present vehicle is calculated (corrected) on the basis of the relative position relation between the position information of an object or an indication as a target and the position of a feature point and the position of the feature point grasped by the present vehicle. Another vehicle also transmits position information of itself. The present vehicle checks an error of the position information of the vehicle grasped by the another vehicle from the position information of the another vehicle and the relation of the position of a feature point detected by the another vehicle and the position of the feature point detected by the present vehicle. In the case where the error is large or unstable, use of the position information from the another vehicle is avoided.