Traffic Light Recognition Using Map-Vision Fusion at Intersections
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional vehicle vision systems rely solely on image sensors for traffic light recognition, which may not be robust enough to accurately determine the relevant traffic light structure for a vehicle approaching an intersection, especially in scenarios with curved roads, leading to potential accidents and an uncomfortable driving experience.
Innovation Solution
A vehicular vision system that combines image data from a front camera with map data and GPS information to determine the most relevant traffic light structure, using a fusion module to aggregate relevancy scores and provide timely alerts or control vehicle speed based on both sensor and location-based data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If vision-based decision making is used alone, then the system simplicity is maintained, but the accuracy of traffic light recognition deteriorates in complex scenarios
Solution Approach 1:
The patent combines map data (providing location-based traffic light structure information) with vision-based image data (providing real-time traffic light status) through a data fusion module. This merging of multiple data sources resolves the contradiction by achieving high recognition accuracy without requiring a completely complex system architecture, as each data source complements the other's strengths and weaknesses.
2Measurement precision
If map data and vision data are fused, then the accuracy of traffic light recognition is improved, but the device complexity increases
Solution Approach 1:
The patent segments the traffic light recognition system into distinct functional modules: a map data processing module that handles location-based traffic light structure information, a vision data processing module that handles real-time image capture and analysis, and a data fusion module that integrates both sources. This segmentation manages system complexity by organizing functions into separate, manageable components with well-defined interfaces.
Solution Approach 2:
The patent introduces a data fusion module as an intermediary component that receives processed information from both map data sources and vision-based image processing, aggregates relevancy scores from multiple sources, and produces a unified traffic light recognition result. This intermediary manages the complexity of integrating multiple data sources by providing a standardized interface and systematic aggregation process.
3Speed
If only vision-based detection is used, then the response time is fast, but the reliability of traffic light status determination deteriorates in complex scenarios
Solution Approach 1:
The patent uses map data to preliminarily determine the expected traffic light structure and location before the vehicle reaches the intersection. This preliminary information from map data provides a prediction framework that guides the vision-based detection process, allowing the system to prepare for upcoming traffic lights and verify detected signals against expected configurations, thereby improving reliability without significantly increasing response time.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the accuracy of traffic light recognition, reduces the risk of accidents, and provides a more comfortable driving experience by ensuring timely and appropriate vehicle maneuvers at intersections.
Implementation Method 1
A vehicular vision system includes a camera disposed at a vehicle equipped with the vehicular vision system that views at least forward of the equipped vehicle. The camera is operable to capture image data.
Data Source
AI summary
A vehicular vision system includes a camera disposed at a vehicle and viewing at least forward of the vehicle. The vehicular vision system, as the vehicle travels along a traffic lane of a road and via processing of image data captured by the camera and based on map data, determines presence of a plurality of traffic lights at an intersection and identifies a target traffic light from the plurality of traffic lights that is associated with the traffic lane along which the equipped vehicle is traveling as the vehicle approaches the intersection. The vehicular vision system determines, via processing of image data captured by the camera, status of the identified target traffic light. The vehicular vision system, responsive to determining the status of the identified target traffic light, alerts a driver of the vehicle of the status of the relevant traffic light and/or controls speed of the vehicle.