Traffic Light State Monitoring for Accurate Lane Assignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face inaccuracies in determining the lane associated with traffic signals and the current state of traffic signals, which can impact their navigation and decision-making processes.
Innovation Solution
A system and method that utilize a processor to receive and correlate visual and vehicle data from multiple sources, employing a joint Hidden Markov Model and trained machine learning models to determine the state of traffic lights and assign them to specific lanes, incorporating time synchronization and vehicle velocity analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If autonomous vehicles use individual sensing devices to detect traffic signals, then each vehicle can independently navigate, but the accuracy of traffic signal detection and lane assignment deteriorates due to limited individual sensing capabilities
Solution Approach 1:
The patent combines visual data from multiple vehicles' cameras with vehicle data (position, speed, acceleration) to collectively determine traffic light states and lane assignments. This merging of data sources improves both detection reliability and measurement precision by leveraging the collective sensing capabilities of the vehicle fleet rather than relying on individual vehicle sensors alone.
Solution Approach 2:
The system introduces a centralized processing system that acts as an intermediary between individual vehicle sensors and the autonomous driving decisions. This intermediary collects, synchronizes, and processes data from multiple vehicles to produce accurate traffic signal detection and lane assignment information, resolving the contradiction between individual sensing limitations and the need for high accuracy.
2Measurement precision
If the system integrates data from multiple vehicles to improve detection accuracy, then traffic signal state determination improves, but the system complexity increases due to data correlation and synchronization requirements
Solution Approach 1:
The system uses the vehicles' own existing sensors and data (cameras, position, speed, acceleration) to collectively solve the traffic signal detection problem. Each vehicle's data serves the collective purpose, and the system processes this self-generated data without requiring external infrastructure, thereby improving accuracy while managing complexity through resourceful use of available data.
Solution Approach 2:
The patent transforms the problem by changing the parameters used for detection - instead of relying on a single vehicle's visual data alone, the system incorporates multiple parameters including vehicle position, speed, acceleration, and visual data from multiple sources. This multi-parameter approach improves determination accuracy while the systematic processing of these parameters manages the complexity through structured data integration.
3Device complexity
If the system processes visual data and vehicle data separately, then data processing is simpler, but the correlation between traffic light state and vehicle behavior is insufficient for accurate lane assignment
Solution Approach 1:
The patent merges visual data processing with vehicle behavior data processing into a unified analysis framework. By combining these previously separate processing streams, the system establishes correlations between traffic light states and vehicle responses (acceleration, deceleration, lane changes), thereby improving lane assignment reliability while accepting the necessary increase in processing complexity as a trade-off for enhanced accuracy.
Data Source
AI summary
Systems and methods are provided for interpreting traffic information. A method includes: receiving, by a processor, visual data from vehicles, wherein the visual data is associated with an intersection of a roadway having one or more lanes; receiving, by the processor, vehicle data from the vehicles, wherein the vehicle data is associated with the intersection of the roadway; determining, by the processor, a first state of a traffic light associated with the intersection based on the visual data; determining, by the processor, a second state of the traffic light associated with the intersection based on the vehicle data; correlating, by the processor, the first state and the second state based on a time synchronization; assigning, by the processor, the traffic light to a lane of the roadway based on the correlating; and communicating, by the processor, the traffic light to lane assignment for use in controlling a vehicle.