Traffic Light Yaw Detection for Relevant Signal Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately detecting and classifying traffic light signals, particularly when multiple traffic light heads are present and yawed relative to the vehicle's sensors, leading to ambiguity in determining relevant signals for the intended direction of travel.
Innovation Solution
Utilizing machine learning algorithms, specifically convolutional neural networks (CNNs), to classify traffic light signals and determine yaw angles, combined with map data and voting schemes to disambiguate relevant traffic light heads and enhance signal detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple traffic light heads are present at an intersection, then the coverage and control capability are improved, but the difficulty of detecting and measuring the relevant signal increases due to yaw ambiguity
Solution Approach 1:
The patent segments the traffic light detection task by classifying each detected traffic light head into multiple yaw categories (first yaw class, second yaw class, third yaw class) based on their orientation relative to the vehicle. This segmentation allows the system to separately process and evaluate signals from different directions, resolving the ambiguity caused by multiple traffic light heads at the intersection.
Solution Approach 2:
The patent introduces a yaw angle dimension to the traffic light detection system. By measuring and categorizing the yaw angles of traffic light heads, the system adds an orientational dimension to the detection process, enabling it to distinguish between traffic lights facing different directions and identify the relevant signal for the vehicle's current direction of travel.
2Measurement precision
If machine learning algorithms and voting schemes are used to disambiguate traffic light heads, then the measurement precision is improved, but the device complexity increases
Solution Approach 1:
The patent implements a voting scheme where multiple traffic light heads cast votes for different signal colors based on their detected states. The system aggregates these votes and determines the final signal classification based on the majority or weighted majority. This feedback mechanism improves measurement precision by cross-validating signals across multiple traffic light heads while managing complexity through a structured voting process.
Data Source
AI summary
Techniques are described herein for determining a yaw value for a traffic light head. The technique comprises obtaining sensor data via the sensor associated with the vehicle, the sensor data including a traffic light head. Based at least in part on the sensor data, a yaw value is determined for the traffic light head indicative of a degree that the traffic light head is yawed with respect to a line of sight of the sensor or the vehicle. Based at least in part on the yaw value, an extent that the traffic light head is associated with navigation of the vehicle is determined. The vehicle is then controlled based at least in part on the extent that the traffic light head is associated with the navigation of the vehicle.