Autonomous Vehicle Stop Location Detection at Changing Intersections
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-driving vehicles face challenges in determining appropriate stopping locations at intersections, especially when there are no painted stop lines, due to changes in road conditions or geometry, which can affect their ability to detect approaching vehicles and objects.
Innovation Solution
A method using onboard processors to evaluate training data and generate classification, localization, and uncertainty estimation channels to predict stop line locations, incorporating sensor data from lidar and camera images, and projecting these points onto a lane tangent to determine a predicted stopping location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If the vehicle relies on pre-mapped stop line locations, then the stopping point can be determined in advance, but the stop location may become inaccurate due to construction, road resurfacing, or road wear and tear
Solution Approach 1:
The system performs preliminary detection of stop line features using classification and localization channels before final stop location determination. This allows the vehicle to prepare stop location candidates in advance while still adapting to actual road conditions observed in real-time, resolving the contradiction between early determination and accuracy.
Solution Approach 2:
The system uses uncertainty estimation channels to continuously evaluate the reliability of detected stop line features. When the stop line is hard to detect or may have changed, the uncertainty metric triggers re-evaluation and adjustment of the stop location, providing feedback that maintains accuracy despite relying on pre-mapped data.
2Adaptability or versatility
If the vehicle stops at a location without a painted stop line, then the vehicle can operate on residential or lower speed streets, but it becomes challenging to determine the appropriate stopping point in real time
Solution Approach 1:
The system applies different detection strategies based on local road characteristics. For roads with stop lines, it uses direct detection; for roads without stop lines, it infers stop locations from crosswalk positions or intersection geometry. This localized adaptation handles diverse stopping scenarios without requiring a single complex solution for all cases.
Solution Approach 2:
Instead of directly detecting stop lines when none are present, the system inverts the approach by detecting related features (crosswalks, intersection boundaries) and inferring the stop location from these alternative references, enabling operation on streets without painted stop lines.
3Productivity
If the stop location is determined without considering uncertainty estimation, then the processing is simpler and faster, but the vehicle cannot adapt to changing road conditions or detect changes in stop line location
Solution Approach 1:
The system implements uncertainty estimation selectively rather than for all stop location determinations. It applies full uncertainty analysis when stop lines are hard to detect or may have changed, while using simpler methods when confidence is high. This partial application maintains reliability for critical cases while preserving processing efficiency for routine situations.
Data Source
AI summary
The technology relates to approaches for determining appropriate stopping locations at intersections for vehicles operating in a self-driving mode. While many intersections have stop lines painted on the roadway, many others have no such lines. Even if a stop line is present, the physical location may not match what is in store map data, which may be out of date due to construction or line repainting. Aspects of the technology employ a neural network that utilizes input training data and detected sensor data to perform classification, localization and uncertain estimation processes. Based on these processes, the system is able to evaluate distribution information for possible stop locations. The vehicle uses such information to determine an optimal stop point, which may or may not correspond to a stop line in the map data. This information is also used to update the existing map data, which can be shared with other vehicles.