Autonomous Vehicle Path Confidence Control Under Planning Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving vehicles face safety risks due to changes in traffic and road conditions, as conventional motion planning generates paths that may not reflect the current environment, leading to potential dangerous situations if the vehicle tracks the original path without updates.
Innovation Solution
The system generates paths with confidence levels for each segment, allowing the vehicle to track only segments with high confidence levels and adjust or modify parameters for uncertain segments, ensuring safety by issuing appropriate control commands based on perception data, map information, and traffic rules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the vehicle tracks the original planned path without considering traffic and road condition changes, then the vehicle maintains simple control operations, but the vehicle exposes itself to dangerous conditions because the path does not reflect the actual driving environment
Solution Approach 1:
The control system dynamically adjusts path tracking based on real-time confidence levels of path segments. The vehicle transitions from rigid path following to adaptive path adjustment, where control commands are modified according to the validity and confidence of each path segment, resolving the contradiction between maintaining simple control and ensuring safety in changing environments
Solution Approach 2:
The system implements feedback by continuously evaluating the confidence level of planned path segments against actual driving conditions. This feedback mechanism allows the vehicle to identify when the original path is no longer valid and adjust control commands accordingly, balancing operational simplicity with safety requirements
2Reliability
If the vehicle adjusts control commands frequently to reflect current traffic and road conditions, then the vehicle improves safety by tracking valid path segments, but the control operations become more complex
Solution Approach 1:
The planned path is segmented into multiple path segments, each with its own confidence level. This segmentation allows the system to apply different control strategies to different segments - tracking high-confidence segments and adjusting or skipping low-confidence segments - thereby improving safety without requiring complex control adjustments across the entire path
Solution Approach 2:
The system applies local quality by treating different path segments differently based on their confidence levels. High-confidence segments receive standard tracking control, while low-confidence segments receive adjusted control commands, creating a nuanced control approach that improves safety where needed while maintaining ease of operation where conditions are stable
3Reliability
If the vehicle tracks only path segments with high confidence levels, then the vehicle ensures safety by avoiding uncertain segments, but the vehicle may lose time by not following the original efficient path
Solution Approach 1:
The system applies partial action by tracking only the necessary high-confidence path segments rather than the entire original path. This selective approach ensures safety by avoiding uncertain segments while minimizing time loss by efficiently following valid portions of the planned path and quickly transitioning to new path planning when needed
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
According to some embodiments, described herein is a method and a system for guaranteeing safety at a control level of an ADV when at least a portion of a planned path generated by a planning module of the ADV is uncertain due to traffics and/or road condition changes. The planning module, when generating a path, also generate a confidence level of each segment of the path based on one or more of perception data, map information, or traffic rules. The confidence levels are decreasing further away from the ADV. When the control module of the ADV obtains the path and the associated confidence levels, the control module issue control commands to track only one or two segments whose confidence levels exceeds a threshold hold, and issue default control commands for the rest of the path.