Adjusting Virtual Road Boundaries for Autonomous Vehicle Navigation Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately determining the drivable area of a road due to uncertainties in their positioning and heading angles, which affects the safety and reliability of their navigation, especially under changing conditions like weather or sensor performance.
Innovation Solution
A system that retrieves measured position and heading angle data, determines uncertainty values based on this data and map information, and adjusts the width and heading angle of the drivable portion of a two-dimensional virtual road representation to account for these uncertainties, ensuring a high probability of the area being within the road boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the drivable area is determined using fixed road boundaries from map data, then the navigation system is simple to implement, but the accuracy and safety of path planning deteriorates due to positioning and heading uncertainties
Solution Approach 1:
The patent applies dynamics by making the road boundaries adjustable rather than fixed. The system dynamically adapts the virtual road representation by modifying boundary positions based on real-time uncertainty values from the positioning and heading angle measurements, allowing the drivable area to flexibly respond to changing confidence levels in vehicle state estimation
Solution Approach 2:
The system changes parameters by adjusting the width and heading angle of the drivable portion based on uncertainty values. When positioning or heading uncertainty increases, the system modifies the geometric parameters of the virtual road boundaries to maintain safety margins, effectively adapting the drivable area definition to current measurement quality
2Reliability
If the road boundaries are adjusted dynamically based on uncertainty values, then the safety and reliability of navigation improves, but the computational complexity and processing time increase
Solution Approach 1:
The system implements feedback by continuously monitoring uncertainty values from the positioning unit and heading angle measurements, then using this feedback to adjust the drivable area boundaries. This closed-loop approach ensures that the virtual road representation remains consistent with the actual confidence in vehicle state estimation, improving safety through adaptive boundary adjustment
Solution Approach 2:
The patent applies preliminary action by pre-defining the relationship between uncertainty values and boundary adjustments. The system has predetermined rules for how much to adjust boundaries based on measured uncertainty levels, allowing rapid computation without complex real-time optimization, thus balancing safety with processing efficiency
3Reliability
If the drivable area width is reduced to account for positioning uncertainty, then the probability of staying within road boundaries increases, but the available navigation space and vehicle maneuverability decrease
Solution Approach 1:
The system applies local quality by adjusting the drivable area boundaries locally based on where uncertainty is most critical. Rather than uniformly reducing the entire drivable area, the system selectively modifies boundary positions in regions where positioning or heading uncertainty would most likely cause boundary violations, preserving maneuverability in low-uncertainty regions while ensuring safety in high-uncertainty regions
Data Source
AI summary
A system for adjusting a road boundary of a two dimensional virtual representation of a road for path planning of an autonomous vehicle includes a retrieval unit to retrieve a measured position and measured heading angle of the vehicle with respect to a reference. The system also includes an uncertainty unit to receive the measured position, the measured heading angle and map data of the road, and to determine a position uncertainty value and a heading angle uncertainty value representing an amount of uncertainty in the vehicle position and heading angle, respectively, on the road based on the input. The system also includes an adjustment unit to adjust a width and a heading angle of a drivable portion of the two dimensional virtual representation of the road, wherein the adjustment of the width and heading angle is based on the position uncertainty value and the heading angle uncertainty value, respectively.