Dynamic Replanning Threshold for Autonomous Vehicle Trajectory Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving vehicles face challenges in determining whether to perform regular or replanning of their route or trajectory due to discrepancies between planned and actual states, such as position, speed, and heading direction, which can lead to inefficiencies and safety issues.
Innovation Solution
A replanning determination mechanism that calculates a replanning score based on differences and trends between current and target states, using sensors and machine learning algorithms to decide between regular and replanning operations, with thresholds for determining when to adjust the route segment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If replanning is performed frequently to improve navigation accuracy, then navigation accuracy is improved, but computational overhead and processing time increase
Solution Approach 1:
The system dynamically changes the replanning threshold parameter based on driving conditions. When the discrepancy between actual and planned states exceeds the threshold, replanning is triggered. The threshold itself can be adjusted according to situation, allowing frequent replanning when needed (improving accuracy) while avoiding unnecessary replanning in stable conditions (reducing processing time).
Solution Approach 2:
The system continuously monitors the discrepancy between actual vehicle state and planned trajectory, using this feedback to determine when replanning is necessary. This feedback mechanism ensures replanning occurs only when needed, balancing navigation accuracy with computational efficiency by avoiding unnecessary replanning cycles.
2Reliability
If replanning is performed frequently to improve safety, then safety is improved, but system complexity and computational load increase
Solution Approach 1:
The system uses adjustable safety thresholds that can be modified based on environmental factors, vehicle speed, and traffic conditions. This allows the system to maintain high safety standards through frequent replanning when risks are detected, while reducing computational load during safe, stable driving conditions.
Solution Approach 2:
Continuous monitoring of trajectory adherence provides feedback that triggers replanning only when safety concerns arise. This feedback-driven approach maintains safety through necessary replanning while avoiding the complexity of constant replanning, as the system responds only to actual deviations from the planned path.
3Speed
If the replanning threshold is set low to improve responsiveness, then responsiveness is improved, but false positives and unnecessary replanning increase
Solution Approach 1:
The system dynamically adjusts the replanning threshold parameter based on driving conditions. In high-risk or rapidly changing environments, the threshold is lowered to improve responsiveness. In stable conditions, the threshold is raised to prevent false positives, thereby maintaining responsiveness when needed while preserving operational efficiency during normal operation.
Solution Approach 2:
The replanning threshold is not fixed but dynamic, adapting to current driving conditions. This allows the system to be highly responsive (low threshold) when conditions warrant attention, while maintaining efficiency (high threshold) during routine driving, thus resolving the contradiction between responsiveness and operational efficiency.
4Productivity
If the replanning threshold is set high to reduce computational load, then computational efficiency is improved, but responsiveness to actual deviations decreases
Solution Approach 1:
The system changes the replanning threshold parameter dynamically rather than using a fixed high value. This allows computational efficiency to be maintained during normal operation (high threshold) while ensuring responsiveness is restored when conditions require it (lower threshold), thus resolving the contradiction between computational efficiency and responsiveness.
Data Source
AI summary
According to one embodiment, a replanning determination mechanism is utilized to determine whether an ADV needs replanning of the route or trajectory based on the driving status and conditions of the ADV at the point in time, such as, for example, the difference between the planned position and the actual position, difference between a current speed and a target speed, and a difference between a current heading direction and a target heading direction. Such differences as a whole represents an error between the planning and the actual response from the ADV. If the error is greater than a predetermined threshold, replanning may be performed; otherwise, a regular planning is performed.