Lane Change Decision Logic With Personalized Longitudinal Speed Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lane change decision-making technologies in autonomous driving face challenges such as complex logic, difficulty in reflecting passenger preferences, and a failure to balance safety and efficiency, particularly in rule-based systems, while learning-based systems are not ready for mass production due to data restrictions and black-box system security.
Innovation Solution
A lane change decision-making method that integrates longitudinal control to formulate an autonomous lane change policy, allowing users to set personalized speed preferences and thresholds, and assess lane change feasibility based on relative speeds, distances, and motion parameters to ensure safe and efficient lane changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If rule-based lane change decision-making is used, then the system is ready for mass production, but the logic becomes complex and fails to reflect passenger preferences
Solution Approach 1:
The patent transforms the complex rule-based decision-making logic into a parameter-driven approach. By defining lane change conditions through quantitative parameters (relative speed thresholds, distance thresholds, acceleration limits) rather than complex logical rules, the system maintains simplicity while being ready for mass production. The controller executes straightforward parameter comparisons instead of complex logical evaluations.
2Ease of manufacture
If rule-based lane change decision-making is used, then the system is ready for mass production, but the ability to reflect passenger preferences is reduced
Solution Approach 1:
The patent introduces dynamic adaptability by allowing the expected speed parameter to be adjusted based on passenger preferences. The system can adapt to different driving styles (conservative vs. aggressive lane changing) by modifying the expected speed threshold, enabling the same hardware platform to serve different user needs without sacrificing mass production readiness.
3Adaptability or versatility
If learning-based lane change decision-making is used, then passenger preferences can be reflected, but data restrictions and black-box system security prevent mass production
Solution Approach 1:
The patent replaces the learning-based artificial intelligence system with a deterministic control algorithm based on classical control theory. By substituting the black-box learning model with a transparent mathematical model (PID controller with parameter adjustments), the system achieves both mass production readiness and the ability to reflect passenger preferences through adjustable parameters, eliminating security and data concerns.
4Device complexity
If lane change decision-making does not consider longitudinal control, then the decision process is simpler, but safety and efficiency are compromised
Solution Approach 1:
The patent merges lane change decision-making with longitudinal control into a unified decision framework. By combining lateral lane change intentions with longitudinal speed and acceleration considerations, the system ensures that lane changes are executed only when both lateral and longitudinal conditions are safe, improving reliability without excessive complexity through integrated parameter evaluation.
Data Source
Figure 1~2
Figure 3
Figure 4A
AI summary
A lane change decision-making method and apparatus, and a storage medium are provided, and relate to the field of autonomous driving technologies. The method may be applied to the fields of intelligent vehicles, new energy vehicles, and the like, and includes: on a first lane on which a target vehicle is located, obtaining longitudinal traveling information of the target vehicle when there is a first obstacle in front of the target vehicle (201); and performing lane change when the longitudinal traveling information meets a preset lane change condition, where the preset lane change condition includes: a first relative speed is less than zero and a second relative speed is less than zero, where the first relative speed is a relative speed between a first speed and a set first expected speed, and the second relative speed is a relative speed between the first speed and a second expected speed (202). According to the method, an intelligent vehicle can fully and intuitively reflect an intent and a preference of a user when satisfying safety, to implement personalized setting of a lane change frequency, and improve lane change decision-making performance.