Automated Lane-Changing Path Planning Using Predictive Motion State
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automatic lane-changing systems in vehicles may result in unsuccessful lane changes due to excessively large steering angles, which can lead to instability and poor user experience.
Innovation Solution
The method involves predicting the vehicle's motion state at the lane-changing destination, planning a lane-changing path and velocity, and iteratively adjusting the steering angle to ensure it falls within an effective range, using a two-degree-of-freedom vehicle model and polynomial fits for lane lines and velocity relationships to maintain smooth steering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the vehicle plans a lane-changing path using only basic lane-changing direction information and current position, then the lane-changing process is simple and fast, but the steering angle may become too large causing unsuccessful lane changing
Solution Approach 1:
The system performs preliminary prediction of the vehicle's motion state at the lane-changing destination before actually executing the lane change. By predicting future position, lane line information, and curvature in advance, the system can plan a more accurate path that avoids excessive steering angles, thus improving success rate without significantly increasing execution time
Solution Approach 2:
The system uses feedback from predicted motion states and actual vehicle state to iteratively optimize the lane-changing path. The predicted curvature and lane line information feed back into the path planning algorithm to adjust the trajectory, ensuring the steering angle remains within effective ranges while maintaining efficient lane-changing performance
2Reliability
If the vehicle increases the number of considered factors in path planning to avoid large steering angles, then the lane-changing success rate improves, but the computational complexity and time increase
Solution Approach 1:
The path planning process is segmented into distinct modules: motion state prediction module, lane line fitting module, curvature calculation module, and path optimization module. Each module handles a specific aspect of the problem independently, reducing overall complexity while comprehensively considering multiple factors for successful lane changing
Solution Approach 2:
Complex calculations such as polynomial fitting of lane lines, curvature computation, and motion state prediction are performed preliminarily before final path determination. This allows the main control system to work with pre-computed parameters rather than raw sensor data, reducing real-time computational burden while maintaining high planning accuracy
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
An automated vehicular lane changing method and apparatus, comprising: a vehicle in an automated driving state receiving a lane changing request, predicting a moving state of the vehicle when arriving at a lane changing endpoint, planning a lane changing path and a vehicle speed at an ith time for the vehicle, determining a steering wheel turning angle of the vehicle at the ith time according to a curvature of the lane changing path, the vehicle speed and a model of a steering wheel turning angle, if the turning angle falls in a range of effective turning angles, determining a moving state of the vehicle at an (i+1)th time according to a vehicle motion model, the moving state of the vehicle and the steering wheel turning angle at the ith time, if the difference between the moving state of the vehicle at the (i+1)th time and the predicted moving state is less than a preset error, driving the vehicle to change the lane according to the lane changing path planned for each time at the planned vehicle speed.