Junction Cut-In Prediction for Autonomous Vehicle Yielding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional lane-changing technologies face challenges in level-4 autonomous driving, particularly in congested junction sections, where predicting a yielding vehicle and maintaining stability with adjacent vehicles is difficult, leading to potential discomfort for users due to slower recognition and inability to identify vehicles intending to yield.
Innovation Solution
A vehicle running control method that collects information on adjacent vehicles in junction sections, determines the possibility of cut-in, and controls the host vehicle's path or velocity to display an intention to yield, using sensors to detect position, velocity, and acceleration, and applying critical velocity values to adjust travel strategies based on congestion and vehicle behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional lane-changing technology is used in junction sections, then the autonomous vehicle maintains a predetermined distance from adjacent vehicles for safety, but the recognition speed is slower than human recognition and the vehicle cannot identify yielding target vehicles in advance, causing user discomfort
Solution Approach 1:
The system performs preliminary identification of yielding target vehicles by analyzing lane change intentions before actual lane changes occur. It predicts which vehicles in the junction section will yield and identifies them in advance, allowing the autonomous vehicle to prepare appropriate responses beforehand rather than reacting after the fact.
Solution Approach 2:
The system extends detection beyond the immediate traveling lane to include adjacent junction section lanes. By monitoring vehicles in neighboring lanes and analyzing their lane change intentions, the system gains early warning capability and identifies yielding targets from an expanded spatial dimension before they enter the main traveling lane.
2Reliability
If the autonomous vehicle maintains a predetermined distance from adjacent vehicles in junction sections, then collision safety is improved, but the ability to predict cut-in vehicles and control traveling path/velocity is reduced, leading to instability in vehicle relationships
Solution Approach 1:
The system predicts cut-in vehicles by analyzing lane change intentions and determines yielding target vehicles in advance. This preliminary identification allows the autonomous vehicle to plan appropriate responses (path adjustment, velocity control) before the actual lane change occurs, maintaining stable vehicle relationships through proactive rather than reactive control.
Solution Approach 2:
The system dynamically adjusts traveling parameters (path and velocity) based on identified yielding target vehicles. When a yielding target is detected, the system modifies velocity or path parameters to display yielding intention or maintain stable spacing, creating adaptive control that responds to predicted vehicle interactions rather than maintaining fixed predetermined distances.
3Reliability
If conventional lane-changing technology is used, then lane changes are performed only when collision avoidance routes are generated, but this approach cannot satisfy level-4 autonomous traveling requirements where traveling must be possible without driver intervention under limited ODD conditions
Solution Approach 1:
The system performs preliminary identification of yielding target vehicles and prediction of cut-in events before actual lane changes occur. This advance recognition capability enables the autonomous vehicle to plan and execute lane changes proactively under level-4 automation requirements, rather than waiting for collision avoidance scenarios to develop.
Solution Approach 2:
The system continuously monitors lane change intentions of surrounding vehicles and uses this feedback to adjust its own traveling decisions. By analyzing the behavior and intentions of other vehicles in real-time, the autonomous vehicle can make informed autonomous traveling decisions that satisfy level-4 requirements without driver intervention.
Data Source
AI summary
A vehicle running control method includes: when a junction section lane is present adjacent to a traveling lane of a host vehicle, collecting, by a processor, vehicle information of at least one vehicle traveling in the junction section lane; determining, by the processor, the possibility of cut-in of junction section lane vehicles based on the collected vehicle information and whether the traveling lane is congested; and controlling, by the processor, at least one of the traveling path or the traveling velocity of the host vehicle based on the result of determination in order to display an intention to yield.