Vehicle Follow-Up Control for Predicted Cut-In Maneuvers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle autonomous driving systems fail to accurately predict surrounding vehicles' lane changes, leading to potential collisions or unnecessary braking due to misjudgment or sensitivity in detecting cut-in maneuvers.
Innovation Solution
A vehicle control device and method that utilizes a receiver for surrounding vehicle detection, a state predictor using a cut-in probability calculation model (Gaussian Mixture Model and Hidden Markov Model) to generate cut-in state information, and a controller to calculate required acceleration for a follow-up target based on this information, thereby reducing risk and driver discomfort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the autonomous driving system sensitively predicts the cut-in operation of the surrounding vehicle, then the response time to actual cut-in is improved, but the driver may step on the brake unnecessarily even when the surrounding vehicle does not actually cut in
Solution Approach 1:
The system dynamically adjusts the follow-up target selection and required acceleration calculation based on the predicted cut-in state. When a cut-in is predicted, the system dynamically switches the follow-up target from the surrounding vehicle to the preceding vehicle, and adjusts acceleration requirements to prevent unnecessary braking while maintaining collision avoidance.
Solution Approach 2:
The system changes key parameters (follow-up target selection and required acceleration) based on the cut-in prediction state. By modifying these parameters dynamically, the system resolves the contradiction between sensitive detection and false alarm response, enabling appropriate braking only when necessary.
2Object-affected harmful factors
If the host vehicle fails to respond to the cut-in surrounding vehicle at an appropriate time, then the driver may be put at risk of collision, but if the system responds too early or too sensitively, unnecessary braking occurs
Solution Approach 1:
The system uses feedback from the cut-in probability calculation model to continuously monitor the surrounding vehicle's lane change intent. This feedback mechanism allows the system to respond at the appropriate time by adjusting the follow-up target and acceleration requirements based on actual cut-in probability, rather than using fixed threshold responses.
Solution Approach 2:
The system performs preliminary action by pre-calculating alternative follow-up targets and acceleration requirements based on predicted cut-in states. When cut-in is predicted, the system has already prepared the alternative follow-up target (preceding vehicle) and adjusted acceleration parameters, enabling smooth transition without sudden braking.
3Device complexity
If the autonomous driving system uses a single follow-up target for vehicle control, then the control logic is simple, but it cannot accurately handle situations where surrounding vehicles change lanes
Solution Approach 1:
The system segments the follow-up target selection into different scenarios: normal following mode and cut-in prediction mode. By dividing the control logic into these segments, the system maintains simplicity in each mode while achieving high accuracy in cut-in detection through the state predictor and probability calculation model.
Data Source
AI summary
The disclosure relates to a technology regarding a vehicle control device and method, in a vehicle follow-up control context, comprising receiving surrounding vehicle detection information received from a sensor of a host vehicle, generating cut-in state information for separately predicting a cut-in state using a cut-in probability calculation model based on the surrounding vehicle detection information, and calculating a required acceleration for a follow-up target of the host vehicle selected according to the cut-in state information.