Driver Assistance Control for Multi-Trajectory Collision Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional driver assistance systems can only control vehicle movement based on a single predicted trajectory of a selected object, failing to account for multiple potential trajectories of surrounding objects, which limits their ability to anticipate and avoid complex traffic scenarios.
Innovation Solution
A driver assistance apparatus equipped with a camera and controller that processes image data to identify multiple trajectories of objects around the vehicle, selects a predicted trajectory based on distance and time of intersection, and generates a target trajectory for the vehicle to avoid collisions, using coordinate systems to convert and predict object positions and vehicle paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the driver assistance system controls vehicle movement based on only one predicted trajectory of a selected object, then the control system remains simple and computationally efficient, but the system fails to account for multiple potential trajectories of surrounding objects, limiting its ability to anticipate and avoid complex traffic scenarios
Solution Approach 1:
The patent segments the trajectory prediction task by dividing multiple potential trajectories into distinct segments or categories. The system identifies and predicts multiple separate trajectories (first trajectory, second trajectory, etc.) for surrounding objects, treating each as an independent segment that can be analyzed and responded to separately, thereby improving collision avoidance without overwhelming the control system with undifferentiated complexity
Solution Approach 2:
The patent adds a temporal dimension to trajectory prediction by considering not just spatial position but also time-based progression of multiple trajectories. The system evaluates objects at different positions along their respective trajectories at different time points, transforming the problem from a single static prediction to multiple dynamic predictions across time, enhancing reliability while managing complexity through structured temporal analysis
2Reliability
If the system identifies and predicts multiple trajectories of surrounding objects, then the system can anticipate complex traffic scenarios and improve safety, but the computational complexity and processing requirements increase significantly
Solution Approach 1:
The patent applies preliminary action by pre-identifying and categorizing multiple potential trajectories before full processing occurs. The system预先 (in advance) establishes the framework for multiple trajectory prediction, identifies objects that may have multiple trajectories, and prepares the computational structure beforehand, reducing the real-time processing burden while maintaining the ability to anticipate complex scenarios
Solution Approach 2:
The patent applies local quality by focusing computational resources on specific local areas or regions where multiple trajectories are most likely to occur. Rather than uniformly processing all objects and trajectories equally, the system concentrates processing power on critical zones or high-risk scenarios, improving efficiency while maintaining reliability where it matters most
Data Source
AI summary
Disclosed herein is a driver assistance apparatus including a camera installed on a vehicle, having a field of view around the vehicle, and configured to acquire image data and a controller configured to process the image data. The controller may acquire center lines of lanes connected to a lane on which the vehicle travels, identify an object positioned around the vehicle based on the processed image data, identify a plurality of trajectories along which the object is movable based on a position of the object and the center lines, select a predicted trajectory among the plurality of trajectories, and generate a target trajectory of the vehicle avoiding the predicted trajectory.