Traffic Conflict Risk Quantification via Trajectory Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for classifying and quantifying traffic conflicts are inadequate, as they rely on assumptions about driver behavior and fail to accurately assess the risk of imminent collisions, particularly due to variations in human perception and the reliance on assumed braking actions.
Innovation Solution
The system analyzes live or streaming traffic video to detect and quantify traffic conflicts by tracking vehicle trajectories within a spatial-temporal domain, calculating collision risk based on minimum separation and relative velocities, without assuming driver behavior, and transmits this risk to computing devices for real-time alerts or control adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If human assessment is used to evaluate traffic conflict, then the process is simple and intuitive, but the accuracy is poor due to variations in human perception
Solution Approach 1:
The patent replaces human visual assessment with an automated image processing system that captures traffic scenes, extracts vehicle positions and trajectories, and calculates collision risk parameters objectively. This substitution eliminates human perception variations while maintaining operational simplicity through automated analysis.
Solution Approach 2:
The patent introduces an intermediate computational system that processes image data and calculates collision risk parameters (such as time-to-collision and impact speed) based on extracted vehicle trajectories. This intermediary transforms raw visual information into quantifiable risk metrics, bridging the gap between simple observation and precise measurement.
2Device complexity
If assumed driver behaviors are used to determine conflict severity, then the computation is simplified, but the accuracy deteriorates due to unrealistic assumptions
Solution Approach 1:
The patent extracts actual driver behavior data directly from image sequences by tracking vehicle positions and velocities over time. Instead of assuming braking actions, the system measures real deceleration patterns from the extracted trajectories, eliminating unrealistic assumptions while keeping the computational model manageable.
Solution Approach 2:
The system uses the image data itself to provide information about driver behavior. By analyzing vehicle trajectories directly from captured images, the system self-determines actual braking actions and speed changes without requiring external assumptions or additional sensors, thereby improving accuracy without proportionally increasing complexity.
Data Source
AI summary
Systems and methods for quantitatively assessing collision risk of bodies, such as vehicles, and severity of a conflict between the bodies are disclosed. A method may comprise receiving image data associated with bodies. Based on the image data, an affinity (proneness) to collision of the bodies may be determined. Based on the determined affinity (proneness) to collision, a proximity (closeness) of collision of the bodies may be determined. Based on the determined proximity (closeness) of collision of the bodies, a collision risk of the bodies may be determined. The determined collision risk may be transmitted to a computing device, such as via a user interface. The determined collision risk may be used to control operations of a traffic control device or an autonomous vehicle.