Vehicle Intersection Yield Prediction via ML Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle systems lack an efficient method to determine the correct order of yielding at intersections, particularly in complex scenarios like four-way intersections, which can lead to potential collisions when multiple vehicles attempt to proceed simultaneously.
Innovation Solution
A computer system equipped with a machine learning program, utilizing image data from sensors to predict the number and order of targets (vehicles, pedestrians, etc.) that a host vehicle should yield to, based on time differences and distance analysis, and communicates this information through alphanumeric messages to ensure safe passage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If vehicles proceed through the intersection in the order they arrived, then the operation is simple and straightforward, but collisions may occur in complex scenarios like four-way intersections where multiple vehicles attempt to proceed simultaneously
Solution Approach 1:
The system performs preliminary analysis of the intersection environment using image data from sensors before the host vehicle reaches the intersection. The machine learning program predicts the number and order of targets to yield to in advance, allowing the vehicle to prepare appropriate yield behavior beforehand, thus preventing collisions while maintaining systematic control
Solution Approach 2:
The patent introduces an intermediary communication system that transmits alphanumeric messages between vehicles at the intersection. This message passing mechanism serves as a mediator to coordinate right-of-way, allowing vehicles to negotiate and establish a safe passage order without direct collision, thus resolving the complexity of multi-vehicle coordination
2Reliability
If a machine learning program is used to predict yielding order, then collision risk is reduced, but the system complexity and computational requirements increase
Solution Approach 1:
The machine learning program is implemented within the host vehicle's own computer system, allowing each vehicle to independently perform its own yield prediction and decision-making. This self-service approach eliminates the need for centralized control infrastructure, reducing overall system complexity while maintaining high prediction accuracy through local image analysis and machine learning inference
Data Source
AI summary
A computer includes a processor and a memory, the memory storing instructions executable by the processor to collect a plurality of images of one or more targets at an intersection, input the images to a machine learning program to determine a number of the targets to which a host vehicle is predicted to yield at the intersection based on time differences between the plurality of images, and transmit a message indicating the number of the targets.