Underride Impact Detection Using Camera Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle safety mechanisms are inadequate in detecting and mitigating underride impacts, where a smaller vehicle crashes into a larger one, with the bumper not making initial contact, leading to insufficient activation of safety measures due to lack of protective elements or their ineffectiveness.
Innovation Solution
A system equipped with a camera and sensors to classify target vehicles, determine open spaces, and predict underride impacts, triggering appropriate safety actions such as alerts and modifying the collision system to deploy airbags and seatbelts proactively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional crash sensors are used to detect impacts, then the system is simple and reliable for standard collisions, but it fails to detect underride impacts where the bumper does not make initial contact
Solution Approach 1:
The system performs preliminary classification of target vehicles and prediction of underride impact risk before the actual collision occurs. By using camera-based vehicle classification and open space detection, the system identifies high-risk scenarios in advance, enabling proactive safety interventions rather than reactive sensor-based detection.
Solution Approach 2:
The patent introduces an intermediary processing system that bridges the gap between simple sensor detection and complex AI analysis. The system uses a camera to capture images, processes them through neural networks to classify vehicles and predict underride risks, and then triggers appropriate safety responses. This intermediary layer enables detection of underride impacts without requiring direct bumper contact sensors.
2Reliability
If the system activates safety measures proactively based on prediction, then safety performance is improved, but false activations may occur reducing system reliability
Solution Approach 1:
The system incorporates feedback mechanisms where the neural network continuously monitors multiple parameters including vehicle classification, open space dimensions, and predicted underride risk. The safety measures are activated only when the predicted risk exceeds predetermined thresholds, providing a feedback-based decision-making process that reduces false activations while maintaining high safety performance.
Solution Approach 2:
The patent utilizes parameter changes in the form of adjusting detection thresholds and classification criteria based on different vehicle types and scenarios. By dynamically changing the parameters for what constitutes a high-risk underride situation, the system optimizes the balance between proactive safety activation and minimizing false positives across diverse operating conditions.
Data Source
AI summary
Devices and methods are disclosed for detecting or predicting an underride impact. An example vehicle includes a camera for capturing an image of a target vehicle, sensors for detecting a speed and heading of the vehicle, and a processor. The processor is configured to determine a target vehicle classification, determine an open space of the target vehicle, determine a closing speed between the vehicle and target vehicle, predict an underride impact based on the classification, open space, and closing speed, and responsively execute an impact action.