Convolutional Neural Network for Vehicle Crash Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle crash detection technologies lack accuracy in distinguishing between crash and near-crash events, which can lead to delayed alerts and inadequate safety measures.
Innovation Solution
A method and apparatus utilizing a trained convolutional neural network to analyze state information from vehicle sensors, including GPS, accelerometers, and gyroscopes, to determine event types such as crash, near-crash, and baseline events, enabling timely alerts and evading operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional threshold-based methods are used to detect crashes, then the device complexity is low, but the measurement precision of event type classification is insufficient
Solution Approach 1:
The patent replaces traditional mechanical threshold-based detection systems with a deep learning-based convolutional neural network system. The CNN model processes sensor data to classify events into crash, near-crash, and baseline categories, achieving high measurement precision (96.8% accuracy) while substituting complex rule-based logic with a trained neural network that automatically learns optimal detection patterns from data.
2Reliability
If simple threshold detection is used, then the ease of operation is high, but the reliability of crash detection is insufficient
Solution Approach 1:
The convolutional neural network performs self-service by automatically learning and optimizing its own detection parameters from training data. The system self-adjusts to distinguish between crash and near-crash events without requiring manual threshold tuning or complex operational intervention, thereby improving reliability while maintaining ease of operation through automated adaptive detection.
3Productivity
If traditional sensor analysis methods are used, then the loss of time in processing is low, but the productivity of safety response is insufficient
Solution Approach 1:
The system performs preliminary action by pre-training the convolutional neural network model with extensive crash and near-crash data before deployment. This pre-training enables the system to rapidly classify new events in real-time without requiring complex processing during actual detection, thereby improving safety response productivity while minimizing processing time loss through optimized inference.
Data Source
AI summary
The present disclosure relates to method, storage medium and electronic device for detecting vehicle crashes. The method comprises: acquiring state information of a target vehicle; and determining an event type of the target vehicle according to the state information and a trained convolutional neural network, the event type being any of the following types: a crash event, a near crash event and a baseline event. The event type of the vehicle is determined using the trained convolutional neural network in the present disclosure, so that the accuracy is high; and near crash events can be detected, thus, when a near crash event is detected, the driver can be further alerted or an evading operation is directly performed on the vehicle, so that the safety is improved and the safety of the driver and passengers is guaranteed.