Vehicle Collision Detection Using Acceleration Peak Time Intervals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing solutions fail to adequately and easily identify minor collisions between vehicles and obstacles, often confusing them with accelerations not causing vehicle damage, and require extensive computational resources and large amounts of input data.
Innovation Solution
A computer-implemented method using a machine learning classification model that collects sensor data on acceleration values, delimits time intervals around peak acceleration, and processes this data to distinguish between minor collisions and other accelerations using time-series statistical data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning algorithms use large amounts of input data and extensive computational resources to detect collisions, then detection accuracy improves, but system complexity and computational burden increase
Solution Approach 1:
The patent extracts and focuses on the most critical features for collision detection: acceleration peaks and their immediate temporal context. By delimiting a first time interval around peak acceleration and comparing it with adjacent second and third time intervals, the system isolates only the essential data needed for detection, eliminating the need to process large amounts of redundant sensor data while maintaining high detection accuracy
Solution Approach 2:
The patent segments the sensor data processing into distinct temporal intervals: a first time interval containing the peak acceleration, a second time interval prior to it, and a third time interval subsequent to it. This segmentation allows the machine learning model to process structured, divided data segments rather than continuous large datasets, reducing computational complexity while preserving detection accuracy
2Reliability
If the system processes extensive sensor data to distinguish minor collisions from other accelerations, then detection reliability improves, but processing time and computational resources increase
Solution Approach 1:
The patent applies preliminary action by pre-defining the temporal structure of data processing around acceleration peaks. The system establishes fixed time intervals (first, second, and third intervals) before actual collision detection occurs, preparing the data framework in advance. This allows the machine learning model to quickly process pre-structured data segments during real-time operation, improving both reliability and reducing processing time
Solution Approach 2:
The patent applies partial action by processing only the necessary temporal portions of sensor data - specifically the peak acceleration interval and its immediate adjacent intervals - rather than analyzing the entire sensor data stream. This selective partial processing maintains high detection reliability by focusing on critical periods while significantly reducing overall processing time and computational resource requirements
Data Source
Figure 1~2
AI summary
A computer-implemented method for detecting a collision between at least one obstacle and a moving vehicle (1) by a machine learning classification model, the method comprising: a) collecting sensor data, comprising acceleration values, from at least one on-vehicle sensor (18) at a plurality of corresponding timings during motion of the moving vehicle; b) delimiting, when a peak acceleration value among said acceleration values is greater than a predetermined threshold, a first time interval comprising the corresponding timing of the peak acceleration value; c) extracting sets of sensor data collected during the first, a second and a third time interval; d) processing the sets of sensor data to extract a first, second and third time-series statistical data; and e) applying said first, second and third time-series statistical data s to the machine learning classification model.