Embedded Acceleration Sensors for Internal Pavement Damage Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pavement damage monitoring methods primarily focus on surface damage and fail to accurately detect internal damages such as voids beneath the road slab, leading to low accuracy in damage recognition.
Innovation Solution
A method involving convolutional neural networks and multi-layer perceptrons to extract acceleration features from internal sensors, combined with working condition data, to determine pavement damage, including internal damage detection through feature splicing and model training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional surface monitoring methods are used, then the monitoring system is simple, but the accuracy of damage recognition is low
Solution Approach 1:
The monitoring system is segmented into multiple acceleration sensors distributed at different positions within the pavement structure, each capturing local vibration characteristics. This segmentation enables comprehensive coverage of internal damage detection while maintaining manageable system complexity through modular sensor placement and independent data collection units.
Solution Approach 2:
A convolutional neural network serves as an intermediary between the acceleration sensors and the final damage classification. The CNN processes raw acceleration data, extracts meaningful features, and transforms them into damage probability distributions, bridging the gap between simple sensor measurements and accurate damage recognition without requiring complex direct measurement systems.
2Reliability
If internal acceleration sensors are deployed, then internal damage detection capability is improved, but the device complexity increases
Solution Approach 1:
The acceleration sensors are designed with multi-functionality, serving both as vibration measurement devices and as structural health monitoring probes. The same sensors detect various types of internal damages (voids, cracks, delamination) across different pavement layers, eliminating the need for separate specialized sensors for each detection purpose and reducing overall system complexity.
Solution Approach 2:
The system transitions from two-dimensional surface monitoring to three-dimensional internal monitoring by placing acceleration sensors at multiple depths and positions within the pavement structure. This dimensional expansion enables detection of internal damages that were previously inaccessible, improving reliability without proportionally increasing complexity through systematic spatial distribution of sensors.
3Adaptability or versatility
If convolutional neural networks are used for feature extraction, then classification generalization ability is improved, but the computational complexity increases
Solution Approach 1:
The convolutional neural network performs preliminary feature extraction and transformation during the data collection and training phases. By pre-processing the acceleration data through CNN convolutions and pooling operations, the system prepares simplified feature representations that capture essential damage patterns, enabling accurate classification without requiring complex real-time computational processes during actual damage assessment.
Data Source
AI summary
A pavement condition determining method, a collection box, a computer device, a storage medium, and a computer program product. The pavement condition determining method comprises: acquiring acceleration data and working condition data, the acceleration data being collected by a plurality of acceleration sensors arranged inside the pavement; performing a feature extraction on the acceleration data according to a convolutional neural network of a pavement condition recognition model to obtain acceleration features; splicing each of the acceleration features with corresponding working condition data to determine a target feature vector; and determining a pavement damage identification result based on target feature vectors and a multi-layer perceptron of the pavement condition recognition model.