Brake Noise Spectrogram Classification for Squeal and Test Anomalies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise analyzers struggle to reliably identify and classify various types of noise generated by vehicle braking systems, including false positives and negatives, and fail to distinguish higher-order harmonics and low-frequency noise, while also failing to detect anomalies in the test environment.
Innovation Solution
A method using artificial intelligence (AI) and machine learning (ML) algorithms, specifically neural networks, to analyze digital audio data from braking systems, filters out low-frequency noise, segments spectrograms into intensity bands, and classifies noise events into categories and sub-categories, including squeals, chirps, artifacts, and anomalies, using tagged training datasets to improve detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard noise analyzers use Fourier transform-based spectral processing methods to identify noise events, then tonal noise detection is improved, but detection of other noise types (collision noise, chirp/wirebrush, artifacts) deteriorates
Solution Approach 1:
The patent segments the noise spectrum into multiple frequency bands and applies different analysis methods to each band. The noise analysis is divided into tonal noise detection (using Fourier transform) and non-tonal noise detection (using other methods), allowing each type to be optimized independently without interfering with the other.
Solution Approach 2:
The patent transitions from traditional spectral analysis to time-frequency analysis by introducing spectrograms and wavelet transforms. This adds a time dimension to the frequency analysis, enabling the detection of transient and non-stationary noise events that cannot be captured by standard Fourier transform methods alone.
2Productivity
If noise analyzers rely on amplitude criteria and spectral processing to identify noise events, then detection speed is improved, but false positives and false negatives increase
Solution Approach 1:
The patent implements feedback mechanisms where detected noise events are validated through multiple analysis stages. The system cross-references detections from different methods (Fourier transform, wavelet transform, spectrogram analysis) and uses confidence scoring to reduce false positives while maintaining detection speed through automated validation pipelines.
Solution Approach 2:
The patent performs preliminary noise characterization and classification before final detection confirmation. By pre-processing the audio data with multiple transformation methods and pre-identifying potential noise patterns, the system reduces the computational burden during final detection and minimizes false positives through early filtering.
3Measurement precision
If engineers manually classify noise events based on spectrograms, then classification accuracy is improved, but analysis time increases
Solution Approach 1:
The patent replaces manual visual inspection of spectrograms with automated machine learning algorithms. These algorithms process spectrogram data and classify noise events automatically, achieving accuracy comparable to or exceeding human experts while reducing analysis time significantly. The system uses trained models to recognize patterns that would be difficult for humans to identify consistently.
Solution Approach 2:
The patent creates digital representations (spectrograms, wavelet coefficients) of the audio noise data that can be processed and analyzed without requiring physical manual inspection. These digital copies enable automated analysis while preserving all the information needed for accurate classification, eliminating the time-consuming manual review process.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
A method for identifying and characterizing noises generated by a vehicle braking system is described. The method first comprises the steps of detecting noises generated by a vehicle braking system under dynamic operating conditions and generating digital audio data representative of the detected noise. The method then provides analyzing the aforesaid digital audio data by a noise analyzer, to identify potential squeal events and respective likely squeal frequencies, and generating squeal frequency information indicative of the squeal frequencies of the identified potential squeal events. The method then comprises the steps of filtering the aforesaid digital audio data by means of high-pass filtering to eliminate spectral components at frequencies lower than a filtering frequency, to generate filtered digital audio data; and generating, based on the filtered digital audio data, a respective spectrogram, which represents, in graphical form, information present in the filtered digital audio data, comprising the sound signal intensity, as a function of time and frequency. The method then involves providing the aforesaid spectrogram and the aforesaid squeal frequency information to a trained algorithm, wherein the algorithm was trained using artificial intelligence and/or machine learning techniques. The method also provides identifying noise events, by the trained algorithm, based on the above spectrogram and squeal frequency information, classifying the identified noise events and finally providing information about the identified noise events, each characterized by the respective category. The aforesaid classification step involves a classification according to at least the following categories: a first category comprising noises to be detected generated by the characteristic dynamic operation of the braking system; and a second category comprising abnormal noises, generated by operational or test anomalies.