Steering Noise Localization Using Neural Network Acoustic Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for locating noise in vehicle steering systems are inefficient, relying on expert estimation and costly analysis equipment, leading to potential erroneous repairs and high analysis times.
Innovation Solution
A device using a neural network model with a sound receiving unit, processing unit, and storage unit to detect and diagnose noise in the steering system by converting time-domain data into frequency-domain data, applying techniques like Mel-Frequency Cepstrum Coefficient, and utilizing a convolution neural network to locate the noise source.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If expert estimation and manual sensor mounting methods are used to locate noise, then noise position can be identified, but the process requires excessive time and high costs
Solution Approach 1:
The patent applies preliminary action by pre-training a neural network model with noise data collected from various steering system components before actual noise diagnosis. The model learns to identify noise sources through offline training, enabling rapid online diagnosis without requiring time-consuming manual analysis or sensor mounting during the actual repair process
Solution Approach 2:
The patent replaces the mechanical system of manual sensor mounting and expert analysis with an automated acoustic analysis system using neural networks. The system automatically processes acoustic signals and identifies noise sources through machine learning, eliminating the need for physical sensor installation and expert intervention
2Measurement precision
If manual noise analysis by experts is performed, then noise position can be estimated, but the analysis cost becomes excessive
Solution Approach 1:
The patent implements self-service by enabling the system to automatically diagnose noise sources without requiring expert intervention. The neural network model independently analyzes acoustic signals and identifies noise positions, making the system self-sufficient and eliminating costs associated with expert labor
Solution Approach 2:
The patent uses cost-effective acoustic sensors and processing systems instead of expensive specialized analysis equipment. The system employs affordable microphones and standard computing hardware running neural network algorithms, replacing costly expert services with inexpensive automated processing
3Reliability
If sensors are mounted on estimated noise positions for analysis, then noise can be detected, but erroneous repairs may occur due to expert variability
Solution Approach 1:
The patent changes the parameter of noise analysis from subjective expert estimation to objective neural network-based acoustic signal processing. By transforming the analysis method from human judgment to machine learning parameter processing, the system achieves consistent and accurate noise position identification without expert variability
Solution Approach 2:
The patent incorporates feedback mechanisms where the neural network model continuously learns from training data and improves its noise identification accuracy. The system uses feedback from training examples to refine its predictions, ensuring high reliability and precision in noise position detection
Data Source
AI summary
A device for locating a noise occurring in a steering system includes: a sound receiving unit detecting noise occurring in a steering system; a processing unit inputting data on the noise in the steering system into a neural network model that performs learning in advance and locating a position or a component at which the noise occurs in the steering system, the noise being detected by the sound receiving unit; and a storage unit in which the neural network model that performs the learning in advance is stored.