AI ADAS Audio Testing for Noise Filtering and Tone Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional ADAS testing is manual and labor-intensive, prone to human error, and lacks accuracy in tone measurements and priority-based testing.
Innovation Solution
An AI-based ADAS testing tool using pre-trained noise cancellation and audio classification ML models to filter noise, identify ADAS audio signals, and classify them into frequency classes for accurate testing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If manual testing is used with pico-scopes and oscilloscopes, then testing can be performed with simple equipment, but the process is time-intensive and prone to human error
Solution Approach 1:
The system performs self-testing by automatically capturing audio signals from the vehicle's audio output, processing them through ML models for noise cancellation and classification, and generating test results without requiring manual operation of testing equipment. The processor autonomously executes the entire testing workflow from signal capture to result generation.
Solution Approach 2:
The patent replaces manual mechanical testing operations with automated electronic processing. Instead of manually operating pico-scopes and oscilloscopes, the system uses a processor to capture, filter, and analyze audio signals electronically, substituting human-operated mechanical measurement tools with automated computational systems.
2Ease of operation
If manual testing is performed, then human intervention allows flexible testing, but the process is labor-intensive and time-consuming
Solution Approach 1:
The system enables continuous automated testing by constantly capturing audio signals from the vehicle's audio output and continuously processing them through the ML-based noise cancellation and classification pipeline. Multiple frequency classes can be tested in sequence without interruption, eliminating the time loss associated with manual setup and measurement for each test case.
Solution Approach 2:
The system performs preliminary classification of audio signals into frequency classes before detailed analysis. The ML model pre-processes incoming audio signals to identify and categorize them, allowing the testing system to efficiently route and process different frequency classes through appropriate test protocols without manual intervention.
3Device complexity
If conventional measurement tools are used, then equipment is simple and accessible, but accuracy in tone measurements and multiple tone testing is insufficient
Solution Approach 1:
The patent introduces ML-based noise cancellation and audio classification models as intermediary processing layers between the audio signal source and the measurement analysis. These intermediary models filter and categorize audio signals before final measurement, enabling accurate distinction between multiple simultaneous tones and improving measurement precision beyond what conventional tools can achieve alone.
Solution Approach 2:
The system transforms the audio measurement approach by changing from direct frequency measurement to ML-based classification of frequency classes. Instead of manually measuring individual tones, the system uses ML models to classify audio signals into distinct frequency categories, enabling accurate measurement of multiple simultaneous tones through parameter transformation rather than direct measurement.
4Productivity
If automated testing is implemented, then testing efficiency improves, but system complexity increases
Solution Approach 1:
The processor serves multiple functions within a single integrated system: it captures audio signals from the vehicle, executes noise cancellation ML models, runs audio classification ML models, performs frequency class identification, and generates test results. This multi-functional approach achieves high testing efficiency without requiring separate dedicated hardware for each function, thereby limiting the increase in overall system complexity.
Data Source
AI summary
The present invention describes an Artificial Intelligence (AI) based Advanced Driver Assistance System (ADAS) testing tool. The testing tool comprises a processing unit configured to filter out, using a pre-trained noise cancellation ML model, the received audio signals to isolate noise signals from the received audio signals. Sequentially, the processing unit is configured to, using a using a pre-trained audio classification ML model, identify one or more audio signal from the filtered-out audio signals relating to ADAS audio signals and classify each of the one or more identified signals into a frequency class based on their distinct frequency characteristics. Finally, the processing unit is configured to evaluate each of the classified signal based on ADAS actions associated with the vehicle and generate test results of ADAS testing on each of the classified audio signals, based on the evaluation.