Acoustic health usage and monitoring system of mechanical automotive subsystems

1. A computer-implemented method for vehicle mechanical component wear detection, comprising:

receiving, by at least one processor, acoustic signals of a vehicle;

processing, by the at least one processor, the acoustic signals to generate training data, the processing comprising:

filtering, by band-pass filters, the acoustic signals;

comparing, based on a spectral density, the filtered acoustic signals to known filtered acoustic signals to identify altered acoustic signals; and

labeling the altered acoustic signals with a corresponding vehicle mechanical component wear level;

training, based on the training data, a predictive learning algorithm to generate a predictive vehicle mechanical component wear model;

classifying, by the predictive vehicle mechanical component wear model, the processed acoustic signals to indicate wear levels of one or more vehicle mechanical components; and

wherein at least one of the receiving, processing, training and classifying are performed by one or more computers to predict the corresponding vehicle mechanical component wear levels.

2. The computer-implemented method of claim 1 , the receiving further comprising:

capturing the acoustic signals from any of pre-stored acoustic data, microphones, acceleration sensors or Micro ElectroMechanical System (MEMS) devices.

3. The computer-implemented method of claim 1 , the receiving further comprising:

continuously monitoring live acoustic signals generated by a vehicle frame or vehicle sub-frame of a vehicle suspension system.

4. The computer-implemented method of claim 3 , wherein the live acoustic signals comprise:

ball joint acoustic signatures.

5. The computer-implemented method of claim 4 , wherein the ball joint acoustic signatures comprise:

ball joint acoustic signatures altered by physical damage.

6. The computer-implemented method of claim 1 , the comparing the filtered acoustic signals further comprising:

comparing the spectral density of each of a plurality of frequency ranges to the known filtered acoustic signals in a same frequency range.

7. The computer-implemented method of claim 1 , the training further comprising:

tuning the predictive vehicle mechanical component wear model based on evaluation metrics.

8. The computer-implemented method of claim 7 , wherein the evaluation metrics include any of:

accuracy, precision, recall, or area under a receiver operating characteristics (ROC) curve.

9. The computer-implemented method of claim 1 , further comprising:

receiving, by a classifier, newly captured acoustic signals of another vehicle mechanical component and predicting, based on the predictive vehicle mechanical component wear model, the wear levels within a range of values.

10. The computer-implemented method of claim 1 , wherein the training data comprises:

a supervised training data set with a sound signal as an input vector and level of wear as an output teacher vector.

11. A system, comprising:

a memory; and

at least one processor coupled to the memory and configured to:

receive acoustic signals of vehicle mechanical components;

process the acoustic signals to generate an acoustic signature, the processing comprising:

filtering, by band-pass filters, the acoustic signals;

comparing, based on a spectral density, the filtered acoustic signals to known filtered acoustic signals to identify altered acoustic signals as the acoustic signature; and

labelling the altered acoustic signals with a corresponding vehicle mechanical component wear level;

train a predictive machine learning wear model with the acoustic signature;

process live vehicle mechanical component acoustic signatures with the predictive machine learning wear model to infer wear levels of one or more of the vehicle mechanical components; and

communicate the wear levels to a cloud-based repository.

12. The system of claim 11 , wherein the at least one processor is configured to receive the acoustic signals from any of pre-stored acoustic data, microphones, acceleration sensors or Micro ElectroMechanical System (MEMS) devices.

13. The system of claim 11 , wherein the at least one processor is further configured to:

continuously monitor the live vehicle mechanical component acoustic signals as generated by a vehicle frame or vehicle sub-frame of a vehicle suspension system.

14. The system of claim 13 , wherein the monitored acoustic signals of the vehicle suspension system originate from a driverless vehicle.

15. The system of claim 11 , wherein the acoustic signals comprise:

ball joint acoustic signatures.

16. The system of claim 15 , wherein the ball joint acoustic signatures comprise:

ball joint acoustic signatures altered by physical damage.

17. The system of claim 11 , wherein the at least one processor is further configured to communicate only wear levels above a failure threshold.

18. A non-transitory computer-readable device having instructions stored thereon that, when executed by at least one computing device, cause the at least one computing device to perform operations comprising:

receiving acoustic signals of vehicle mechanical components;

processing the acoustic signals to generate an acoustic signature, the processing comprising:

filtering, by band-pass filters, the acoustic signals;

comparing, based on a spectral density, the filtered acoustic signals to known filtered acoustic signals to identify altered acoustic signals as the acoustic signature; and

labelling the altered acoustic signals with a corresponding vehicle mechanical component wear level;

training a predictive machine learning wear model with the acoustic signature;

processing live vehicle mechanical component acoustic signatures with the predictive machine learning wear model to infer wear levels of one or more of the vehicle mechanical components; and

communicating the wear levels to a cloud-based repository.