Deep learning could predict car trouble before you notice it
Ultrasonic microphones can detect aberrations in mechanical sounds and alert the operator before a problem happens.
Trying to replicate the sounds your car is making to the mechanic you're talking to is equal parts frustrating and embarrassing. But neural networks could put your best "worn out wheel bearing" impression into retirement. Israeli company 3DSignals uses deep learning to keep track of sounds, listening for deviations from the norm in musical and mechanical acoustics. According to an interview with IEEE Spectrum, the company is in talks with "leading European automakers" to put its ultrasonic microphones in their cars.
But rather than consumer vehicles, the focus seems to be on applications for autonomous taxis. The idea is that before a problem happens, the operator will know as soon as there's an aberration in the car's normal operating sounds.
"When trained, the 3DSignals deep learning algorithms are able to identify [and] predict specific problems in advance with 98 percent accuracy," IEEE Spectrum reports. The tech is already in place in industrial applications, but 3DSignals' clients have to do some of the heavy lifting in terms of training the detection algorithms.
People are still manually labeling specific sound signatures that indicate problems, so it might be a bit before we see widespread adoption of the system. Given BMW's eye toward the future and Volkswagen's aim to launch self-driving taxis, those might be where it'll show up first.
