Use phone app to know where the car is broken

A new study by the Massachusetts Institute of Technology (MIT) in the US says it has found a way for phones to detect damage in cars.

As cars become smarter, they could automatically alert users when they should take their car in for maintenance, or notify them of a problem with a part of the car.

However, with older cars, which make up a very high percentage of those on the road, it's a bit more difficult.

While it is possible to plug a dongle (diagnostic port) into the vehicle to automatically run diagnostics, this can expose the vehicle to hackers.

But new research by Joshua Siegel and colleagues at MIT says they've discovered how microphones and accelerometers in smartphones can detect vehicle malfunctions.

According to the paper published by the research team in the journalTechnical Applications of Artificial IntelligenceThen the audio data collected from the smartphone can diagnose whether an air filter needs to be replaced.

The idea behind this is pretty simple. When a filter is dirty or clogged, the sound that passes through it will be different than the sound that comes through it when it’s not clogged. And that difference will produce different vibrations.

This is important because a dirty filter will not be able to bring air to cool the engine, resulting in reduced fuel economy and increased wear.

In the near future, we may be able to use smartphones to diagnose car problems. Photo: Popular Mechanics

In their experiments, Siegel and colleagues used an iPhone 6 to measure engine noise from a Mazda 2 and a Honda Civic.

They collected engine sounds from two cars with normal air filters and with the air filters covered with a 2mm layer of dust. They also covered the air filters with a square piece of paper (to simulate a leaf stuck to the filter) and then collected the engine sounds.

After carefully studying these sounds, the team created an algorithm to identify the air filter under normal conditions and the air filter when it is clogged with dust.

Of course, this is just the first step. From detecting an air filter abnormality to understanding the entire vehicle’s components is a long journey. Siegel’s article cites several previous studies that have found ways to use sound to detect intake air leaks, ECT, faulty axle sensors, and more.

Not only that, other studies have shown that algorithms are also capable of detecting wheel imbalances.

Siegel now intends to commercialize this research through a new company he founded, Data Driven. The first commercial product is expected to be released within the next 18 months.

According to TTO/Ars Technica