Engineers from MIT have wound up developing an app which uses sensors in a smartphone to figure out why someone's car might be making a funny noise.
You might take pride in your power to diagnose an engine problem that's developing just by sound itself. You might even amuse and/or annoy colleagues, friends, and family by replicating that mystery sound to them, Click and Clack style.
ClingClanger is the name of a new app for smartphones that is designed for everyone else, the folks that can't immediately know the sound of a piston slap or get surprised by the clicking noises of a battery that's dying. Most drivers have a natural impulse to just crank up their radio and then hope the problem or noise goes away. Typically though, they don't, which is just how a small problem can turn into a big one.
You should be forewarned that some particular audio clips are especially grating, such as the failing water pump or the worn-out brakes. Others are also very disturbing, such as 'engine seizing up' or an 'engine running without oil'. You have to remember these were required by torturing cars that likely died soon thereafter.
The ClingClanger app theoretically can use the microphone of your phone to actually listen to your ailing vehicle and then compare the sounded cries to entries in its sound library before providing a tentative diagnosis.
You can put Cling Clanger on your Android or iPhone immediately for just $1.99. The sound-recognition version is going to cost more once it's released. We looked up and down through Google's app store and found nothing comparable at this price point or any other, so it does seem like a form of cheap insurance for those not mechanically inclined. It might even be a powerful learning guide.
Surely there's been a time you were cruising down the road when your vehicle started making a sound like 'plink, plink'. Or was it more like 'pring, pring'? Could it be the gearbox? Or is the clutch? Is your alternator the culprit behind this mystery? Your air conditioner? Will your engine fail you and leap out of your car? Most drivers have no idea what's going on in their cars. They rely on mechanics and manuals to let them know when something needs to be replaced or fixed.
However, what if you could use your phone to diagnose the car on your own? Researchers at MIT are working on an app they believe can analyze the noises a car makes, as well as its vibrations, and then tell you the difference between imbalanced wheels or a clogged air filter.
This app should be able to empower daily drivers to diagnose their own problems, instead of calling Click and Clack, the funny hosts of NPR's long-running Car Talk, who were famous for diagnosing car issues just by listening to telephone callers imitating the strange noises their cars made at the time.
The app uses parts of the smartphones it is installed on, including GPS, accelerometers, and microphones. The microphone can 'hear' a clogged air filter's whistling sound. A GPS system can combine vibration data with vehicle speed to determine if the tires are inflated properly.
This app uses what is called machine learning to discover the vibrations and sounds that signify particular problems or issues. The team claims it has better than 90 percent accuracy.
Siegel and his colleagues developed the app by renting many different types of cars and 'breaking' them temporarily to deliberately induce the very kinds of problems they were hoping to study. Before returning the cars, they'd restore them to premium condition with oil changes and tire rotations.
The journal titled Engineering Applications of Artificial Intelligence recently published a paper about their work.
The current app is limited to particular and common problems that smartphone sensors might detect easily, such as clogged air filters, improper tire pressure, engine misfires, and wheel imbalances.
Don't fire your neighborhood mechanic just yet. Many issues need fault tracing that is far more nuanced, or they might have intermittent occurrences. Worse yet for the app, their vibration pattern might not be repeated or characteristic to match the database.
A number of mechanics have skepticism about just how much can be done using a smartphone app.
Experienced mechanics are still necessary given that one make or model of car might make a different noise for the same problem that many other cars do a different noise for.
For instance, if a human mechanic is trying to diagnose the air conditioning and its noise, he might sit in the driver's seat first, then the passenger seat, then stick his head underneath the dashboard, and then play with all the climate control settings to check for changes in the noise.
An app microphone can detect such changes, but there still needs to be a human that knows how to fix things. Personally, I believe there's a broad divide between how vehicles get diagnosed in real-world shops and an academic laboratory.
Still, many mechanics might be eager to embrace newer technology, and apps might wind up playing a large diagnostic role in the future.
Still, I'm not sure we're close to that yet.
On the other hand, Siegel and the team around him think their app might save drivers over a hundred bucks a year, and that's something. Also, truckers might save six times that much, and that's notable if accurate. It might also help drivers save money on gas by ensuring their vehicles are operating efficiently, or just avoid breakdowns and blowouts.
Siegel decided to found a startup that he named Data Driven which he intends to use to bring the idea to the market. At the time of writing, it was thought that field testing with a prototype app could start in six months or so, followed by a full commercial version roughly a year after that. Don't be shocked if a mobile mechanic Houston professional you call winds up using such technology to accurately diagnose your car far away from the environs of a garage.
Terrence Clarke is a person who loves everything about cars - from tires to keys. As a matter of fact, he has this collection of different vintage cars right on his garage. In his spare time, he shares his fascination in the online community by writing.