Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Ford brakes the mould with car maintenance prediction algorithm

Margi Murphy | June 23, 2015
Ford has developed an algorithm using readily available car data to predict when its brakes need maintaining – and will send a text to the driver to let them know.

Ford drive on demand carsharing CES
Ford has a range of new tech like car sharing, car apps and is working on its driverless tech. Image: Ford

Ford has developed an algorithm using readily available car data to predict when its brakes need maintaining - and will send a text to the driver to let them know.

The manufacturer has begun trialling brake maintenance alerts as part of a wider 'conditioned based maintenance project' by closely monitoring 25 different vehicle signals in a prototype fleet of 25 Ford Transits. Using an algorithm developed internally it said it can send drivers more accurate notifications than current brake-replacement alerts based on mileage alone.

Improving car safety is high on the agenda for car manufacturers, with brands like Volvo and Jaguar Land Rover embedding prototypes with sensors and LiDARs to detect obstacles and potholes in the road and communicating that to other cars and local authorities to improve maintenance.

But Ford's new technology is not reliant on sensors but accurate analysis of a car's duty cycle in order to predict problems in enough time for a driver to take a car to be serviced.

Vasilly Krivtsov, technical specialist on the project told Techworld: "The idea is to now collect the data and analyse the data statistically in order to evaluate the needs for maintenance for every single vehicle. For example, depending on how you drive, we would recommend how soon you change your brake pads. If you imagine the difference between a taxi driver in contrast to a retired citizen driver - we can recommend personalised oil change notifications for two very different types of drivers."

It can crunch data relating to how heavy a driver pushes on the pedals and other traits specific to the driver, to predict car part wear and tear.

"This is more than just embedding specialised sensors that tell you if something has gone wrong. It's like the difference between a heart monitor that tells you that you are having a heart attack, and having a system in place that monitors how you are using your body so it can predict that heart attack", Arun Chopra, a technical specialist on the project told Techworld.

What makes this project innovative is its non-invasive, low cost technology.

"Using the example of heart monitoring again," said Krivtsov, "that would require an operation with electrodes placed in proximity of your heart - it's a very invasive procedure.

"In contrast Ford is not equipping each car with a spatial or heat based laser to observe braking conditions. We are collecting regular driving habits like speed or acceleration which means we don't need external hardware solutions. We are simply taking advantage of current car signals and inferring the car's duty cycle very accurately, relative to each of the components - like the brake pedals or engine, for example."

 

Sign up for MIS Asia eNewsletters.