MIT is working on a self-diagnosis system for machines

In the next couple of years, MIT engineers plan to finish work on the car's auto-diagnostic system. Caught in a car equipped with such a system, the driver will receive a report in the mobile application about the condition of the machine, notifications of necessary repairs, mileage data and breakage history.

According to the creators, the system will allow car owners to save an average of $ 125 a year due to the timely replacement of worn and defective parts; for trucks, this figure is $ 600 per year. The calculation did not take the lost profit in case of sudden car breakdowns.

Testing the prototype showed that to analyze the information coming from the sensors, enough power of the smartphone processor, and only for some types of diagnostics you will need a computer, naked-science.ru  reports.The accuracy of the diagnosis exceeded 90%; In addition, the system has passed a false positive test, never having stated a non-existent problem.

The prototype distinguishes between the sound and vibration of new tires and the sound of tires that it's time to change, warns of the need to replace the air filter, the plugs and the oil filter. The machine learning algorithm compares the data with the database and diagnoses it. To test the system, the engineers took cars for hire, brought them to the right emergency condition, diagnosed it, and then repaired the car and returned it.

To bring the system to market, engineers are assembled through the Data Driven startup, founded by one of the authors of the development, Joshua Siegel. About the development tells the site of the Massachusetts Institute of Technology.

© All rights reserved. Citing to www.ict.az is necessary upon using news