Оформите заявку на сайте, мы свяжемся с вами в ближайшее
время и ответим на все интересующие вопросы.
Today, our company has more than 15,000 objects in service (our servers stores the telemetry history of these objects for the last 24 months), 40% of them are equipped with trackers connected to the standard CAN bus. We did a survey of our customers about solution Mielta Auto Predict and we collected data on breakdowns and spare parts purchases for 6000 units of tracked equipment over the past 2 years.
In that way, we performed initial training of the system on a large amount of data, classified the types of vehicles and special equipment by the type of work performed and operating conditions. During a creating the system, we did a training on 2,000 incidents of breakdowns and service operations and 34 billion telematics data.
We estimate the accuracy of trained models based on this data about 75%, since most of the analyzed equipment has an insufficient set of data that directly affects the probability of failure of vehicle/special equipment.
The main consumer quality of our solution is to improve the accuracy of predicting breakdowns with a mathematical model, by obtaining relevant data from our m7pro tracker and connected additional sensors that directly and indirectly affect the failure of mechanisms. According to our estimates, with our solution, we will be able to achieve 95% accuracy of prognostic.