Smart data preprocessing method for remote vehicle diagnostics to increase data compression efficiency

摘要

The increasing number of functions in modern vehicle leads to an exponential increase in software complexity. The validity and reliability of all components must be ensured, making the use of appropriate vehicle diagnostics systems indispensable. The purpose of such systems is to collect and process data about the vehicle. To find issues during vehicle development, the OEMs will usually have a development fleet of thousands of vehicles. The challenge for diagnostic systems is to detect issues during these tests, as well as collecting as much data as possible about the circumstances that led to the fault. A singlevehicle produces hundreds of gigabytes of data per month. The required data bandwidth cannot be fulfilled by current mobile network subscriptions as well as WIFI or cable-based infrastructure. This limits the amount of data that can be collected during field tests and hinders big data analysis like AI training or validation. Hence a software solution for data reduction is necessary. The authors present a method for data handling that drastically reduces the amount of data consumption and optimizes the transfer delay between a remote-diagnostic systems and the cloud. Using a pipeline of data preprocessing as well as an established compression algorithm, the amount of transmitted data is reduced by a factor of nearly ten. This method will allow to collect more data in field testing and improve the understanding of issues during vehicle development.