This paper deals with the processing of vehicle data in a stream using machine learning. In modern vehicles, a significant amount of sensor data are generated, which are only used temporarily before being discarded. The toolchain presented here aims to historize the data and evaluate it in near real time. Stream machine learning is used to process the data. Requirements for the toolchain are the platform independent use of these and the free provision of the tools used. The result is a complete and innovative toolchain that maps everything required. From the reading of data on the vehicle to the use of stream machine learning and the evaluation of the data. An illustrative use case is presented and an outlook on extensions of the toolchain is given.
Jacob, T., Kubica, S., Rocco, V. (2018). Stream machine learning on vehicle data. In CANDO-EPE 2018 - Proceedings IEEE International Conference and Workshop in Obuda on Electrical and Power Engineering (pp.55-59). IEEE [10.1109/CANDO-EPE.2018.8601157].
Stream machine learning on vehicle data
Jacob, T;Rocco, V
2018-01-01
Abstract
This paper deals with the processing of vehicle data in a stream using machine learning. In modern vehicles, a significant amount of sensor data are generated, which are only used temporarily before being discarded. The toolchain presented here aims to historize the data and evaluate it in near real time. Stream machine learning is used to process the data. Requirements for the toolchain are the platform independent use of these and the free provision of the tools used. The result is a complete and innovative toolchain that maps everything required. From the reading of data on the vehicle to the use of stream machine learning and the evaluation of the data. An illustrative use case is presented and an outlook on extensions of the toolchain is given.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
