The Impact Of Distributed Data Preprocessing On Automotive Data Streams

Vehicles have transformed into sophisticated com- puting machines that not only serve the objective of transporta- tion from point A to point B but serve other objectives including improved experience, safer journey, automated and more efficient and sustainable transportation. With such sophistication comes complex applications and enormous volumes of data generated from diverse types of vehicle sensors and components. Automotive data is not sedentary but moves from the edge (the vehicle) to the cloud (e.g., infrastructure of the vehicle manufacturers, national highway agencies, insurance companies, etc.). The exponential increase in data volume and variety generated in modern vehicles far exceeds the rate of infrastructure scaling and expansion. To mitigate this challenge, the computational and storage capacities of vehicle components can be leveraged to perform in-vehicle operations on the data to either prepare and transform (prepro- cess) the data or extract information from (process) the data. This paper focuses on distributing data preprocessing to the vehicle and highlights the benefits and impact of the distribution including on the consumption of resources (e.g., energy).