Improving Data Fusion in Big Data Stream Computing for Automotive Applications

Connected vehicles are capable of generating huge amounts of data at a very high frequencies. This big data has a great value for a large broad of services ranging from road safety services to aftermarket services (e.g., predictive and preventive maintenance). Nevertheless, they raised new challenges in terms of big data real-time or near-real time processing, storing, etc. Within this paper, we address the issue of online data fusion of automotive data. More precisely, we focus on the performance of the big data infrastructure to process collected data from several millions of connected vehicles. To this end, we propose novel approaches, based on spatial indexation, to speed-up our automotive application. To validate the effectiveness of our proposal, we have implemented and conducted real experiments on PSA-Group big data platform. The experimental results have demonstrated the efficiency of our spatial indexing and querying techniques.