Abstract
Advances in scientific research related to electric vehicles have led to generation of large amounts of data. This data is mainly logger data collected from various sensors in the vehicle and stored as flat files. It is predominantly unstructured and non-relational in nature, also called Big Data. Analysis of such data needs a high performance information technology infrastructure that provides superior computational efficiency and storage capacity. It should be scalable to accommodate the growing data and ensure its security over a network. This research proposes an architecture built on Hadoop to effectively support distributed data management over a network for real-time data collection and storage, parallel processing, and faster and random read access for information retrieval for decision-making. This system provides a simplified way of extracting from sensor data loggers and transforms this raw data into classified buckets. Once imported into a data store, the system supports data analytics over the data for knowledge discovery, and these analytics can help understand correlations between parameters under various circumstances. This system provides scalability to support data accumulation in the future and still perform analytics with less overhead, and its design can be employed to other fields with similar kind of data analytic challenges. Overall, open problems in data analytics are taken into consideration and a lowcost architecture for data management is proposed.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have