Arrowhead Framework Database System for Storage and Retrieval of Vehicle Signals

Abstract: Logging data is important in the digital world, both in the present and in the future. Using a database is a common way to store and retrieve data. Database prosumer system is a concept that may be the solution to many of the problems regarding data logging in the fourth industrial revolution. Earth-moving machines are evolving to being automatically operated, which demands scalable means for data management. Several software systems are needed to achieve a solution of this sort. A system to collect sensor data from the actual machine, a system to store the data and retrieve it, and systems that analyze, predict and actuate on that data. This thesis has its focus on storage and retrival of vehicle signal data, by implementing a Database system with the ability of easy update, adaptation and reuse in mind (quality attributes). The Arrowhead Framework, based on Service Oriented Architecture, is designed to handle communication between heterogeneous systems with cybersecurity. The first objective was to investigate if the Arrowhead Framework could guide towards a solution of the system, with the quality attributes in mind. The second objective was to examine what the consequences regarding batching data points to a database could be. InfluxDB and MongoDB are the databases used during the test. The Database System is implemented in Java containing four different modules. Client Common which holds necessary Arrowhead Framework resources. The Database Module, responsible for all database related functionality. The Database Provider which provides services regarding access of sensor data, and the Database Consumer responsible for collecting sensor data from different provider system(s). The system was tested by consuming a service at a Can Bus Provider, which is polling RPM data from a car. The rpm data is then stored in the implemented database. The test indicated that Arrowhead Framework can provide such a solution to achieve easy update, adaptation and reuse, on a small scale, old and in some cases non-existent documentation of the framework is holding the practical usage back. The result of batching data points concluded that the performance of InfluxDB writing operation improves with data point batching regards to performance, by writing several data points at once. MongoDB was not affected in the same way as InfluxDB, due to its fast single writing operation.