Development of a Visualization Framework for Evaluating Data Accuracy in an Embedded IoT Device

Sammanfattning

Husqvarna is a manufacturer of gasoline powered outdoor products that use embedded IoT devices, known as the Husqvarna Connectivity Device, to monitor equipment. These IoT devices collect and store data, which is later retrieved via Bluetooth and transmitted to Husqvarna fleet management cloud. Husqvarna currently lacks a framework to test and visualize whether operational data is accurately captured by their connectivity device. This bachelor thesis investigates the accuracy of the data reported by the IoT device by comparing it with simulated reference data using various visualization techniques. This thesis was conducted using a simulation box designed to replicate the behavior of real gasoline powered equipment. Engine revolution speeds were simulated and logged using the Inter Module Communication simulation box, and the resulting data was compared to histogram data collected from the IoT device. To enable analysis, data from the IoT device was retrieved from a server via a Bluetooth connection. Both the simulated and captured data were processed and manipulated to allow for effective comparison and visualization. Several visualization methods were employed such as bar plots, heatmaps, and step plots to analyze discrepancies between the data. The results indicate that discrepancies exist between the simulated data and the data captured by the IoT device at certain engine speeds. To display individual numerical values such as serial number and battery level, tables were used. The most effective method for visualizing histogram data was the use of bar plots, which allowed side by side comparison of engine revolution speed data. This thesis has explored various visualization techniques and developed methods for simulating and capturing data from the IoT device.