Due to the increasing difficulty of bridge maintenance, bridge monitoring is used increasingly to detect damages at an early stage. This paper deals with the development of a low-budget monitoring system for expansion joints. The aim of this paper is to select the hardware of a low-budget monitoring system, create an artificial neural network model, that can distinguish between a good and a broken expansion joint and finally transmit the data set to a dashboard in a web application. The idea is to record audio data directly under the expansion joints using a microcontroller. For the selection of the right hardware, two microcontrollers were tested, on the one hand Arduino Nano 33 BLE Sense and on the other hand Arduino Portenta H7 with Vision Shield. Then the audio samples are uploaded to Edge Impulse and the data set is assigned to two classes: good and damaged. An artificial neural network is then created in several steps, starting with the creation of a so-called impulse, testing the model and loading the model onto the Arduino device. The results of the monitoring system are transmitted to a dashboard in a web application in real-time.
CITATION STYLE
Ambros, L., Binder, N., Hölzl, C., & Vill, M. (2023). Development of a Low-Budget Monitoring System for Expansion Joints with Real-Time Data Analyses. In RILEM Bookseries (Vol. 43, pp. 891–901). Springer Science and Business Media B.V. https://doi.org/10.1007/978-3-031-33211-1_80
Mendeley helps you to discover research relevant for your work.