This project aims to develop and test an IOT-based vibration sensor that would later be deployed on houses to measure the vibrations experienced. This research is conducted at the request of Heijmans which is a Dutch leading firm in urban planning and infrastructure development. The project was undertaken to answer the client’s question of the type of vibrations that houses next to construction sites experience and determine if those are causing any structural damages.
Project description
The process of developing the electronics, mounting system and software follows the requirements outlined by the client, in which the system should be able to run autonomously for one year, the electronics must be matchbox size, and the mounting system should not leave any marks when removed.
As there are many unknowns, research was conducted to identify possible solutions for the problems. The hardware research is about topics such as microcontrollers, real time clock and battery protection. The mounting department focused on researching mounting systems that do not leave marks such as a drainpipe mounting and vacuum for rough surfaces, as well as possible materials to design the housing out of. The software department investigated data structures signal processing and machine learning algorithms.
Project results
One of the conclusions of our research and experiments is that developing a classification model requires multiple independent sensors recording at the same time. Each with validated sample frequencies and calibrated references. Ideally the use of pairs of several types of sensors would be useful in comparing data and excluding possible inconsistencies and outliers. Furtermore, for more reliable trends to emerge out of the data a significantly larger data set, compared to the one currently avaialable, will be required.
Video
Acknowledgements
The student project team (Cosmin, Stan, Megan, Lars and Bas) would like to thank the Heijmans construction company for their valuable contribution the project, as well as the Fontys project mentors of this proejct: Ing. J van Rens and Dr. C. Lee.