Low-cost passive wireless water leak sensor for the automotive industry
memoria.pdf (17,63Mb) (Restricted access) Request copy
Què és aquest botó?
Aquest botó permet demanar una còpia d'un document restringit a l'autor. Es mostra quan:
- Disposem del correu electrònic de l'autor
- El document té una mida inferior a 20 Mb
- Es tracta d'un document d'accés restringit per decisió de l'autor o d'un document d'accés restringit per política de l'editorial
Document typeMaster thesis
Rights accessRestricted access - author's decision
The evolution of wireless technologies has shifted the expectations on how sensors anddata collection is performed. By reducing the complexity to deploy and connect sensors, anincrease in the variety and quantity of data collected is leading to the concept of ubiquitoussensing.However, there are measurements that still rely on manual work. In particular, the auto-motive industry still performs manually the rain isolation quality tests. These tests requirenowadays for an employee to manually place a handheld moisture sensor device in con-tact with the area to analyse, taking long time, low scalability and prone to human errors.Additionally, workers are exposed to these humid conditions which is hazardous for theirhealth.In order to tackle this problem, low-cost wireless identification tags are proposed in thiswork to be used instead. These tags, known as Radio Frequency Identification (RFID),are bundled in pairs, one waterproofed and one exposed to work as moisture detectionsensor. By interacting with them with a RFID telecommunications system, low level datais collected to perform predictions about their status. This way, it is possible to measuresignal differences when subjected to different quantities of water exposure, both in staticand in-movement scenarios, which lead to predictive methods of water leaks.In this work, different methodologies and setups will be evaluated on their performanceas predictors of wet RFID tags. The project involves researching the state of the art onthis topic, empirical experimentation of the different proposed methods, development of anautomated software platform and potential contribution to the academia with the resultsfound, by means of paper publication in conferences and/or journals. In particular, twodifferent experimental setups were tested, a single antenna setup and dual antenna setup.These setups were implemented in a controlled scenario, where 48 different experimen-tal iterations were performed. Later, the setup was implemented on a in-vehicle scenarioto emulate the real industrial operating scenario. In this scenario up to 70 different ex-perimental iterations where performed. When analysing the data samples, four differentmethods are proposed and implemented. Finally, the project also includes the evaluationthrough machine learning algorithms of the best method which achieved higher than80%accuracy test on predicting water leakages in real life scenario such as automotive “raintests”, through the developed sensors and system.With this research, the viability of this methodology for wireless detection of leaks hasbeen shown. Next steps are to explore different Tags manufacturers or complex scenariosinside production lines.