Citation

Lokman, Amar (2022) Scheduling and predictive maintenance for smart toilet. Masters thesis, Multimedia University.

Full text not available from this repository.

Abstract

Modern society requires toilets. Worn-out appliances and expensive cleaning services cause a low sanitation rating. In addition to maintenance and cleaning, the strategy calls for a cheap, reliable sensor. There were three objectives established for this study. The primary objective is to create an IoT (Internet of Things) administration platform. The benefits and drawbacks of various existing systems have been analyzed through literature reviews. Second, propose effective predictive maintenance to help anticipate when certain pieces of bathroom machinery may break down. Finally, a scheduling algorithm was deployed to calculate how many janitors are needed to keep the restrooms clean. It will quantify the model's efficacy and draw conclusions about its future use. Connected infrared, temperature, and humidity sensors help establish a bathroom's Internet of Things setup. Sensors have been researched in depth to determine the best way to adapt them to the hygienic and private conditions of a bathroom. With further development and testing, the sensor's accuracy and cost-effectiveness could be improved. The ARIMA model has been shown to accurately predict subsequent lags in time series, making it a promising candidate for use in predictive maintenance. Time series predictions are also a specialty of LSTM, so it is only fair to evaluate the proposed model against LSTM. In this study, it employs the ARIMA model to address the predicament of RUL prediction procedures by adjusting the parameters MA and AR. Finally, a genetic algorithm for scheduling is utilized to develop a cleaning schedule for the janitorial staff. In order to solve the problem of scheduling cleaning staff, a genetic algorithm is proposed. This algorithm improves the design of the classic genetic algorithm by studying the constraints of both soft and hard scheduling. The Greedy algorithm also applied to the same data set, is used for comparison. The study concludes with experimental evaluations showing that the proposed model ARIMA and GA provides a better fit for both objectives.

Item Type:	Thesis (Masters)
Additional Information:	Call No.: TK5105.8857 .A43 2022
Uncontrolled Keywords:	Internet of things
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5101-6720 Telecommunication. Including telegraphy, telephone, radio, radar, television
Divisions:	Faculty of Computing and Informatics (FCI)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	10 Jan 2024 05:40
Last Modified:	10 Jan 2024 05:40
URII:	http://shdl.mmu.edu.my/id/eprint/12025

Downloads

Edit (login required)