Design and Comparison of Low-Cost Urine Level Detection Systems for Critical Patients

Abstract

With the rise in size of the elderly population suffering from urinary incontinence and the number of bedridden patients, individuals providing assistance to the aforementioned groups suffer from the “caregiving burden” due to the time-demanding nature of home health care. The current study proposes a low-cost Internet of Things (IoT)-based urine monitoring system for detecting a critical urine level in a urine collection container and sending an SMS text message to alert the caregiver. Remotely informing the caregiver of the patient’s urine level provides freedom of mobility. The proposed system consists of an Arduino microcontroller connected to a sensor measuring system and a SIM800L module with a SIM card. Three sensor measuring systems are suggested: Water Level Sensor, Ultrasonic Distance Sensor, and Weighing Scale Sensor. To compare the performance of the systems, 110 detection recordings from each system are used to construct a confusion matrix. The accuracy, precision, specificity, recall, and F-measure are computed for each system. Numerical and graphical analysis of the measurements show that the Water Level Sensor measurement system (accuracy: 0.96, precision: 0.98, specificity: 0.98, recall: 0.95, F-measure: 0.96) is the most reliable system, followed by the Weighing Scale Sensor (accuracy: 0.93, precision: 1, specificity: 0.936, recall: 0.88, F-measure: 0.93).