Sistem Pemantauan Dan Peringatan Volume Cairan Intravena (Infus) Pasien Berbasis Edge Computing

Abstract

Manual infusion fluid volume monitoring in hospitals has the risk of delays, which can endanger patients such as dehydration or electrolyte disturbances. This problem is caused by the limited time and labor of nurses in monitoring infusions in real-time. This research aims to develop an edge computing-based infusion fluid monitoring and alert system to monitor the volume and speed of infusion droplets in real-time through the website. This system uses NodeMCU ESP32 to collect data from load cell and photodiode sensors, which are then processed on the edge server and displayed on the website. Tests were conducted in three conditions, namely Line of Sight (LoS), Non-Line of Sight (NLoS), and without edge computing. The test results show that the LoS condition provides the best performance with an average delay of 113 ms, while the NLoS condition increases the delay to 146 ms. Without edge computing, the delay is higher than with the use of edge computing. Based on the test results, the use of edge computing in LoS conditions provides the best performance with the lowest delay, thus significantly improving the efficiency of infusion monitoring and reducing the risk of delayed infusion fluid replacement. The system has a measurement accuracy of 99.48% on the first infusion device, 99.58% on the second infusion device, and 99.57% on the third infusion device.