Klasifikasi Penyiraman Tanaman di Lahan Vertical Menggunakan Metode Perceptron

Abstract

This research was conducted to improve the process of watering the seeding stage in conventional vertical farming. The method of watering the seedlings has been changed to be automated using a watering device based on the Arduino Mega 2560 microcontroller. The watering cycle is connected to the conditions of temperature, humidity, and soil humidity obtained from the readings of the DS18B20 sensor for temperature values, the DHT11 sensor for air humidity values, and the YL-69 sensor for soil moisture values. This sprinkler implements the Perceptron algorithm to determine the status of the water pump and whether it should be activated or deactivated based on the temperature, air humidity and soil moisture values. Information from temperature, air humidity, and soil moisture can be monitored by plant owners through a web-based application. Based on the test results, applying the Perceptron algorithm to this automatic watering device has an accuracy value of 87.5% with an error percentage value of 12.5%.