Design of intelligent sprayer control for an autonomous farming drone using a multiclass support vector machine

The increasing need for agricultural commodities poses a serious threat to the cultivating ability of agricultural supplies. The use of autonomous farming drones to support the cultivation process is a rapidly growing trend. In an effort to improve efficiency and accuracy in watering using drones, this research strived to propose a control design for the drones using the multiclass support vector machine (MSVM) method. The proposed system was an improvement compared to the common approach of using constant watering strength in all drone conditions. By obtaining suitable watering strength based on the drone's altitude, wind speed, and speed sensor data as the input, the optimal solution between water usage and water efficiency was expected to be achieved. An experimental trial that consisted of 12 flights was conducted to acquire a data set with 3,750 data. The results of classification with MSVM obtained an accuracy of 90.82%. The efficiency of using the water resource and the accuracy of delivering the correct amount of water based on the drone condition were achieved. These results show that the proposed technology has great potential for using drones as an automatic watering system in agriculture.