Design and development of industrial prototype of spot specific orchard sprayer using ultrasonic sensing system and advanced pressure control mechanism

Variable rate technologies (VRT) and spot-specific applications are one of the key research areas of precision agriculture with the ability to facilitate the farmers by automatically targeting only plant areas to avoid unnecessary losses of agrochemicals. Spot-specific spraying not only saves costly agrochemicals, but it can also help to protect environment. The primary objective of this study was to develop a spot specific orchard sprayer (SSOS) system based on tree detection through ultrasonic sensors. For this purpose, a traditional two vertical booms horticulture sprayer with six nozzles on each boom was modified to a spot specific spraying using automatic control of individual nozzle. Two sensor booms were installed, one on each side of the SSOS, in front of the sprayer booms. Six ultrasonic sensors were mounted, three on each boom. The developed sprayer was tested for fast switching of electronic valves using power transistors and relay-based operation, and testing results showed a switching time of only 28 ms. Similarly, the sprayer was tested for accurate sensor spacing mount position and sensor mount angle to minimize the miss detection of the target tree canopy and testing results showed that a mount angle of 0° worked best during lab testing. Furthermore, we have developed a Machine Learning (ML) based Decision Tree model for a pressure control mechanism to protect the hydraulic piping system in the traditional sprayers. The rule-based decision tree algorithm precisely controls the Overflow Connections (OFCs) through accurate predictions for switching of such connections, resulting in 94 % accuracy. The developed SSOS was tested in different citrus orchards at University Research Farms, Rawalpindi, Pakistan. The test results showed that spot specific orchard spraying system saved 38 % agrochemicals compared to traditional spraying for citrus orchards.