Navigation of autonomous vehicle for rubber tree orchard

Abstract

This research has studied the navigation of autonomous vehicles in rubber tree orchards. It was emphasised on the rubber tree plantation in Thailand. Due to the high labour costs in the rubber industry and the demand for natural rubber has continued to decline over the past several years. The farmer were struggled to reduce the labour costs in production. The development of autonomous vehicle technology in rubber plantation can be a way to ease the production costs, such as reducing the amount of labour in fresh latex harvesting. This work was aimed to develop algorithm of a vision system for replacement of expensive sensors used in typical autonomous vehicles. The vision system model was implemented using only one camera installed on the vehicle to search for calibrated targets which were put on the rubber tree trunk. There were two important parameters, tree row offset and target distance, which were concerned when determining the performance of the autonomous vehicle in the rubber tree orchards. In order to evaluate the accuracy of the system, a set of experiments was conducted at a distance of 1-5 meters between the target and the camera. The results were compared between the controlled environment in the laboratory and the uncontrollable environment in the real rubber plantation.