Autonomous hybrid GPS/reactive navigation of an unmanned ground vehicle for precision viticulture-VINBOT

Yield forecasting can provide important benefits for wine industry in terms of quality and efficiency. Vineyard yield estimation can be obtained using several methods being the most widely used the method based on visual assessment and/or counting/weighing the yield components The increasing importance of yield forecast has lead to automated solutions for the data acquisition and allowed the first service robotics applications in viticulture. In this paper we aim to present the preliminary results obtained in the European research project VinBot: “Autonomous cloudcomputing vineyard robot to optimize yield management and wine quality”. The paper focuses in the robot navigation. Robot navigation for agriculture has been a continuous research topic in the last years. Even there is a wide number of RTK-DGPS and PPP based navigation solutions available for precision farming, navigation in vineyards has some particularities and can benefit from other navigation techniques. The high cost and in some cases other limitations as fix ratios (determined by baseline distances to base stations), or communication limitations in the field make alternative solutions desirable. In this paper, we present a hybrid reactive/GPS based navigation scheme tested successfully in vineyard navigation. The proposed solution makes use of a laser range finder and RGBD device to perform reactive row following and obstacle avoidance, while it can make use of other reactive behaviors or GPS waypoint navigation for changing from row to row or field to field, thus supporting different levels of automation. The paper includes also some experiences with recently introduced new generation low-cost RTK-DGPS devices, that in the coming years will enable the progressive introduction of viticulture robots