Abstract
Fundamental agriculture operations can benefit from the adoption of advanced robotic platforms for enabling precise, repeatable and autonomous operations. Among others, precise plant treatment and monitoring tasks are critically important for estimating plant yields or trying to prevent and detect diseases. However, the uneven and irregular terrain of vineyards cause jumps during robot motion that pose challenges to precision caring activities. For this reason, our mobile robot manipulator uses a camera mounted on its end-effector to track a landmark and to control the end-effector pose with respect to it. We show the effectiveness of our approach through lab experiments, where we created an in-lab example of a vineyard row with synthetic grapes, leaves, and ramps to simulate the rough terrain.