We propose a new active vision of a heavy material handling agricultural robot emphasizing the initial cost. A key point is a new combination of a structure system and a control system, not only one of the two. First, a camera configuration with only one camera is proposed to achieve lower initial cost. This configuration requires more robustness. Second, in order to realize the vision, a robust control system is designed in the presence of uncertainty and constraint. Finally, the validity is confirmed by experiments in an actual open field. The camera configuration can not work well without the designed controllers. That is, the robust controllers reduce the initial cost.

Global performance of agricultural robots is discussed to design the structure system reasonably. First, problems of the theoretical field capacity are summarized. Then a new macroscopic working environmental model is proposed. Third, the theoretical field capacity is extended to solve the problems based on two fundamental equations of working space and working time. Finally the validity of the extended theoretical field capacity is confirmed by deriving two design guidelines for structure systems of agricultural robots.

An experimental study is presented on control of a heavy material handling agricultural manipulator using robust control. The harvesting of heavy vegetables and organic compost fertilization requires hard manual labor. For heavy material handling, an agricultural robot was proposed in the first step. And then a control system using LQ control was designed and high-speed manipulation was realized in the second step. In this paper, we designed a control system to achieve robust performance in the presence of uncertainty. The robotic harvesting experiment was done in the watermelon field and the use of robust gain scheduling and /spl mu/-synthesis was confirmed.

This paper presents a manipulator which is able to handle heavy materials for agricultural applications. The characteristics of agricultural operation are discussed and extracted. As the manipulator for handling heavy materials is analyzed using kinematic indices, the parallel type manipulator is shown to be superior to the other manipulators (i.e. the polar coordinate type, articulated type and cylindrical coordinate type manipulators). A parallel type manipulator has therefore been designed and developed. The robotic harvesting experiment was carried out using the parallel type manipulator in a watermelon field.