Machine Intelligent System Techniques for Automatic Harvest Systems

Publisher Summary This chapter describes some intelligent system techniques for an automatic harvest system with a stereoscopic observation function—an automatic harvest system with artificial fingers made of silicon rubber tubes that employs a three-dimensional measurement system by a single charge-coupled device (CCD) camera. This system estimates the position of the objects, directions to approach them, and the optimum path for gathering. The positions and directions for approach that avoid collisions with the branches and leaves of a fruit tree have to be known for the robot to gather objects without a human operator. The minimum distance of the gathering path to effectively get the objects needs to be known as well. This chapter also elucidates a new method termed virtual sensing, for collision avoidance of plural-hand robots. This system has an air-controlled robot hand called a soft hand for grasping a piece of fruit without damaging it. The 3D measurement system in this case is controlled by visual feedback from a single camera. This chapter provides an overview of the whole systems, the method of 3D measurement, the image processing, and the control method of the soft hand system. Furthermore, it describes a new technique for collision avoidance to avoid mutual collision when using plural robots.