Development of a robotic hand for harvesting cucumber [Cucumis sativus] fruits

• Published in: Journal of Agricultural Meteorology Vol. 45; no. 2; pp. 93 - 97
• Main Authors: Amaha, K. (Tokyo Univ. (Japan). Faculty of Agriculture), Takakura, T
• Format: Journal Article
• Language: English; Japanese
• Published: Tokyo The Society of Agricultural Meteorology of Japan 1989; Japan Science and Technology Agency
• Subjects: AUTOMATAS; AUTOMATE; Cucumber; CUCUMIS SATIVUS; EQUIPO DE RECOLECCION DE FRUTAS; ESSAI; Harvesting robot; MATERIEL DE RECOLTE DES FRUITS; PRUEBAS; Robotic hand; Shape recognition
• ISSN: 0021-8588; 1881-0136
• DOI: 10.2480/agrmet.45.93

A robotic hand for harvesting cucumber fruits has been developed. The hand is capable to follow the curved shape of located fruit and to detect and cut its peduncle. The hand consists of three parts, the gripper, the sensor and cutter, and the free connector. The gripper serves to grasp and hold a fruit. The sensor and cutter travels following the shape of the grasped fruit. The sensor detects where the peduncle is by measuring the diameter of the fruit continuously until the diameter is sensed to be less than 8(mm), at which point, it is judged to be the peduncle. The cutter then rotates and cuts the peduncle to release the fruit from the canopy. If the fruit is curved, the sensor and cutter can move upwards following the curved shape by means of the free connector. The free connector has tow degrees of freedom of rotation and one of expansion and contraction. The hand is also designed to handle fruits gently without inflicting damage. Motion tests have been made with real cucumber fruits in the laboratory. The fruits in these tests were placed in the hand manually. It was concluded as follows. 1) The hand could follow the shape of most fruits except for those that were very curved. 2) The sensor could detect all peduncles except for those that were heavily curved, which the hand could not follow. 3) The cutter was successful in cutting the peduncles. It is necessary to improve the hand so that severely curved fruits can be followed, and to do additional motion tests with randomly selected fruits in order to establish the percentage of fruits which can be harvested. The most important subject for future study would be to test the hand in a real canopy