A Setup of Mobile Robotic Unit for Fruit Harvesting

A. A. Aljanobi*, S. A. Al-hamed*, and S. A. Al-Suhaibani*

* King Saud University/ Department of Agricultural Engineering, Riyadh, Saudi Arabia janobi@ksu.edu.sa, alhamed@ksu.edu.sa and suhaiban@ksu.edu.sa

Abstract—The description of a mobile robotic unit for fruit harvesting was illustrated in this paper. The setup of the system was developed to harvest date palm fruit which is the most common fruit in Saudi Arabia. The system was based on readymade industrial robotic arm.

I. INTRODUCTION

Fruit production is the most labors demanding due to the lack of mechanization of harvesting and the short of the harvesting season. Several attempts to mechanize fruit harvesting specially by tree trunk shaking were achieved, which affect the tree on the long run [1;2]. In the last fifteen years, several works have been developed to design robotic systems for harvesting fruits especially orange [3;4] and apple [5;6]. These works did not succeed yet due to the size of the machine and the cost of development. On the other hand, several works have been done to develop small robotic units for harvesting agriculture products in greenhouse such as tomato and cucumber [7]. The development of these systems was based on robotic arm that can be guided to the right object through video camera. Fig. 1 and Fig. 2 show the design of such system for harvesting agricultural products in controlled environment.

The robot arm can distinguish between fruits and leaves by using video image capturing. The camera is mounted on the robot arm, and the robot picks the detected fruit based on image features. Air jet may be used to blow leaves that hide the fruit. The shape of the gripper depends on the fruit being picked, as some fruits, such as plums, crush very easily, while others, like oranges are not so susceptible to bruising. The robots should have access to all areas of the orchard in order to reach all of the fruit [8].

Figure. 1. The fundamental blocks of agricultural robots [8]

Figure.2. A robotic arm for agricultural purposes [9]

An eggplant-harvesting robot achieved a harvesting rate of 29.1%, averaging 43.2 seconds per fruit. The stereoscopic vision system of the tomato-harvesting robot could detect individual fruit, and detected the closest ripened tomato with an accuracy of approximately 85% [10].

Reference [11] stated that for automating the harvesting operation, an intelligent robot that can emulate the judgment of human labor is necessary. A robotic harvesting system that performs recognition, approach, and picking tasks should be developed. In order to accomplish these tasks, three essential components should be included in the system. First, a machine vision algorithm combining a color segment operation and a vertical dividing operation where the algorithm could detect the fruit even under different light conditions. Next, a visual feedback fuzzy control model to actuate a manipulator should be designed. Furthermore, an end-effector composed of a fruit-grasping mechanism, a size-judging mechanism, and a peduncle-cutting mechanism should be developed. These components should produce enough force for grasping the fruit and cutting the tough peduncle.

A procedure and the results of an optimal design of the kinematic structure of a manipulator to be used for autonomous cucumber harvesting in greenhouses was identified. The design objective included the time needed to perform a collision-free motion from an initial position to the target position as well as a dexterity measure to allow for motion corrections in the neighborhood of the fruits. A four link PPRR type manipulator was found to be most suitable. For cucumber harvesting four degrees-of-freedom, i.e. three translations and one rotation around the vertical axis, are sufficient [12].

The main objective of this work is to use a readymade industrial small robotic arm for harvesting fruits in

105

19th International Workshop on Robotics in Alpe-Adria-Danube Region – RAAD 2010 •June 23–25, 2010, Budapest, Hungary

978-1-4244-6886-7/10/$26.00 ©2010 IEEE

uncontrolled environments. The date palm tree was selected to apply this technique on due to it is the most common fruit in Saudi Arabia where there are more than 23 million tree producing more than one million tons, and the most costly in harvesting which my reach to half of the production cost.

II. COMMON METHODS OF DATE HARVISTING

The unique nature of date palm tree, require highly labor-intensive orchard operation under hazardous condition. As all of the fruits in a bunch do not ripen at the same time, it is necessary to make several picking during the short season [13]. Harvesting has historically involved selective hand picking of individual ripe fruit. There are two methods of harvesting the date fruits, traditional method and mechanical method. In traditional method, a labor climbs the tree with a safety belt around him and the palm. When he reaches the fruit bunches, begin to pick the ripe fruits, and then picked fruit are placed into carried basket which are then lowered to the ground with a rope [13]. Fig. (3) illustrates traditional method for date harvesting. Furthermore; in this method labor has to stay at the top of the tree, for a long time to pick ripe fruits. This practice, due to its difficulty is a dangerous job, and working for a long time in that height is very difficult, and labors risk serious accidents. So, it is done by skilled labor that has long time experience [13]. However, analysis of manual date palm climbing operation is achieved as listed in [14]. In a majority of the date growing regions, manual harvesting is the predominant method but labor shortages and human hazards during harvest operations motivated a need for improved mechanical harvesting methods and equipment [13].

Figure 3. Traditional method of date harvesting

In mechanized harvesting and related fruit separation several systems for elevating the workers to the fruit, removing the fruit from the bunch, and handling the harvested fruit between the field and packing house were investigated. Usual date harvesting machines are vehicle equipped with a long arm, at the end of which a man could stand in a basket and harvest the fruits [13]. In [ 15, 16, 17, 18 and 19], a date palm machine (Fig. 4) designed and tested in the field. It consists of an electro-hydraulic controlled basket mounted on a four-wheel-drive chassis. This is driven by a diesel engine used to enable the worker to be elevated into the tree or to power the wheels. The output of harvested dates in kg/man hour is greater with the machine than with traditional hand harvesting. In [20], a tractor-mounted date palm tree service machine was developed and tested. A hydraulically driven winch is available to lower the harvested dates in a clean basket to the ground. Compared to the traditional way of serving palm trees, where a worker climbs the tree using a piece of rope, the system provides a safer and more comfortable working environment. Although many date farms still use ladders, a U-Shaped basket on a forklift to reach the dates could be used [21].

III. THE MOBILE ROBOTIC UNIT FOR DATE HARVISTING

Saudi Arabia is at the forefront of the world production of dates, producing about one million tons of dates annually [22]. With an increasing number of palm trees planted and the limited availability of skilled labor and increased the cost of harvest dates, this makes the interest in raising the efficiency of operations and reap the harvest of dates in Saudi Arabia is important work. So, develop a Mobile Robotic Unit for Date Harvesting (MRUDH) is our research interest in the College of Agriculture and Food Sciences, Department of Agricultural Engineering, King Saud University. This unit takes different stages to be visual. The first stage in developing the MRUDH is to study the characteristics of fruit bunches based on their position on palm tree. These characteristics include diameter of fruit bunch (D), the distance of fruit bunch from the trunk (L), the radius of a pregnant fruit bunch (R), and the horizontal distance between the two fruit bunches (S) as shown in Fig. 5. Measurements in accordance with a changing age and dates type in the area of Riyadh were achieved after ripping the dates on palm tree. The dates types include different varieties. The age palm tree ranging between 8 to 15 years.

Figure 4. Date harvesting could be done by mechanical ladder

[15]

A. A. Aljanobi et al. • A Setup of Mobile Robotic Unit for Fruit Harvesting

106

Figure. 5. Characteristics of fruit bunch based on its position on palm

tree

The statistical analysis, which show a moral difference between the varieties of palm studied with an overall average of diameter of fruit bunch (D), the distance of fruit bunch from the trunk (L), the radius of a pregnant fruit bunch (R), and the horizontal distance between the two fruit bunches (S) is 375 mm, and 482 mm and 699 mm and 346 mm, respectively. The horizontal distance between the two fruit bunches (S) increased by increasing tree age with linear relationship with R2 of 0.924. The robotic arm suitable to deal with this characteristics has length of minimal 600 mm, and degree of freedom not less 4.

The second stage in developing the MRUDH is selecting the robot arm that meet the requirements of characteristics of date bunch. By browsing through manufactures of robotics, a robotic arm of Neuronics company [23] was found to be meet the required specifications (Fig. 6) in primary works. The specifications of that robotic arm are shown in Table 1.

The third stage in developing the MRUDH is selecting the lifter to hold the robotic arm and reach to the right position at the palm tree. It is Caterpillar Telehandlers TH62. Some specifications of the used Caterpillar Telehandlers are shown in Table 2. With the wide range of lift capacities and lift heights, Cat Telehandlers give us the strength and power to take the robotic arm and loads up and out. Also, it has two forks.

The fourth stage in developing the MRUDH is assembled the robotic arm on the Caterpillar Telehandlers forks. Special platform was designed with width of 1200 mm and depth of 300 mm to carry the robotic arm and control box. This platform was assembled and sliding on the two forks. The distance between the robotic arm and the wall of the forks was 170 mm.

Figure 6. The robotic arm from Neuronics company

TABLE 1. SPECIFICATIONS OF THE ROBOTIC ARM

Drive DC motors with position encoders Repeat accuracy ±0.1 mm Degree of freedom 5 to 6 Working radius Up to 60 cm Mechanical design High-strength aluminum, anodized Net Weight 4.8 kg (without control box and gripper) Max power 96 W(24 V/4A) Speed 90º/sec Working space 517 mm (without gripper) Payload from flange 400 g Force 4 N (in all directions) Point speed < 1 m / sec Standard interfaces Digital I/0, Ethernet, Modbus TCP/IP, (additional I/0 and PLC-connection), USB, LAN, Soft Stop, UPS, C++ interface, support for Python

TABLE 2.

SOME SPECIFICATIONS OF THE CATERPILLAR TELEHANDLERS TH62

Engine Model Cat 3054E Gross Power 74.5 kW Operating Specifications Rated Load Capacity 3600 kg Maximum Lift Height 7200 mm Load at Max Reach 1250 kg Load at Max Height 2000 kg Maximum Forward Reach 3800 mm Dimensions Height 2270 mm Width 2350 mm Ground Clearance 425 mm Transmission Speeds Forward - 1 7 km/h

IV. RESULTS AND DISCUTION

The MRUDH was tested for measuring the movement

of the arm to select and pick up the right one. At each position, the system was calibrated to a zero level, then the unit measure the movement of the arm around dates on bunch. The system shows promising efficiency as shown in Fig. 7 and Fig. 8.

Figure 7. The arm reached the date

107

19th International Workshop on Robotics in Alpe-Adria-Danube Region – RAAD 2010 •June 23–25, 2010, Budapest, Hungary

Figure 8. The arm hold one date

ACKNOWLEDGMENT

The authors are grateful to King Abdulaziz City for Science and Technology (KACST) for funding the project MT-1-3.

REFERENCES

[1] FAO;" Date palm cultivation," FAO Plant production and protection Paper 156 Rev. 1 Edited and compiled by Abdelouahhab Zaid, Rome, 2002.

[2] S.M. Mazloumzadeh, M. Shamsi, and H. Nezamabadi-pour," Evaluation of general-purpose lifters for the date harvest industry based on a fuzzy inference system," Computers and Electronics in Agriculture, 60(1),2008, pp. 60-66.

[3] M. Recce, J. Taylor, A. Plebe, and G. Tropiano," Vision and neural control for an orange harvesting robot," Proceedings of the 1996 International Workshop on Neural Networks for Identification, Control, Robotics, and Signal/Image Processing (NICROSP '96.p. 467,1996.

[4] G. Muscato, M. Prestifilippo, N. Abbate, and I. Rizzuto," A prototype of an orange picking robot: past history, the new robot and experimental results," Industrial Robot: An International Journal,32 (2),2005,pp.128-138.

[5] L. Zhaoxiang, and L. Gang," Apple maturity discrimination and positioning system in an apple harvesting robot," New Zealand Journal of Agricultural Research Abstracts 2007, vol. 50,2007, pp.1103–1113.

[6] D.M. Bulanon, T. Kataoka, Y. Ota, and T. Hiroma, " A machine vision system for the apple harvesting robot," Agricultural Engineering International: the CIGR Journal of Scientific Research and Development. Manuscript PM 01 006, vol. III,2001.

[7] N. Kondo, M. Monta, and T. Fujiura," Fruit harvesting robots in

Japan," Advances in Space Research, 18 (1-2), 1996, pp.181-184. [8] http://kernow.curtin.edu.au/www/Agrirobot1/frutrob.htm. [9] A. R. Hirakawa, A. M. Saraiva, and C. E. Cugnasca," Wireless

robust robot for agricultural applications," Proceedings of the World Congress of Computers in Agriculture and Natural Resources (13-15, March 2002, Iguacu Falls, Brazil),2002, pp. 414-420.

[10] S. Hayashi, T. Ota, K. Kubota, K. Ganno , and N. Kondo," Robotic harvesting technology for fruit vegetables in protected horticultural production," Information and Technology for Sustainable Fruit and Vegetable Production FRUTIC 05, 12-16 September, Montpellier France,2005.

[11] S.Hayashi, K. Ganno, Y. Ishii, and I. Tanaka," Robotic harvesting system for eggplant," JARQ 36 (3),2002,pp. 163 – 168.

[12] E.J. Van Henten, D.A. Van’t Slot, C.W.J. Hol, and L.G. Van Willigenbur,"Optimal manipulator design for a cucumber harvesting robot," Computers and Electronics in Agriculture 65,2009,pp. 247–257.

[13] M. Abounajmi, "Mechanization of dates fruit harvesting," Paper Number: 041028,ASAE/CSAE Annual International Meeting,2004.

[14 ] A. S. Hamam, "Analysis of manual date palm climbing operation," Misr J. Ag. Eng., 20 (1),2003,pp.253-267.

[15] S.A. Al-Suhaibani, A.S. Babeir, J. Kilgour, and J.C. Flynn, "The design of a date palm service machine," Journal of Agricultural Engineering Research 40(2),1988, pp. 143-157.

[16] S. A. Al-Suhaibani, A. S. Babier, and J.Kilgour,"Design specification of a date palm service machine," Agricultural Mechanization in Asia 21(4),1990,pp. 53-60.

[17] S. A. Al-Suhaibani, A. S. Babier, J. Kilgour, and B. S. Blackmore, " Field tests of the KSU date palm machine," Journal of Agricultural Engineering Research 51,1992,pp. 179-190.

[18] S. A. Al-Suhaibani and A. Babaeer," Field tests of the KSU date palm service machine," Arabic Gulf Journal of Scientific Research. 9(3),1991,pp.55-73.

[19] S. A. Al-Suhaibani, A. S. Babaeir, M.L. A. Bascombe, and J. Kilgour," The KSU date palm service machine portable bridge units: design and test program," J. King Saud Univ., vol. 5, Agric. Sci. (2),1993,pp.127-139.

[20] M. A. Fadel, "Development of a tractor-mounted date palm tree service machine," Emir. J. Agric. Sci. 17 (2),2005,pp.30-40.

[21] http://www.eingedi.dead-sea.k12.il/EinGedi/datepalm.htm. [22] http://faostat.fao.org/site/339/default.aspx. [23] www.neuronics.ch.

A. A. Aljanobi et al. • A Setup of Mobile Robotic Unit for Fruit Harvesting

108