GSJ: Volume 10, Issue 1, January 2022

ISSN 2320-9186 369

GSJ: Volume 10, Issue 1, January 2022, Online: ISSN 2320-9186

www.globalscientificjournal.com
Tomato Harvesting Robot Based on Solid Work
*Md Newaz Sharif 1, **Muhammad Uzair 2
*Department of Mechanical Engineering ,Wuhan Textile University, China
**Departments of Mechanical Engineering, Nanjing University of Science and Technology, China
Email: *[email protected]

Abstract
Designing and development of agricultural robot is always a challenging issue, because of robot intends to work
an unstructured environment and at the same time, it should be safe for the surrounded plants. Therefore,
traditional robots cannot meet the high demands of modern challenges, such as working in confined and
unstructured work spaces. Based on current issues, we developed a new tomato harvesting robot arm with a
flexible backbone structure for working in confined and extremely constrained spaces. Moreover, we optimized a
tomato detaching process by using newly designed gripper with passive stem cutting function. Moreover, by
designing the robot we also developed ripe tomato recognition by using machine learning. This paper explains
the proposed continuum robot structure, gripper design, and development of tomato recognition system.
Keywords: Tomato harvesting, gripper, tomato detection, design, agricultural robot.

1 Introduction of harvesting robot are a mobile platform, an arrival

manipulator, a grasping tool and a tomato recognition
The harvest of fruits and vegetables is an activity algorithm. In this study, we propose a robot arm
of human existence. Fresh tomato is a good product, manipulation, grasping tool and tomato recognition
and it is widely used in China and the world every system. In this study, we propose a new tomato
year. Tomato is one of the most popular and widely harvesting robot system with arm and tomato
cultivated vegetables in the world, with an annual detection system[5]. This paper is organized in the
output of about 60 million tons. Fresh fruits and following order: design concept, motion / kinematics
vegetables are very important sources of vitamins and formula, development of recognition system and
minerals, which are very important to human health. experimental results.
Due to the rapid growth of population, the domestic
consumption and demand of tomato are increasing, so 1.1 Project Background
it is necessary to harvest properly to ensure the
quality of consumption. In many parts of the world, it This current project aims to design a robot that
is considered to be a major cash and industrial can recognize and harvest mature tomatoes without
crop[1]. Due to the continuous rebound of labor costs, human help or guidance. If successful, the pilot
the cost of harvest has increased significantly in project will be further developed in agriculture and
recent years, or the price of harvest has reached 1 / 4 greenhouse industry. The idea of self harvesting fruit
of the total cost. With the decrease of agricultural is not new because it has been used in different
population and the increase of labor cost, automatic countries. However, the implementation of these
harvesting technology has attracted people's attention technologies is not very common in the region.
and may be applied in the future. It's a fact that every Today, depending on the type of crop that
year in China there is plenty of land for farming, and produces fruit, it can be harvested manually or by
then good crops. The country is one of the leading machine. Most commercial robots are used in
producers of tomatoes and other vegetables. Tomato industry, while robots for agriculture, forestry and
is one of the most common vegetables, because of its fishery are being developed, but not yet
delicious taste, rich nutrition, easy to eat, and its yield commercialized. The reasons for the delay are the
accounts for 15% of the total yield of vegetables. cost of robots, the safety of using robots, the
Tomatoes are grown as fresh market and processed difficulty of outdoor operation and the knowledge
tomatoes. They are essential to the food industry transfer from farmers to computers. If we can
because they are raw materials for the production of overcome these difficulties and apply robots to
value-added products. Harvesting robots are designed agriculture, robots can contribute to saving labor,
to use a variety of sensors to sense the complex improving production and automation of production
agricultural environment and use this information lines. Moreover, it is possible to manage the quality,
together. Harvesting robot is designed to use various quantity and environmental conditions of agricultural
sensors to sense the complex agricultural products, and intelligent agriculture will be realized.
environment, and use this information with the target Although harvesting robot has a broad prospect for
to perform certain operations. The main components the future, the overall performance of harvesting

GSJ© 2022
www.globalscientificjournal.com
GSJ: Volume 10, Issue 1, January 2022
ISSN 2320-9186 370

robot can not be compared with manual operation[2]. are designed to harvest tomatoes at a certain height
The bottleneck of commercialization of harvesting along the guide rail. Two wheel drive control and
robot is low efficiency and high cost. The differential steering control can not meet the demand
combination of human and robot is a feasible way to of automatic harvesting of fruits and vegetables.
improve the success rate of robot harvesting line Moreover, the picking arm is usually replaced by the
method. If we can promote agricultural robots, we industrial robot arm, which has the disadvantages of
can predict labor saving, automation, production high cost and control complexity. The harvesting time
growth and so on. In addition, it is possible to of citrus is 3-7 s, melon is 15 s, cucumber is 10 s,
transform facts such as management of crops, eggplant is 1 min, and the harvest rate is less than
environment, cultivation, state and quality. From the 90%[4]. Fruits and vegetables are mainly identified
perspective of safety, appropriate guidance can be by color features, gray threshold and geometric shape.
provided to consumers. There are many serious These algorithms may be affected by illumination and
challenges to overcome, but it is possible to establish environmental factors, and it is difficult to identify
a new robot market in agriculture. Among all kinds of overlapping fruits.
fruits and vegetables, tomato is one of the main fruits
that consume a lot of fruits and vegetables. Many 1.4 Research Content
tomatoes are planted in houses and other facilities,
but due to the high temperature and humid working The task is to build a tomato harvesting robot,
environment and long harvest period, labor is large which can detect and harvest mature tomatoes, and
and labor needs to be saved. learn the automatic system in the process to pick
tomatoes from tomato plants safely and accurately. It
1.2 Research Status is expected that modern farms will produce higher
yield per unit area, provide high-quality products in a
In Japan, the Netherlands, Britain, France, Italy, sustainable way at lower cost, and have less
the United States, Israel and other countries, various dependence on labor force. Implementing digital
robots have been developed to harvest tomatoes, agriculture and the response of precision management
cucumbers, grapes and oranges. Kyoto University has in a specific location to this expectation, which not
developed a tomato harvesting robot manipulator only depends on the sensor technology, but
with five degrees of freedom. Okayama University continuous collection of field data only through the
has developed a 7-DOF robot, which is composed of proper use of agricultural robots is feasible.
motion system, vision system, end effector, Agricultural scientists, farmers and growers also
manipulator and control system. Later, a new tomato face the challenge of sustainably producing more
harvesting robot was developed, which was food on less land in a way that meets the projected
composed of vision system, manipulator, control needs of 9.8 billion people by 2050. This is
system and rotating arm. It takes about 15 seconds for equivalent to feeding a new city with 200000 people
each tomato from recognition to input, and the a day. Integrated digital tools, sensors, and control
success rate is 50% - 70%. In addition, the Korean technology accelerates the design and development of
Agricultural Mechanization Research Institute has agricultural robot technology, showing great potential
also developed a series of tomato harvesting and potential the benefits of modern agriculture.
robots[3]. The vision system can accurately These developments start from the collection of
determine the surface color of tomatoes, so that the accurate and detailed information to digitize the
robot can selectively pick mature tomatoes. However, timely spatio-temporal information of plants and
the potential disadvantages of these robots are slow fields to complete the complex nonlinear control
response and clumsiness. tasks of robots navigation. Robots and manipulators
in the field of agriculture have become an important
1.3 Domestic Research Status part of digital agriculture and precision
agriculture[6].
In China, the research of tomato harvesting robot
has made great progress, such as manipulator, image 2 Overall Design of Tomato Picking Robot
recognition and motion control. Panasonic
demonstrated the initial version of the harvesting 2.1 Gripper Design
robot and exhibited it at the latest international
robotics exhibition in December 2015. The robot can We use solid works software to design the gripper
pick tomatoes and move around the place. The robot of tomato harvesting robot. Solid works is a highly
puts tomatoes in the basket and replaces them once efficient 3D CAD software tool with integrated
they are full. The robot uses image sensing to detect analysis tools and design automation functions,
the color and shape of tomatoes. It also uses sensors which can help stimulate various physical behaviors
to identify the location of tomatoes. Usually, the such as motion, dynamics, stress, deflection,
current robot is not intelligent enough, and the vibration, temperature or fluid flow to adapt to all
success rate and pitching rate are far lower than types of design. Designers and engineers from all
expected. In addition, most tomato harvesting robots over the world and industries are using solid works to

2
GSJ© 2022
www.globalscientificjournal.com
GSJ: Volume 10, Issue 1, January 2022
ISSN 2320-9186 371

create innovative products.

Fig 1: Solid grip design Fig2:Installation of gripper on Manipulator

2.1.1 Installation of Gripper on Manipulator

2.2 Working Process of Gripper
The manipulator has a system, which can quickly
switch tools, supply power and exchange information Designing grippers for harvesting tomatoes is a
between the manipulator and tools with the help of challenging problem because tomatoes are a soft,
electronic connector. The gripper is instaled on the juicy fruit. Grasp the tomato gently to prevent any
manipulator, which can pick up and pull out mature malfunction such as over voltage. In addition, the
tomatoes without damaging them. To prevent separation of tomato and tomato stem should be
tomatoes from falling, the gripper is equipped with a considered in tool design. Therefore, we designed a
supporting spherical round clamp. The picture below hemispherical grab tool for grasping spherical objects
is straight. The design of the gripper is based on the such as tomatoes. To separate the tomatoes, we added
fixture of two fixtures. Two clamps enable the clamps a cutting blade to the edge of the cup. This design
to adapt to the shapes of tomatoes of different sizes. makes it possible to separate tomatoes in a passive
When the robot arm pulls the tomato out of the plant, way and shorten the harvest time
both fingers hold the tomato in the fixture. However,
instead of putting the tomatoes in the basket, place
and hold the ball clip at the bottom so that the
tomatoes do not fall. Once the tomato is firmly fixed,
the robotic arm pulls and twists the tomato until it
leaves the plant. The spherical clamp of the fixture
shall be designed so that its surface is smooth and
there is no edge that may damage the tomato or plant.
The two s in the grippers are connected to each other
at 120 degrees in opposite ways and are actuated by a
single actuator using a gear train[7]. The servo motor
is mounted on the base and drives the directly
connected worm gear. The worm gear transmits Fig 3: Working process of gripper
rotational motion to the worm gear set, which is
connected to each fixture. Clamping is of the highest The average size of cherry tomato was calculated to
importance because without this feature, the robot define the size of grip cup. According to our
will not be able to collect tomatoes. Removing calculation, the diameter of tomato is 48-54mm.
tomatoes from plants seems to be a cleaner way. The Before designing the gripper, the requirements must
tomato will not be damaged by partial pressure, but be set. Since the grabber will be used to pick
the tensile force will lead to the peak pressure on the tomatoes, the earlier named features must be
top of the tomato. But it's more important to hold on mentioned. The diameter of tomato can be classified
to the tomatoes because they slip during twisting. The as:
main advantage of the robot arm fixture is flexibility. Extra small: 48-54mm
Small: 54-58
Middle: 58-64
Large: 64-73
Super large: 73-88
Up to 88 mm.
This means that the range of the gripper should be at

3
GSJ© 2022
www.globalscientificjournal.com
GSJ: Volume 10, Issue 1, January 2022
ISSN 2320-9186 372

least 48mm to 88mm. A wider range is recommended,

but not too large. The weight of a tomato depends on
its diameter and type. Most tomatoes contain between
50 and 280 grams. Exceptions can be as high as 450
grams[8]. This means that the paper teeth should be
able to carry at least 500 grams. Slightly
recommended. The maximum weight of the fixture
and tomato should not exceed 4.5 kg as it is within
the range of the manipulator.

2.2.1 Fixture Specifications

The fixture we designed will meet the following

specifications:
Grasp tomatoes with a diameter of 60mm to 80mm. Figure 1. The design of the manipulator consists of
Load bearing 500g three links, three joints, a parallel plate fixture, a
It will be waterproof. rotary table, an acrylic disc and a rectangular
The tomato will not be damaged in any way. platform[10]. The chain link is a pair of two
It should be durable rectangular beam structures connected by supports.
It can remove tomatoes from plants, There are two brackets for the U-shape. The
Flexible movement rectangular platform is used to stabilize the turntable
Measure when picking tomatoes. The pressure sensor and support the whole structure. The acrylic disc is
and servo system are managed by the robot or the used to mount the turntable to the platform.
external electronic board which only manages the
motion and sensors of the servo system. The joints in the chain of motion between the
manipulator and the hand or tool are called
mechanical wrists. Depending on the application, the
3 Mechanical Design of Tomato Harvesting Robot wrist may have one or more degrees of freedom. The
arm and wrist components of the robot are used to
3.1 System Overview position the end effector. In the end, it's the effector
that really does the job. It must grasp, lift and
To create an overview, the robot is divided into manipulate the workpiece without damaging it or
different subsystems. In any production sector, there loosening it.
is a growing need to increase production and revenue
while reducing time and costs. One of the solutions to 3.2.1 Master the Process
this challenge is to develop an automatic system,
which can replace manual tasks when human It can be divided into the following steps:
precision and work cycle are worse than automatic
equipment. Harvesting beautiful tomatoes takes a lot I. Approach the object: this step is to place the
of time and manpower, so the cost is very high. gripper near the object.
Therefore, it will be a great leap for the agricultural II. Contact: at this time, the clamping claw contacts
industry to develop an automated system that can with the workpiece. In the case of non-contact
perform personalized collection, harvest only processing, the workpiece is within the force field of
tomatoes with the required conditions by using the the gripper.
selective strategy, and provide a system that can work III. Increasing force: the force increment should be
24 hours a day[9]. The vision system used in the within a certain range, so that the pre installed parts
automatic harvester is designed to detect tomatoes will neither be damaged nor slide out of the claw.
and provide information, tomato fruit size and other IV. Fixed object: when enough force is applied, the
parameters (such as ripening stage) to the robot total degree of freedom of the object is removed, and
controller. The main purpose of this paper is to the object is fixed stop moving independently of the
introduce the research carried out in order to evaluate gripper.
the latest technology level of automatic system V. Move object: this step involves moving the object
development, which aims to detect, identify and to the desired location. vi 。 Release objects: at the
locate mature tomato fruits without manual help or macro level, objects are usually released by gravity.
guidance.
3.2.2 Section Modulus
3.2 Three Degree of Freedom Design Mechanism
and calculation The manipulator consists of three joints. The
lumbar joint is represented by the rotation of the
The conceptual design process consists of several turntable, the shoulder joint is represented by the
stages that lead to the final design, as shown in rotation of the connecting rod (1), and the elbow joint

4
GSJ© 2022
www.globalscientificjournal.com
GSJ: Volume 10, Issue 1, January 2022
ISSN 2320-9186 373

is represented by the rotation of the connecting rod provide maximum torque of 0.941n. M and 2.422n.
(2). Each joint uses a specific servo motor to provide M respectively.
the torque required for movement. Each motor
receives shaft position control signal from matlab
code through Arduino microcontroller. The Cartesian
position of the object to be picked is detected by
using the camera and the available matlab plug-in,
which is identified by its color.

1 Configuration of free body figure fully extended in

s robot arm
c (1)
I am the second moment in the region, and C is the W1 W2
maximum distance from the neutral axis (mm). Where and represent the weights of links (1)
As can be seen from the generated graph, for any W
value of cross-sectional area, two and (2), respectively m 3 . It's the weight of the

The s value of the parallel rectangular beam is the elbow muscle. Represents the total weight of the
highest, followed by the hollow cylindrical beam. payload and gripper.
This means that two parallel assembled rectangular
beams have the highest structural strength, 3.3 Recognition of Mature Tomato
considering that they are all made of the same
material with the weight ratio of the three links[11]. The goal of the recognition system is to find all
Therefore, the rectangular beam is used for the link of the tomatoes in the plant and know which are ripe
the manipulator and which are not. Therefore, it is very important to
understand the characteristics such as the color, shape,
type, weight and water ratio of tomatoes. The color of
the tomato is a good way to know if the tomato is
harvested. However, when tomatoes are fully ripe,
they must be eaten within a few days. As a result,
tomatoes are collected when they continue to ripen
during transportation, and the flavor of tomatoes is
less than that of tomatoes that have been preserved
for a longer time. One of the main difficulties in
using color as a visual discrimination parameter is
that it may suffer from uncontrolled or changing
lighting and shadows. There are many different
varieties of tomatoes and many different shapes[12].
The relationship between cross-sectional area and So let's focus on the round tomatoes. Weight is
cross-sectional area for three beams considered another factor in determining whether a tomato is ripe.
As tomatoes grow, their size increases with their
In order to determine the most appropriate weight. If the volume and density of a tomato are
cross-sectional shape of the link, three possible known, its weight can be easily determined.
geometric shapes are considered. Solid cylindrical
beam, hollow cylindrical beam and a group of two
rectangular beams. The section modulus s is a
geometric property that represents the strength based
on the shape or area of the cross section of the beam.

3.2.3 Static Torque

In order to ensure that the torque of the elbow and Fig 4: tomato detection
shoulder motors is sufficient to rotate the joint, the
maximum payload is 200g, the following calculation
is carried out. The figure shows the number of 3.3.1 YOLO Tomato Classifier
maximum required torque corresponding to the worst
case of full arm extension. From the moment of Mature tomato recognition is also a key part of
equilibrium equation, equation. The results show that the robot. In this study, we used machine learning to
the elbow needs at least 0.351n. M and the shoulder train neural networks to distinguish mature tomatoes
needs at least 1.259n. M. Therefore, according to the from immature tomatoes and other similar fruits. As a
servo motors available in the market, the servo tomato classifier, we use YOLO We developed three
motors selected for elbow joint and shoulder joint main filters to distinguish tomatoes. The first filter

5
GSJ© 2022
www.globalscientificjournal.com
GSJ: Volume 10, Issue 1, January 2022
ISSN 2320-9186 374

detects the shape of the object, the second filter inflate the air bag, while the DC motor drives the
detects the color or (RGB) filter, and the third filter is clamp to clamp the stem properly.
machine learning. This is necessary because there are The pressure sensor can detect the pressure
many fruits or objects that may be very similar to between the fruit and the air bag under the premise of
tomatoes, such as apples or oranges. In addition, it reliably clamping the fruit, so as to ensure that there
should be possible to calculate the distance between is no damage on the fruit. Until it is stable, then drive
the camera and the tomato using the classification the double acting cylinder to shrink the sleeve to cut
camera. For this technique, we use a measurement and pick the tomato. Our harvesting tools are able to
method for a single camera by using the focal length grab tomatoes with suction cups and then use
of the camera. swinging blades. Natural changes in crop size, shape
and direction make it challenging and reliable to
select a single end effector pose to grab and cut each
tomato at the same time. In order to overcome this
difficulty, a key feature of our harvesting tool design
is the passive decoupling mechanism, which allows
sequential grasping and cutting operations. The
decoupling mechanism is a kind of flexible strip,
which is fixed on the main body of the end effector
together with the suction cup. The suction cup is also
magnetically connected to the lower side of the
cutting blade, allowing the manipulator to guide the
suction cup during the installation phase. After fixing,
lift the cutting blade to separate the suction cup from
the cutting blade. The suction cup is then attached to
the end effector only by a flexible piece, allowing the
cutting blade to move independently of the suction
Fig 5: fixture calibration (a) geometric formula (b) cup during the cutting operation.

Here are (o) opposite, (H) hypotenuse and (a)

adjacent. The relative distance is fixed, and the
adjacent distance is provided with the camera. Based 4 Motion Analysis
on this formula, we can calibrate the position of the
gripper to grasp accurately. For kinematic analysis, the coordinate frame 23
has been rigidly attached to each link.
As shown in the figure, the initial frame is connected
to the base of the manipulator,
According to the D-H coordinate system

Fig 6: tomato test in real environment

The camera measures the distance and detects the

ripe tomatoes[13]. The experiment has been carried
out in a tomato greenhouse in Kawasaki, Japan. In
this study, we used a 2 megapixel resolution duct
mirror camera.

3.4 End Effector

The end effector is mainly composed of frttom of

the sleeve to detect whether the fruit is in the sleeve.
When the fruit enters the sleeve, the infrared sensor Table 1 DH parameters of various links
will detect the fruit and control the air compressor to

6
GSJ© 2022
www.globalscientificjournal.com
GSJ: Volume 10, Issue 1, January 2022
ISSN 2320-9186 375

T10  A1
link
d i  i a i  i
T20  A1 A2
T30  A1 A2 A3
L1 0
 i a 1
0

T
0
L2 0
 2 a 2
0
3 is a transformation matrix representing the
L3 0
 3 a 3
0 visibility of the end effector relative to the basic
framework
A. Inverse kinematics of manipulator
The figure above shows the method that can be used
for inverse kinematics analysis

Here a i is the length of the ith connection and is

constant, while 
i is the angle of the (i)th
connection, which is variable.
Is the homogeneous transformation moment that

represents the position and direction of

ox yzi i i i

relative to
o x y z i 1 i 1 i 1 i 1

 c i s ic i s is i

a c i
 i
 
Ai   0 i s i  i
s c c c s a s i i i i


i c d i i


In order to determine the required joint angle of the
0 0 0 1  end effector and the desired position (x, y) of the
 manipulator in the work area[15], it is necessary to
(1) know the coordinates of the wrist and the direction of
the end effector
c i  cos i c a i  cos a i Give Way,
and
x y w w Wrist coordinates
 c 1 s 1 0 a1c 1
 
 1
s c 1 0 a1s 1  z is the angle made by the end effector relative to

A I
 0 0 1 0 
the x-axis, and then
  x w
 x  a 3 cos
 0 0 0 1 
y w
 y  a 3 sin 
 c 2 s 2 0 a c 2
 2 y 

  a  tan 1  w
 s 2 c 2 0 a s 2
x 

A2  0

0 1
2
0   w

 
r  a  a  2a a
2
cos 
 0 0 0 1  1 2 1 2 (1)

r x y
2 2 2

TJI  AI 1 AI  2 .....AJ w w
From I<j
β is obtained from equation 1

Give help to
oxyz j j j j
The position and  2
 
direction of is converted to since

r a  2r a 1 cos   a 2
2 2 2
o x y z Transformation matrix
i i i i TJI 1 (2)

If we get γ from equation (2), we can find  1

7
GSJ© 2022
www.globalscientificjournal.com
GSJ: Volume 10, Issue 1, January 2022
ISSN 2320-9186 376

  
1
 a 3c 3s 12  a 3s 3c 12 
 
    
2 3 1 2  a 3c 3c12  a 3s 3s 12 
 0 
4.2 manipulator Jacob j3   0

The upper joint angle is for the elbow down  
configuration  0 
Jacobian matrix is one of the most important tools in  
 1 
robot motion control because it has been developed
joint and end effector speed. Jacobi's law transforms The whole matrix of the whole system is compiled by
the ratio of forces into joint coordinates. Given an Jacobi
N-linked manipulator
Jacobian
j
 j j j  .
1 2 3

j
 j j .... j
1 2 n



ith column in Jacobian of ith link

j i 5 Conclusion
 z  a  o   This paper introduces a tomato harvesting
j   i 1 n i 1
 manipulator for tomato harvesting. In addition, the
i 
 z i  1

 (3) robot design, motion / kinematics formula, tomato
recognition system and robot control architecture are
From the forward kinematics analysis, we have discussed. The proposed manipulator can control
objects less than 200 grams, and the slender part of
0  a 1c1 the robot achieves high rigidity in the harvesting
   
o o  0 o1  a1s1 process[16]. In terms of speed, the proposed
manipulator is slower than the commercially
1   0  available prototype, but the main advantages of the
 a 1c1  a 2c12  a 3c 3c12  a 3s 3s 12  robot are accessibility and selectivity. Experiments
  have proved the ability of the manipulator to work in
o 2   a1s1  a 2s12  a 3c 3s12  a 3s 3s12  a very narrow environment. In addition, the proposed
claw simplifies the harvesting process by passively
 0 
cutting the stem.
Now put the above vector into equation (3) above and As a plan, we plan to use machine learning for
evaluate the Jacobian matrix of different links as manipulator control. According to the data obtained,
 a 1s 1  a 2 s 12  a 3c 3s 12  a 3s 3c12  the manipulator spends more time to avoid obstacles
  and grasp tomatoes. Therefore, by using machine
 a 1c1  a 2c12  a 3c 3c12  a 3s 3s 12  learning tools, we can shorten the harvesting time and
  achieve the best continuous robot shape, so as to
0
j1   0
 achieve rapid harvesting. The robot is equipped with
a gripper and measuring tools. The vision system can
 
  accurately determine the coordinates, but the tomato
0 must be suspended at a fixed distance because a
 
 1  single camera cannot see the depth. Because of the
low resolution of 3dcamera, the precision of
 a 1s 1  a 3c 3c12  a 3s 3s 12  reference point is low, but it can determine three
  coordinates. The gripper can grab tomatoes and
 a 2c12  a 3c 3c12  a 3s 3s 12  transport them to the desired destination. The robot
 0  can prevent the tomato from being damaged when a
j2   0
 pressure sensor is installed in the finger of the gripper.
  Though, tomatoes have to be regular in shape[17]. In
 0  the development process of harvesting robot, the
  knowledge of controlling robot arm has been greatly
 1  improved, and the processing technology has been
developed.

8
GSJ© 2022
www.globalscientificjournal.com
 GSJ: Volume 10, Issue 1, January 2022
ISSN 2320-9186 377

& Aldon, M. J. (1987). MAGALI – A Self Propelled

REFERENCES Robot to Pick Apples. ASAE Paper, No. 871037.

[1] R. Hussin*, M. RizonJuhari, Ng Wei Kang, R.C.Ismail, [14] Edan, Y., (1999). Food and agriculture robotic.
A.Kamarudin ―Digital ImageProcessing Techniques for Handbook of Industrial Robotic, Second edition. Chapter
Object Detection From Complex Background 60, pp. 1143-1155
Image‖International Symposium on Robotics and
Intelligent Sensors 2012 (IRIS 2012). [15] C. Ji, J. Zhang, T. Yuan, W. Li Research on key
technology of truss tomato harvesting robot in greenhouse
[2] Dameshwari Sahu, Ravindra Manohar Potdar. "Defect
Identification and MaturityDetection of Mango Fruits [16] Schertz, C.E. and Brown, G.K. (1968). Basic
Using Image Analysis."American Journal of Artificial considerations in mechanizing citrus harvest. Transaction
Intelligence. Vol. 1, No. 1, 2017, pp. 5-14. DOI: of ASAE. 343-346.
10.11648/j.ajai.20170101.12.
[17] Zion, M. Mann, D. Levin, A. Shilo, D. Rubinstein, I.
[3] Rahul Gautam, Ankush Gedam, Ashish Zade, and Ajay Shmulevich Harvested-order planning for a multiarm
Mahawadiwar ―Review onDevelopment of Industrial robotic harvester Comput. Electr. Agric., 103 (2014), pp.
Robotic Arm‖ International Research Journal of 75-81.
Engineeringand Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 03 | Mar -2017www.irjet.net p-ISSN:
2395-0072.

[4] Srinath, A., Rao, C.S. and Jagadeesh, A., 2011. Genetic
algorithm based Hammingnumber technique for detection
of isomorphism in planetary gear trains. Int. J. of Applied
Engineering Research,6 (1), pp.33-42.

[5] Surekha, B., Vundavilli, P.R., Parappagoudar, M.B. and

Srinath, A., 2011. Design ofgenetic fuzzy system for
forward and reverse mapping of green sand mouldsystem.
International Journal of Cast Metals Research,24(1),
pp.53-64.

[6] Ibrahim, M. M., Amin, E., Farag, A.“Developing a

Multi Purpose Digger forHarvesting Root Crops”. Mis J.
Ag. Eng., 25 (4):1225-1239, 2008.

[7] Munde, P.A Nadre, R.G., Sawant B.P.“Performance

evaluation of bullock drawngroundnut digger suitable for
MAU multipurpose tool carrier”. International journal
ofAgricultural Engineering, 2 (1): 157-159, 2009.

[8] Zhengzhou, G. “Peanut harvester, Peanutgroundnut

harvester”. P: 1-10. Machinery CO.LTD, 2010.

[9] Kondo, N., Monta, M., Shibano, Y., Mohri, K.,

Yamashita, J. & Fujiura, T. (1992). Agricultural Robots (2):
Manipulators and Fruits Harvesting Hands. In ASAE
Paper, No. 923518.

[10] Monta, M., Kondo, N., Shibano, Y., Mohri, K.,

Yamashita, J. & Fujiura, T. (1992). Agricultural Robots (3):
Grape Berry Thinning Hand. In ASAE Paper, No. 923519.

[11] Okamoto, T., Shirai, Y., Fujiura, T. & Kondo, N.

(1992). Intelligent Robotics. Jikkyo Shuppan, Japan:
Tokyo.

[12] Ting, K. C., Giacomelli, G. A. & Fang, W. (1993).

Decision Support System for Single Truss Tomato
Production. Proceedings XXV CIOSTACIGR V Congress:
70–76.

[13] D'Esnon, A. G., Rabatel, G., Pellenc, R., Journeau, A.

GSJ©9 2022
www.globalscientificjournal.com