Autonomous Robotic System for Pumpkin Harvesting

Round 1

Reviewer 1 Report

The presented work shows the results that will be used as future research in the field of automation in agriculture harvesting. The paper is well organized and contains all the necessary information to sustain the results of the research

Author Response

Reviewer 1

The presented work shows the results that will be used as future research in the field of automation in agriculture harvesting. The paper is well organized and contains all the necessary information to sustain the results of the research.

Answer: Thank you so much for your encouragement and kind response.

Reviewer 2 Report

For automatic pumpkin harvesting, the prototype of picking robot was developed, and the performance of picking robot arm was evaluated by experiments.

It is a very interesting research and has application value. However, the paper still has the following problems:

As described in the title of the paper, the topic is about the automatic harvesting robot system, however, it should include target recognition unit, manipulator, vehicle, end-effector and controller. But the performance of the manipulator is mainly tested in this paper, so I personally think that the title of the paper needs to be consistent with the content.

As stated by the authors, harvesting heavy fruits requires a powerful mechanical arm, so the first thing to evaluate is the load performance of the mechanical arm.

Compared with small fruits, the positioning resolution and positioning accuracy of the robot arm are not the most important.

The main problem of this paper lies in the experiment. The actual measured values and the deviation from the target values, how they obtained need to be introduced in detail.

For other comments, please refer to the notes in the manuscript.

Author Response

Reviewer 2

-For automatic pumpkin harvesting, the prototype of picking robot was developed, and the performance of picking robot arm was evaluated by experiments.

-It is a very interesting research and has application value. However, the paper still has the following problems:

-As described in the title of the paper, the topic is about the automatic harvesting robot system, however, it should include target recognition unit, manipulator, vehicle, end-effector and controller. But the performance of the manipulator is mainly tested in this paper, so I personally think that the title of the paper needs to be consistent with the content.

Answer: Thank you so much for your valuable comment. As you already mentioned, a harvesting system for agriculture industry includes many necessary components including platform (vehicle), manipulation system (mostly robotic arm plus its end effector), controlling unit, and sensors (recognition, safety, positioning, attitude, …). The presented paper is a part of a progressive and continuous research about autonomous vehicles in which cited as “Roshanianfard, A.; Noguchi, N.; Okamoto, H.; Ishii, K. A review of autonomous agricultural vehicles (The experience of Hokkaido University). Journal of Terramechanics” in the paper. Please check figure. 21 of the cited paper. Based on this logic and also because most components and their progresses presented in previous paper, we tried to focus on the part of systems which is directly related on harvesting. The details and related paper mentioned in 111 to 139.

-As stated by the authors, harvesting heavy fruits requires a powerful mechanical arm, so the first thing to evaluate is the load performance of the mechanical arm.

Answer: Thank you for your valuable comment. The torque calculations, evaluation, and optimization described in the paper “Roshanianfard, A.; Noguchi, N. Kinematics analysis and simulation of a 5DOF articulated robotic arm applied ‎to ‎heavy products harvesting. Tarim Bilimleri Dergisi-Journal of Agricultural Sciences” as cited in line 76 ~ 78. The desired payload was 25 Kg (with FOS of 2) as described in the paper “Roshanianfard, A.; Noguchi, N.; Kamata, T. Design and performance of a robotic arm for farm use. International Journal of Agricultural and Biological Engineering (IJABE)” as cited in line 113. More clarification and desciptions included in line 76 ~ 80.

-Compared with small fruits, the positioning resolution and positioning accuracy of the robot arm are not the most important.

Answer: Thank you so much. The position resolution is not the first priority but for harvesting, the minimum resolution is required which described in table 8.

-The main problem of this paper lies in the experiment. The actual measured values and the deviation from the target values, how they obtained need to be introduced in detail.

Answer: Sorry the question was not clear for us. Could you please describe it more?

-For other comments, please refer to the notes in the manuscript.

-These review is Not very relevant to the topic of this article. The harvesting robots existed fot heavy fruit should be focused more:

“GPS/INS system consisting of 3 gyroscopes, two inclinometers, and a low-cost GPS that provides orientation and roll/pitch inclinations. The designed system had a heading, roll, and pitch accuracy of 0.64, 0.43, and 0.61 degrees, respectively [3]. As the accuracies were bigger than RTK-GPS and IMU systems, the researcher recommended using them in a low-speed and low-cost application [4]. Then, Liu, et al. [5] developed a low-cost IMU using sensor fusion to estimate attitude angle, which had errors of 1.35 and 0.73 degrees in roll and pitch directions, respectively. Later, real-time motion detection using a three- dimensional camera [6]; using omnidirectional stereo vision [7], and specific predeter- mined point recognition [8] were applied to previously developed systems to increase their efficiency and application. After all modifications, Noguchi and Barawid [9] presented a multiple RTs system. This system included a paddy fielder robot, an RT, and a robotic combine harvester. They considered the safety, economic parameters, management, and real-time monitoring aspects. After that, Zhang and Noguchi [10] developed a multi-robot tractor for fieldwork to reduce work time and improve efficiency. The system could follow several predetermined patterns with a circular and rectangular safety zone. The simulation results indicated that the efficiency of three and seven robots varied between 59.4% and 89.8%. They concluded that the efficiency could increase to 85% in the large fields. The developed ARs in the laboratory of Vehicle Robotics – Hokkaido University (VeBots) started with the development of a path planning system (1997);”

Answer: Thank you so much. The revision was applied.

-According to “A real-time controlling system includes (1) the central controller located in the controlling station, (2) ECU of RTHRHC + PC, (3) ECU of RTcarrier + PC, and (4) controlling unit of manipulation system based on PLC system. The controlling unit as a compact circuit consists of five servo motors and amplifiers, a position board installed on a PC, a controlling program, and optical cables for data transfer. The PLC system was pow-ered by 200ACV, which was generated by a gasoline generator. Servo motors com-mand transferred to the position board (connected to the PC by a PCI Express protocol) via optical cables. The controlling program was written using C++. ‎The mathematical equations and related algorithms were calculated and designed using the D-H method [38] because of its simplicity, minimum response time during operation, and good changeability during experiments.” How the pumpkins is identified and located? Is there a visual unit in the robot?

Answer: Thank you so much. The revision applied was in line 121 ~ 124.

 

-According to “The results indicated that the , , and  of the designed system was ,  , and , respectively. Nevertheless, after development, these parameters were reduced to , , and  of designed parameters, respectively, which were , , and , respectively.” How these values were measured?

Answer: This is very accurate and good question. Excess description of this section is not included due to the need for brevity and compliance with standards. The evaluation has done using the following steps:

1. Several random points selected inside the working space (in this paper 11 points inside harvesting possibility zone).
2. A red laser pointer attached exactly perpendicular to the ground in the (0, 0) coordination of the EE.
3. The target positing determined on ground using the scaled papers with tolerance of 1mm.
4. In GPn of each test, was stopped and the determined position (by laser pointer) was highlighted on the paper. To assess the parameters, they measured after a complete stop of the EE from the previous pose.
5. After experimentation finish, the papers collected.
6. Photos were taken from the sheets.
7. Photos were imported in the Solidwork softer to evaluate the parameters with high accuracy. The accuracy was measured by the error between the target position and the obtained position. For evaluation. Geometrically, it was the maximum errors distributed inside the reference frame. And also, the repeatability was the ability to achieve the repetition of a position. For evaluation, it is hitting one position repeatedly. Geometrically, it was the radius of the smallest sphere that encompasses all the positions reached for the same requested position.

For more information, please check ANSI/RIA R15.05.

 

-According to “After this stage, the system was installed on another RT (stage-IV), and the modification was applied. In this stage, the , , and  increased by , , and  in comparing by previous development, respectively, which was , , and , of desired parameters of the designed system, respectively. Why they increased in another RT?

Answer: Explained in lines 246 ~ 249 and lines 253 ~ 256.

 

-the experiment should be introduced more, how the resolution was measured?

Answer: As mentioned in two comments before. The explanation can damage the brevity of paper. If needed, we can include.

-according to “there was no significant differences indicated. However, the resolution tolerance de-creased due to mobile platforms, but this tolerance had no significant effects on the general performance of the designed system” How the arm was controlled during the test? Automatic or manual command control?

Answer: The RA was maneuver automatically during all operations. Just the position of pumpkin import at the beginning of test.

-“ A resolution of 5 mm is acceptable for an actual agricultural field. The presented robotic system is more accurate than the required and desired indicator” For a big target, if so small resolution is important?

Answer: Absolutely a small resolution is not the main propriety. But, the robot can help with other harvesting applications such as cabbage, melon, … harvesting which will need more accurate operations.

-According to “As results show, in stage II, the  was 10.91mm in the x-direction, 9.52 mm in the y-direction, and  was 12.74 mm (Figure 9). The in x and y directions were belonged to point-2 by 2.55 mm, and point-9 by 0.83 mm, respectively. The  belonged to point-4 by 8.1 mm. The  of points 4, 6, 7, and 9 were more than the , and the of points 1, 2, 4, and 7 were more than the .” how the errors was measured?

Answer: Thanks for your comment. It has been mentioned in the previous comments.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper focused on the development of the unique and unified system including a mobile platform, a manipulation system, an end-effector, and its integrated algorithm for pumpkin, which is very significant to efficient automatic harvest heavy-weight crops. Since the system the paper proposed can increase the speed and improve the process of harvesting because it has a precise robotic manipulation and end-effector, and a programmable controlling system that can work autonomously, what is the scope of application of this system? Is the evaluation system or algorithm specific to the robot in the paper or can it be applied to the relevant harvesting heavy-weight crops robot? Please give a discussion or explanation.

Author Response

Reviewer 3

The paper focused on the development of the unique and unified system including a mobile platform, a manipulation system, an end-effector, and its integrated algorithm for pumpkin, which is very significant to efficient automatic harvest heavy-weight crops. Since the system the paper proposed can increase the speed and improve the process of harvesting because it has a precise robotic manipulation and end-effector, and a programmable controlling system that can work autonomously, what is the scope of application of this system? Is the evaluation system or algorithm specific to the robot in the paper or can it be applied to the relevant harvesting heavy-weight crops robot? Please give a discussion or explanation.

Answer: Thank you so much for your accurate and smart question. The system includes several components, each of which can have a different scope of application as follow:

The mobile platform: The RTs are commercialized tractors which can maneuver autonomously. The RTs can have various applications for carrying objects, harvesting, plowing, seeding, cultivation, and most of the farm applications.

Robotic arm: The developed robotic arm is mostly designed for farm applications which is not so accurate for very precise application such as car production line, or circuit assembly. But it obviously is a practical system for farm application, carrying objects, horticulture application, and so on.

End effector: The designed end-effector is specifically designed for pumpkin harvesting. By changing the fingers, it can be applied for many more objects including agricultural products.

Controlling system: This unit can have many applications in many industries with some minimum modifications.

Author Response File: Author Response.pdf