Source: PADMA AGROBOTICS LLC submitted to
AUTONOMOUS HARVESTER FOR CILANTRO WITH BUNCHING AND TYING CAPABILITY
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
SMALL BUSINESS GRANT
Reporting Frequency
Annual
Accession No.
1030602
Grant No.
2023-33530-39846
Cumulative Award Amt.
$181,500.00
Proposal No.
2023-00937
Multistate No.
(N/A)
Project Start Date
Jul 1, 2023
Project End Date
Feb 29, 2024
Grant Year
2023
Program Code
[8.13]- Plant Production and Protection-Engineering
Project Director
Brauer, C.
Recipient Organization
PADMA AGROBOTICS LLC
3795 S WINDSTREAM PL
CHANDLER,AZ 85286
Performing Department
(N/A)
Non Technical Summary
This project aims to develop an automated robotic harvester for cilantro and other specialty crops that are harvested in bunches. Currently, this is a labor-intensive task that takes up to 50% of total costs. Automating this task would mean reducing OPEX for farms by 40%-50% and, more importantly, tackling the huge labor shortage currently affecting the agricultural sector. This will permit to grow produce that is nowadays economically unviable due high labor costs, thus reducing the need to import cilantro and other specialty crops and making US internationally more competitive.Phase I will focus on developing a proof-of-concept prototype and test the operation with real cilantro in an iterative design process. The principal technical challenge is the stem collection and bunching system, which requires complex and precise mechatronics to accurately detect, cut, extract, transfer and bunch the cilantro stems into a precise size, which are later tied. Both the tying system and the autonomous moving platform that traverses the cilantro beds will represent a lesser challenge, as the company has already developed working prototypes that only need to be adapted.To our knowledge, no such system exists in the market, in spite of the growing need, representing a high-risk endeavor.This proposal addresses the Topic Area 8.13 "Plant Production and Protection (Engineering)", specifically its Research Priority 1. "Improved Crop Production Methods or Strategies", and its topics:"Technologies that enhance commercial horticulture production","Production, harvesting, and postharvest handling of specialty crops", and"Cyber-physical systems to support precision agriculture".
Animal Health Component
80%
Research Effort Categories
Basic
10%
Applied
80%
Developmental
10%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
40214302020100%
Knowledge Area
402 - Engineering Systems and Equipment;

Subject Of Investigation
1430 - Greens and leafy vegetables;

Field Of Science
2020 - Engineering;
Goals / Objectives
The overall objective of the research described in this Phase I proposal is to demonstrate the feasibility of the robotic mechanical system for detecting, cutting, collecting, bunching and tying stems of cilantro and other specialty crops. Padma AgRobotics has set the following specific objectives of the Phase I work plan to demonstrate the feasibility of the described system:O1. Construct a testing station to allow for controlled experimentation and design iterations.O2. Develop a bunching system that meets demanding performance and reliability criteria.O3. Improve the current version of the tying station and integrate it with the bunching system. The combined system must meet demanding performance and reliability criteria.O4. Integrate the combined bunching and tying system with a mobile robotic platform to enable field testing with real cilantro. Padma AgRobotics will perform field testing to improve the system as needed until obtaining the desired outcome.Padma AgRobotics will address the following questions to establish the technical feasibility of the proposed approach: (1) Can the system bunch > 99% of plants in an organized and neat manner?; (2) Can the system work at a speed of tying one bunch every three seconds?; (3) Can the system tie cilantro bunches that fall within a specified diameter tolerance > 99% of the time?; (4) Can the system prevent or reliably clear plant debris buildup that might occur during the operation?; (5) Can the system reliably work non-stop for 8 hours in the field?; (6) Can the system reliably deal with loose soil, dust and inclement weather?; and (7) Can the system accommodate imperfect plant inputs, including gaps in the plant row and plants in a range of heights and orientations?
Project Methods
TASK 1 - Testing Station Construction. The team will construct a testing platform able to move a mounted mechanism at a consistent velocity over a stationary row of plants. This will provide a controlled and repeatable test setup for the mechanisms developed in tasks 2 and 3 to facilitate the ease of design iterations. The platform will consist of a simple motorized cart that operates at ground level, or a rail-mounted carriage affixed to a tabletop or raised planter. The stationary row of plants will consist of real cilantro plants on a test bed or a suitable cilantro substitute, previously grown by Maricopa Ag. Center. Overall, this task will only require relatively basic mechanical engineering to build the moving cart, its moving structure (ground or rails) and an aligning means to ensure the correct positioning along the plant lines.KPIs: (1) Consistent speed at any setpoint between 0-3 miles per hour; (2) Able to carry mounted mechanical systems of up to 50 pounds in weight.TASK 2 - BUNCHING SYSTEM DEVELOPMENT. The team will develop a bunching system that collects the plants at ground level and outputs separated bunches of a consistent size. The anticipated primary process steps are cutting, separating, and accumulating stems. Each of these steps requires the development of a corresponding mechanism that is reliable and robust. Additional minor process steps for transporting and reorienting plants may be added to any of these mechanisms.2.1. Cutting Device. The team will develop a cutting device that accepts an incoming row of plants and cuts them off at ground level. The cutting mechanism must generate a clean cut, accommodate varying densities of plants, and not throw any soil into the plants or subsequent devices. The team has identified two viable cutting approaches: a band-saw for all size rows, and two counter-rotating discs for each row. The team will design and build a version of each cutting mechanism, mount it on the system and test it on cilantro grown on the ground for efficiency and the cleanness of the cut. Any way, the cutting system should not be a major challenge, since the team has already developed and implemented similar systems for other agricultural robots.2.2. Extraction and Separation Device. The team will develop the device for extracting and separating individual or small groups of plant stalks, with plants at various orientations. Extensive research and design iteration is needed regarding suitable approaches to this task. The team will design a conveyor belt system (food grade) that will be able to cleanly transfer the collected cilantro stems to the accumulation station. Based on the efficiency of the accumulation subsystem, this extraction and separation device can be a two-belt system each running at different speeds or a single belt system. The team will build prototypes of both designs and test them to identify the best option in terms of efficiency of transfer and damage minimization.2.3. Accumulation Device. The team will develop the device that accumulates plant stalks until the desired quantity is achieved in a bunch, then inserts a separator between these stalks and any other incoming stalks. This device must include an appropriate sensing mechanism to ensure consistent bunch size. The team has defined two theoretically valid designs: a funnel system that is able to collect the bunch precisely and release it after a bunch size is reached; and a two-belt system with different speeds that will control the movement of the cilantro and thus create the bunch. Both concepts will be developed and experimented with in the testing station, to clarify which alternative provides better results in terms of accumulation precision and consistency, as well as in terms of damages eventually inflicted to the collected plants.KPIs: (1) Able to collect rows of plants with a left-right position variation of +/-1 inch; (2) Able to accumulate bunches within the desired size tolerance 99% of the time; (3) Able to retain at least 99% of the collected cilantro stems within bunches and with no observable damage; (4) No observable soil or plant debris buildup.TASK 3 - TYING SYSTEM INTEGRATION. The team will integrate a system to affix ties around each bunch created by the bunching system. Overall, the technical complexity of these developments will be relatively low.3.1. Update Tying Station. The tying system will be largely based on technology previously developed by Padma AgRobotics. The team will update the existing tying station as needed to accommodate the bunching system. This may include changing the system's dimensions, the tie channel's axis of movement, or the types of actuators used. The team will explore these options through multiple experiments, seeking to achieve a tying system that is effective, uses unexpensive materials and minimizes energy consumption, while tying the stems with no damage.3.2. Integrate Tying Station with Bunching System. The updated tying station will be integrated to the bunching system in a position that adequately installs ties around plants captured in the accumulation area and allow plants to enter and exit the area while the tie channel is retracted.KPIs: (1) Able to create a tied bunch every 3 seconds; (2) No observable damage to plants when securing ties; (3) Able to operate continuously with no observable buildup of tie trimmings or plant debris.TASK 4 - Field Testing. The team will integrate the prototype bunching and tying system in a mobile robot chassis and conduct field testing to test the system's harvesting capabilities and long-term durability.4.1. Integrate Bunching & Tying System with Mobile Robot. The team will integrate the bunching and tying prototypes to the moving platform built in Task 1, in a manner that accommodates for variable terrain height. This will require only relatively basic mechanical engineering, such as the incorporation of additional joints, mounting linkages, or tracking wheels.4.2. Capability Field Testing. The team will use 2 acres of cilantro on 84-inch beds, with 6-8 rows per bed and each row being at least 100 feet long. The cilantro beds will be readily available thanks to the services of Maricopa Ag. Center, which will grow with enough anticipation such plants. The testing station developed in Task 1 will be implemented on these beds and the prototype developed in Tasks 2 and 3 will be run on such station. The team will build one arm to test the operation on all rows. Multiple experiments will be conducted, simulating different conditions such as dirt, water, and other debris introduced at various points in the system. Also, the rows of plants will be modified to represent different possible conditions including inconsistent left-right positioning, gaps in the plant row, and variations in the plant height and orientation. Data collection will focus on the operation speed and observed points of debris buildup.4.3. Durability Field Testing. Along with the capability field testing, the team will subject the prototype to durability testing, covering 8+ hours of operation and multiple acres of plants. Throughout this testing, the team will monitor the accuracy of the generated bundle sizes and the percentage of plants lost or damaged. As well, the team will inspect the system periodically for wear points and heat buildup in components.KPIs: (1) Able to handle conditions representative of inclement weather; (2) Able to collect rows of plants with a left-right position variation of +/-1 inch; (3) Able to create a tied bunch every 3 seconds; (4) No observable soil or plant debris buildup in cilantro cut, bunched and tied when running on a row of cilantro at least 100 feet long; (5) Able to operate continuously for 8 hours in the field; (6) Able to accumulate bunches within the desired size tolerance 99% of the time; (7) Able to retain at least 99% of the collected cilantro stems within bunches and with no observable damage?.

Progress 07/01/23 to 02/29/24

Outputs
Target Audience:Agricultural sector Changes/Problems:The major change in approach during the project was the reevaluation and redesign of the harvesting sequence. Originally, the project had the following sequence: cutting the cilantro from ground, transporting the cut cilantro, bunching the cilantro and finally tie the cilantro bunch. However, However, significant challenges arose during the integration of these steps, leading to dropped or stuck plants between mechanical components and resulting in unsuccessful attempts to produce clean tied bunches of cilantros. To address these challenges, the team made the decision to radically change the approach. Instead of cutting the plants first, they redesigned the system to bunch and tie the cilantro while still on the ground, before cutting and transporting them. This required a completely new bunching and tying design. The main reason for this change was to make the plants more controllable since they were rooted to the ground, resulting in successful outcomes in testing. This change in approach was necessary to overcome technical hurdles and achieve the desired outcomes of the project. By adapting the sequence of operations and redesigning the system accordingly, the team was able to address the challenges and make significant progress towards developing a functional automated cilantro harvesting system. In terms of reporting requirements specified in the award Terms and Conditions, no special or additional reporting requirements were mentioned in the provided project description. However, it's important to note that ongoing communication and reporting to the funding agency, USDA, would likely be required to provide updates on project progress, milestones achieved, challenges encountered, and any deviations from the original project plan. This ensures transparency, accountability, and alignment with the project objectives outlined in the award agreement. What opportunities for training and professional development has the project provided?The project has provided several opportunities for training and professional development for the team involved: Technical Skill Development: Team members had the opportunity to extensively develop and test new technologies in handling plants, especially specialty crops for harvesting operations. With the challenges encountered during the integration of the various components, the team got an opportunity to build on their mechanical engineering expertise and learn new ways and designs for mechanical design for agricultural machines through the automated cilantro harvesting system. Problem-Solving Skills: There were many challenges where the team had to iterate through the design process to accommodate for the real-world conditions such as, cilantro plants getting stuck in the mechanical components, plants bending over and creating problems while tying the knot etc. The work required a consistently thinking out-of-the-box approach to resolve these issues. The solution that was successful required the team to alter the approach ground up and redesign from scratch. Field Testing Experience: Although the initial testing was done on the testing station in the lab, the metric for success was the field testing done in Cilantro Farms in Phoenix. The hands-on experience allowed them to gain insights into the challenges and requirements of field operations, as well as the opportunity to refine and optimize the system based on the results from the field and from customers who were supervising the machine testing in the field. Project Management Skills: Managing a complex research and development project like this required effective project management skills. There were objectives which were executed in parallel (objective 2 and 3), Team members had the opportunity to develop skills in project planning, scheduling, budgeting, resource allocation, and coordination, which are valuable for future career advancement. Communication and Presentation Skills: The team worked closely with cilantro farms in Phoenix and Yuma in presenting their project updates and showing the results of their testing. This also included demonstrations to effectively communicate project progress and outcomes and providing them with field testing updates and receiving their feedback. Overall, the project has offered a rich environment for training and professional development, equipping team members with valuable skills and experiences that will benefit their careers in robotics, mechanical engineering and agriculture. How have the results been disseminated to communities of interest?The dissemination of project results to communities of interest has likely involved various channels and methods, including: Technical Reports: Detailed technical reports documenting the project objectives, updates and results that have been prepared and distributed to stakeholders, including funding agencies and collaborators (cilantro farms). Demonstrations: As part of the field testing, the collaborators were invited to observe the bunching and tying mechanisms that were developed as part of the project. This provided the opportunity for the stakeholders to interact with the project team, ask questions, and provide feedback. By employing a multi-faceted approach to dissemination, the project can effectively reach and engage communities of interest, ensuring that project results are accessible, understandable, and actionable for relevant stakeholders. What do you plan to do during the next reporting period to accomplish the goals?During the next reporting period, we plan to focus on the following activities to accomplish the project goals: Iterative Design and Testing: We will continue to iterate on the design of the automated cilantro harvesting system based on feedback from Phase I testing. This includes refining the bunching and tying system, improving integration with the mobile robotic platform, and addressing any remaining technical challenges. Field Testing and Validation: We will conduct further field testing of the integrated system to validate its performance under real-world conditions. This will involve testing the system's ability to accurately detect, cut, bunch, and tie cilantro plants in various field environments. Optimization and Refinement: Based on field test results, we will optimize and refine the system to improve its efficiency, reliability, and accuracy. This may involve adjustments to mechanical components, software algorithms, and operational procedures. Documentation and Reporting: We will document all design iterations, test results, and improvements made during the reporting period. This includes preparing technical reports, presentations, and other documentation to communicate project progress and outcomes to stakeholders. Collaboration and Outreach: We will continue to collaborate with industry partners, agricultural experts, and other stakeholders to gather feedback, share knowledge, and ensure alignment with industry needs and standards. Additionally, we will explore opportunities for outreach and dissemination to raise awareness of the project and its potential impact. By focusing on these activities, we aim to advance the development of the automated cilantro harvesting system and move closer to achieving the project goals outlined in Phase I.

Impacts
What was accomplished under these goals? Under the goals outlined, several accomplishments were achieved: Objective 1 - Construct a testing station: The team successfully built a testing platform capable of controlled experimentation and design iterations. This station allowed the team to extensively test multiple design iterations quickly and efficiently before testing on the real cilantro fields. This reduced the need for multiple cilantro beds for testing and the costs substantially. After the bunching and tying mechanism has been tested for consistency and speed on the test station, the mechanism was tested on the cilantro beds. Objective 2 - Develop a bunching system: Padma AgRobotics designed and developed a bunching system capable of gathering plants and splitting them into consistent-sized bunches. The first bunching system design was an integrated system which performed the bunching (objective 2) and tying (objective 3) sequentially in one operation. However, due to integration issues with the transportation issues, the bunching system was redesigned from the ground up (after 7 design iterations). The final design worked effectively to bunching and tie and was successfully tested in the cilantro farms with a new approach of bunching and tying the plants while they were still on the ground. Objective 3 - Improve tying station and integrate with bunching system: The team worked on improving the tying station and integrating it with the bunching system. As with bunching system (objective 2), this tying system was redesigned from ground up to accommodate for the issues, such as plants getting stuck or dropped between mechanical components. After multiple design iterations, the team successfully designed a bunching and tying system that was successfully tested on the cilantro farm. Additionally, the project addressed key questions to establish the technical feasibility of the proposed approach, including the system's ability to bunch plants accurately, work at a specified speed and environmental conditions. With regards to accommodating variations in plant inputs and handling plant debris, Padma is confident that that the proposed mechanical updates to the system in Phase 2 will help move the development forward. Overall, Padma AgRobotics demonstrated significant progress in developing an automated cilantro harvesting system that met demanding performance and reliability criteria, paving the way for further optimization and eventual commercialization in Phase II.

Publications