Progress 09/01/23 to 08/31/24
Outputs
Target Audience:The target audience of our research is mushroom farms who are interested in improving productivity and profit through automating the mushroom harvesting process. Changes/Problems:The development of the mushroom harvesting system has encountered some delays due to necessary changes in the robotic system structure to enhance productivity and reliability. As we advance, we have discovered more technical challenges than initially anticipated. While we remain confident in our ability to develop a functional system, we recognize that transitioning from laboratory technology to field application will require significantly more effort. To bring the technology to real-world adoption, we will need to pursue additional funding. What opportunities for training and professional development has the project provided?This USDA project has significantly contributed to recruiting and training undergraduate senior design students from mechanical engineering programs, capstone students from biotechnology programs and supporting their research projects. It has also facilitated the recruitment of Ph.D. students working on their dissertation in mushroom robot system development and computer vision algorithms. Additionally, we receive an additional UH infrastructure grant to establish a permanent test best at UH for training and educating students and small-scale mushroom producers. Some of the students and their completed projects are given below. Abdollah Zakeri (Doctoral student) - Researching computer vision algorithms for identifying mushrooms. Biru Koirala and Jiming Kang (Doctoral students) - Working on gripper design, mechanical system design, and mechatronic system hardware and software development. Samson Stern and Leo Rodriguez (Undergraduate students) - Worked on mushroom gripper design in summer 2024. Patricia Kurniawan (Undergraduate student, Computer Science) - Worked on robot simulation in summer 2024. Sandesh Risal and Ezra Wari (Doctoral students) - Worked on modeling and simulation of the mushroom substrate system. Both have graduated with Ph.D. degrees this fiscal year. Sandesh was hired as a post-doc under a grant from the Department of Education. How have the results been disseminated to communities of interest?During this funding period we have published both a conference paper and a journal paper. Additionally, we have also submitted another journal paper and an extended abstract to a conference. We are currently in the process of writing five more manuscripts to be published in various peer-reviewed scientific journals. To disseminate our research finding to the mushroom growing community, we have developed a website (https://mushroom.egr.uh.edu/). What do you plan to do during the next reporting period to accomplish the goals?In the third year, we aim to finalize the development of the mushroom robotic harvesting system and proceed to field testing with our industrial partner, Monterey Mushroom.
Impacts
What was accomplished under these goals?
Mushroom Robot Development: We have been working on structural design of the mushroom robot (Figure 1). The robot consists of the following components: 1) a mobile robot chassis capable of navigating complex environment; 2) a lift structure for moving the picking arms vertically; 3) a SCARA robotic arm for positioning the gripper over the mushroom growth tray; 4) a mushroom gripper for picking mushrooms; and 5) computer vision deep learning algorithms and a camera for identifying mushrooms ready for picking. Three doctoral students are involved in this project. Two undergraduate students funded by National Science Foundation REU program assisted with the gripper design and testing in the summer of 2024. Another undergraduate student, funded by the Department of Education REU program, worked on developing the robot and IoT simulation through a web-based user interface. In this cycle, we transitioned from a Cartesian robot arm to a SCARA robot arm because the SCARA's motion is significantly faster, which is essential for achieving productivity levels comparable to human workers. The first set of SCARA's robots, known as Hitbot was acquired in January 2024. However, the Hitbot arm's length was insufficient for reaching the entire the mushroom growth bed. To address this, we designed the robot's end effector with a camera mount and gripper to extend its reach and cover more space (Figure 2). As of now, each component has been developed but still requires improvement and integration into a cohesive system. A system engineering approach is being used to guide this integration process. Significant effort has been invested in designing and testing the mushroom gripper which has undergone numerous revision cycles (Figure 3). Our industrial partner, Monterrey Mushroom, prepared a small mushroom growth tray testbed for our lab experiments. One conference paper about the gripper has been accepted for presentation at ASME IDETC/CIE conference in Washington DC in August 2024. This conference paper has been further refined and accepted by a journal. After several months of programming and testing, we recognized the critical importance of power management. The mobile robot chassis called Reeman can power itself with a battery (and automatic charging) to move around. However, the Hitbot requires 110V power. Since we need an untethered mobile solution, we negotiated with Reeman and Hitbot vendors, who agreed to equip the Reeman robot with a larger battery capable of powering the Hitbot directly. In June 2024, we acquired an integrated set of Reeman (with a large battery) and Hitbot allowing us to test the entire mobile system. Besides the Reeman and Hitbot, there are other motors and controllers that require 24V, 12V, 5V, and 3.3V. Ensuring all components are correctly powered is crucial. A web-based user interface is being designed to visualize and control the whole system, as well as facilitating the system integration process. Using image processing to harvest mushrooms: Image processing and data analysis are emerging technologies used to assess the size and quality of mushrooms, determine the growth rates, and measure yield. We have collected and processed mushroom image data and tested numerous image processing algorithms to detect mushrooms. A manuscript detailing our mushroom dataset for experiments and mushroom identification algorithms has been submitted for publication. Due to the switch from a Cartesian to SCARA robotic arm, we had to revise the mushroom bed scanning algorithm. The location of the scanning camera (Intel RealSense D405) has been modified . The constraints are: 1) working distance from the camera to mushroom bed must be between 7 and 50 cm; 2) mushroom scanning needs to be efficient and quick; and 3) image processing needs to be efficient and preferably real-time. These constraints are addressed through a combination of hardware and software solutions. Currently, we are using YOLOV8 algorithm for mushroom detection. Since mushrooms may not grow straight up and can go any direction, we are also working on the mushroom pose estimation. Besides tuning different algorithms, we have acquired the latest microcontroller NVIDIA Jetson Orin AGX to get more cores for faster GPU computing. Modeling mushroom production system: We have also modeled a medium-scale mushroom farm with substrate storage and substrate processing in a commercial setting. Energy-efficient production and processing of agricultural products are crucial for reducing fossil fuel use, which contributes to global warming and climate change. Computer simulations can determine energy consumption for stochastic processes. We developed a simulation model for a large-scale mushroom farm to audit energy expenditure and improve process efficiency in the substrate processing facility. Mushroom substrate composting involves steps like bale transportation, crushing, watering, nutrient supply, turning, and compost transport. The model analyzes energy requirements for each stage using industry equipment and vehicle specifications collected from literature. This work has been submitted to the Winter Simulation Conference as an extended abstract and will be presented as a poster. This work is being extended to simulate the other mushroom production related processes such as harvesting. This simulation will help determine the impact of automation on mushroom production.
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2024
Citation:
1. Biru Bikram Koirala, Abishek Kafle, Huy Canh Nguyen, Jiming Kang, Abdollah Zakeri, Venkatesh Balan, Fatima Merchant, Driss Benhaddou, Weihang Zhu*, A Hybrid Three-Finger Gripper for Automated Harvesting of Button Mushroom, Actuators 2024, 13(8), 287; https://doi.org/10.3390/act13080287
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2024
Citation:
2. Biru Bikram Koirala, Guoyang Shen, Huy Nguyen, Jiming Kang, Abdollah Zakeri, Venkatesh Balan, Fatima Merchant, Driss Benhaddou, Weihang Zhu*, Development of a Compact Hybrid Gripper for Automated Harvesting of White Button Mushroom, ASME IDETC/CIE2024-143056, Washington DC
- Type:
Journal Articles
Status:
Submitted
Year Published:
2024
Citation:
3. Abdollah Zakeri, Mulham Fawakherji, Jiming Kang, Bikram Koirala, Venkatesh Balan, Weihang Zhu; Driss Benhaddou, Fatima Merchant, M18K: A Comprehensive RGB-D Dataset and Benchmark for Mushroom Detection and Instance Segmentation, Computers and Electronics in Agriculture
- Type:
Conference Papers and Presentations
Status:
Under Review
Year Published:
2024
Citation:
Developing a Simulation Model for Energy Consumption in Mushroom Substrate Composting, Ezra Wari, Sandesh Risal, Weihang Zhu, Venkatesh Balan, poster at Winter Simulation Conference, Orlando, Florida, December 2024
|
Progress 09/01/22 to 08/31/23
Outputs
Target Audience:We will work with mushroom growers to automate the mushromm harvesting process. Specifically, we work with Monterey Mushroom, Inc., at Madisoonille, Texas. We have visited its mushroom farms to better understand the requirement and contraints for the harvesting system design. Changes/Problems:The development of the mushroom harvesting system is on schedule. What opportunities for training and professional development has the project provided?This USDA project has helped to recruit and train undergraduate senior design students from mechanical engineering programs, capstone students from biotechnology programs and complete their research projects. It has also helped to recruit Ph.D. students to complete their dissertation in mushroom robot system development and computer vision algorithms. It also helped us to receive an additional UH infrastructure grant to establish a permanent test best at UH to train and educate students and small scale mushroom producers. Some of the students and their completed projects are given below. One Ph.D. student Abdollah Zakeri is researching computer vision algorithms for identifying mushrooms. Two Ph.D. students Biru Koirala and Jiming Kang are working on the mechatronic system hardware and software development. One undergraduate student Huy Nguyen worked on the mushroom gripper design. In addition, the relevant project activities include: Three biotechnology undergraduate students (Dinithi Jayasinghe, Jazmin Valle and Rob Schmyler) completed a capstone project on the topic 'Producing mycelium and recycling spent mushroom substrate' in May 2023. Six biotechnology undergraduate students (Aaron Martinez, Jarod Nguygen, Aiden Tran, Caitlin Tran, Ha Truong, Swami Sundaravel) completed a capstone project on the topic 'Composting and Mixing SMS to Use as Casing Layer Materials for A. bisporus Cultivation' in December 2022. Four senior design students from mechanical engineering program (Adrian Gonzalez, Efrin Guerrero, Alex ?Reyes, Rumaldo Villaseñor) has jointly developed a 'Automatic Tumbling composter system to convert spent mushroom substrate as fertilizer and casing material to replace peat moss' using cement mixer in May 2023. Three Ph.D. students from (Mahsa Alian, Sandesh Risal, Ezra Wari) are researching 'Modeling small scale mushroom cultivation in shipping containers.' How have the results been disseminated to communities of interest?For this funding period we have published a conference paper and are in the process of writing five other manuscripts that will be published in various peer reviewed scientific journals. We have developed a website to disseminate our research findings to mushroom growing communities (https://mushroom.egr.uh.edu/) and presented some of our research findings at the recently concluded ACS spring conference held in Indianapolis, IN. As the COVID traveling restrictions have been lifted, we will present our research finding in domestic and international conferences. What do you plan to do during the next reporting period to accomplish the goals?In the second year, we are looking forward to completing the mushroom robotic harvesting system development and go to the field testing at our industrial partner Monterey Mushroom.
Impacts
What was accomplished under these goals?
Mushroom Robot Development: We have been working on the mushroom robot structural design. The robot consists of the following parts: 1) a mobile robot chassis that can navigate in a complex environment; 2) a lift structure that can move the picking arms up and down; 3) an arm mechanism that can move the wrist and gripper in and out of the mushroom trays; 4) a wrist mechanism that can move the mushroom gripper up and down; 5) a mushroom gripper that can pick up mushroom; and 6) computer vision deep learning algorithms and camera for identifying mushrooms ready for picking. Three doctoral students are working on this project. An undergraduate student funded by NSF REU program also worked on this project in summer 2023. As of now, each component has been developed. The components need to be improved and integrated as one system. A system engineering approach is adopted to guide this integration process. A manuscript on mushroom dataset for experiments is being prepared for publication. Shipping container testbed: We were successful in receiving a UH infrastructure grant ($320,000) that will be used to build a test bed comprising of two shipping containers. This funding will help to establish a permanent testbed at UH campus. Several small and medium scale mushroom producers will receive hands-on training by participating in a certification program. We have finalized the design, identified the vendor and are in the process of receiving approval from UH facilities to place order and install the containers at UH Sugarland campus at the end of fall 2023. Using image processing to harvest mushrooms: Image processing and data analysis is an emerging technology used to assess the size and quality of mushrooms to determine the growth rate and measure yield. In this paper, we report different steps involved in collecting and processing image data in a large-scale commercial mushroom farm. We have demonstrated the technical feasibility of measuring the quantity and yield at different stages of mushroom cultivation which can be automated in the future. Developing this technology would allow mushroom producers to adjust the microclimate conditions to get better yield and coordinate harvesting schedules when quality is at its peak to maximize their profit. Modeling small scale mushroom cultivation system: We have modeled a medium-scale mushroom farm with substrate storage and substrate processing. Unit operations such as milling, mixing, packing, sterilization, and mycelium inoculation to produce hundreds to a few thousand pounds of mushrooms per day using a 40-foot-long shipping container as a growth chamber. We used a computer-aided design (CAD) model in a software called SOLIDWORKS. We used substrate and processing condition information of oyster mushrooms growth conditions reported in the literature. We also evaluated the organization of different substrate beds packed in bottles, bags, and buckets to maximize the utilization of space and calculated the amount of oyster mushrooms that could be produced per container. This modeling study has helped us to determine the minimum footprint requirement to build a medium scale mushroom farm including the total number of required shipping containers. This will lay the foundation to estimate the cost of establishing such farm and return of investment based on the volume of mushroom produced. Similar modeling studies can be used to evaluate the cultivation of different varieties of mushrooms at different scales by varying the substrate combinations, packing methods, and growth conditions. We are in the process of summarizing this work and submitting two manuscripts.
Publications
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
1. Sandesh Risal, Venkatesh Balan, Weihang Zhu, Ezra Wari, Masha Alian, Modeling Urban Medium-scale Oyster Mushroom Cultivation using Shipping Container. Proceedings of the IISE Annual Conference & Expo 2023, New Orleans, LA, May 2023
|