Progress 05/01/24 to 04/30/25
Outputs
Target Audience:• Academic researchers: faculty, post-docs, and graduate/undergraduate students in plant pathology, metagenomics, agricultural automation, robotics, horticultural science, and agricultural economics. • Technical workforce: engineers, scientists, technicians, and extension agents who translate research into field-ready tools. • Industry stakeholders: greenhouse growers, crop advisors, input suppliers, and CEA business owners seeking data-driven solutions. • Education & outreach: K-12 educators, community groups, and local administrators championing hydroponic school or community gardens. • Governance & policy: state commodity boards, CALS advisory members, and senior policymakers shaping the future of indoor agriculture. Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?For aim 1. Agraduate student pursuing a Master's degree in Robotics has led the development of the robot and its perception system, enabling the autonomous measurement of key physical parameters of the plants. In parallel, an undergraduate student was trained to operate the robotic system and to deploy the biomass prediction models developed in this project. As part of the training and technology transfer effort, the undergraduate intern visited PI Song Li's lab, where they successfully assisted in deploying both the robotic system and the trained predictive models. At VT, a master's student was trained to replicate the robotic system for growing hydroponic lettuce developed by the CMU team. The student was able to successfully build a replica of the robot and the new system at VT can evaluate the effect of two different nutrient solutions side by side. For aim 2. Several graduate and undergraduate students worked on Objective 2. At VT, one MS level graduate student from the School of Plant and Environmental Sciences, as well as a PhD level graduate student were involved part time working on the IoT component of the project. At VT, another MS student in Computer Science continued his efforts in understanding indoor air quality--specifically, CO2 variation patterns across different buildings--and strategies for effectively modeling those patterns. At Clemson, two undergraduate students (one from Horticulture and another from Mechanical Engineering) worked on the project part time to help build a fully automated hydroponic setup. For aim 3. Drs. Boris Vinatzer, Blake Johnson, and Juhong Chen: As part of the Smart Health objective, one undergraduate student continued to be trained and five graduate students were trained at the intersection of plant science, plant pathology, bioinformatics, and engineering in the PI's respective departments. For aim 4. The project has provided support for training and professional development for two Ph.D. students (Lee and Qi) whose dissertations will incorporate essays related to hydroponics and CEA. For aim 5. Dr. Kaylee South: One Horticulture M.S. student (Colin Fite) was hired whose thesis work focuses on herb production in indoor vertical farms and community garden applications of CEA production systems. How have the results been disseminated to communities of interest?For aim 1, CMU: Co-PI Silwal and Kantor shared research methodologies and findings through seminars at multiple institutions including Penn State University, Iowa State University, West Virginia University, ICRA conference etc. For aim 2, (a) Our VT group presented the results of this work in an i2GROW-wide meetingin Feb. 2025. We also presented the generated results in the annual international meeting of American society of agricultural and biological engineers, July 2024, Anaheim, CAwith the title of"Intelligent Greenhouse Gas Removal in Indoor and Controlled Environment Agriculture". Furthermore, we disseminated the results by presenting in a symposium held at the School of Plant and Environmental Sciences (SPES), Virginia Tech (Sept. 2024), as well as a seminar, in which graduate students were the primary audience (Feb. 2025). As noted previously, a conference paper was accepted and presented at the ASCE International Conference on Computing in Civil Engineering, which took place on July 2024, in Pittsburgh, PA. The paper was presented at the conference and published later in 2025. Finally, an MS thesis in computer science was completed and published by the Virginia Tech library in 2024. A journal paper has been being prepared to be submitted. (b) From Clemson's team, co-PI Dr. Velni was invited to present at two events in 2025. In June, he gave a talk at the International QUAD-AI ENGAGE Workshop on AI and Digital Agriculture, hosted by Georgia Tech. This workshop attracted over 60 participants from academia, industry, and government, from US and abroad. Additionally, in January, Dr. Velni delivered a presentation during the Graduate Seminar Series of the Clemson Department of Plant and Environmental Sciences. Both presentations focused on cost-efficient lighting solutions for indoor agriculture, aligning with the i2GROW project's goals. For aim 3, For objective 3, the undergraduate student and graduate students shared their results at university-level research symposia and results were also presented at the Controlled Environment Agriculture Summit East. Danville, VA. For aim 4 and 5?to begin disseminating the results, we engaged with other teams within the I2Grow project and presented our research proposals, including the objectives and methodological framework. Our first paper under this aim (Lee and Bovay, "Impact of Hydroponic Technology Adoption on Vegetable Production in the United States") was accepted for presentation at the 2025 Agricultural and Applied Economics Association (AAEA) Annual Meeting in Denver (scheduled for July 28, 2025). What do you plan to do during the next reporting period to accomplish the goals?For aim 1, CMU: We plan to further evaluate the efficacy of the digital twin model by conducting additional experiments and quantifying performance metrics across a broader range of conditions. This will allow us to thoroughly assess the robustness and generalizability of the predictive models. Additionally, we intend to publish a systems paper detailing the hardware, software architecture, and modeling approaches, and to publicly release the full dataset and associated algorithms to support reproducibility and foster further research in this area. For aim 2, Our research on CO? recycling systems is progressing with continued focus on optimizing indoor production environments. With IRB approval secured, we are now positioned to investigate the impact of several critical parameters, including production space configuration, human occupancy, air ventilation, and HVAC systems. In parallel, we are developing cloud-enabled CO? sensors that will allow growers and researchers to continuously monitor key environmental variables (such as CO? levels) in real time. This advancement supports precise environmental control and maximized crop yields. In the upcoming reporting period, we will explore the correlation between various bacterial inoculations and CO? concentrations. This line of inquiry aims to provide growers with more effective strategies for disease control in controlled-environment agriculture. Finally, we plan to disseminate our findings through presentations at both university-wide and international conferences.We will also continue our efforts to investigate CO? modeling using emerging approaches in scalable machine learning, including generalized time series models trained on heterogeneous datasets. These efforts may involve local data collection to support model validation and performance assessment. At Clemson, the group will focus on developing new strategies to co-optimize light intensity and CO? concentrations in order to enhance both the growth rate and quality of indoor-grown vegetables. In particular, we plan to quantify CO2and light interactions and use them to co-optimize these two critical environmental factors. We select two crops - lettuce and radish - with contrasting morphology and sink strengths to further investigate how crop sink-source relationships influence their response to light intensity and CO2. For aim 3, after repeating the experiment to determine the effect of human touch on the lettuce leaf microbiome in spring 2025, Dr. Vinatzer's group plan toextracted all DNA and sent it for 16S rRNA amplicon sequencing. We will analyze the data and repeat the experiment a third time in fall 2025. We are planning to write up a manuscript on the risk of human touch on the lettuce leaf microbiome and its consequences on plant and human health risks in spring 2026 and submit it before the end of the next reporting period.Dr. Blake Johnsongroupplansto submit at least two manuscripts based on the research accomplished during the last two years and continue extending applications to continuous monitoring formats that can interface with indoor hydroponics systems. Dr. Juhong Chenplansto publish one to two peer-reviewed manuscripts based on the results obtained so far. For aim 4 and 5. In the next period of the project, Lee and Dr. Bovay will conduct the analysis applying a random forest model to identify the most influential predictors of hydroponics adoption at the state level. Variables of interest include producer age, production experience, electricity and natural gas prices, and average temperature, etc. We will examine both the importance of each predictor and its marginal effect on adoption. Based on these results, we will sketch the adoption landscape and identify regions or producer characteristics with higher potential for adopting hydroponics. In addition, we will develop a partial equilibrium simulation model to assess how hydroponics adoption may affect vegetable markets--particularly supply change. We will also finalize a manuscript for journal submission to disseminate the results to broader communities of interest.In addition, Qi and Dr. Bovay plan to finalize the CEA site selection framework in the next reporting period. Qi also plans to present his work and preliminary results at the upcoming 2025 CEA Summit East Graduate Student Research Poster competition where he can share it with broader communities of interest. In the next reporting period, Dr. South and Fite will complete the projects investigating basil production under varying CO2 conditions and other production practices and summarize findings in presentations, publications, and trade articles. This project and the general I2GROW project will be showcased at the CEA Summit East 2025. South and Fite will also work on developing Extension publications on topics and findings related to Aims 1 - 3. To continue with the goal of industry collaboration, South will continue to collaborate with industry and outside university partners on outreach and research efforts that were established during reporting period 2. Dr. Li will organize three major activities, including the second advisory board meeting, the second round of mini-grants and writing a manuscript describing the scope and progress of the I2GROW project.
Impacts
What was accomplished under these goals?
For Aim 1. In the previous reporting period, we developed an autonomous gantry-style hydroponic robot equipped with RGB-D imaging and a suite of environmental sensors, including CO?, humidity, and light intensity sensors. Leveraging the multimodal data collected by the robot along with ground truth biomass measurements, we developed a digital twin model of lettuce plant growth. This digital twin architecture continuously integrates real-world sensor and image data into the simulation model, enabling more accurate biomass prediction over time. Using this framework, we achieved a preliminary biomass prediction accuracy within ±2 grams for a forecasting horizon of 4 days. To evaluate performance, we trained and compared several models, including the Nicholet model, Multi-Layer Perceptron (MLP), and deep neural networks such as Long Short-Term Memory (LSTM) networks. For Aim 2. Using an array of low-cost CO2 sensors, we have collected data from multiple different indoor agriculture environments - (a) green house facilities located on the campus of Virginia Tech; (b) vertical farming facilities located at the institute for advanced learning and research (IALR) at Danville, VA. We have been able to develop learning-based models and use time series to predict the next hour/day CO2 concentration in indoor environment. The input parameters are relative humidity, temperature, and atmospheric pressure in order to predict the CO2 concentration as the output. In one study, an LSTM model was trained to predict greenhouse CO2 concentration over a 2-hour period, achieving high precision with an R² value of 0.82. Additionally, we have extended the study to conduct season-based data acquisition in order to enable us to generate longer prediction horizons for CO2 concentration and the need for the supplements. Additionally, we have been pursuing the idea of developing testbeds for other researchers and scientists. We have initiated collaboration with other academic departments, research institutions and companies. This pursuit yielded two min-grants under the umbrella of i2GROW project, focusing on better understanding of the impact of CO2 concentrations on plant health and growth in various indoor environments. Further investigations into CO? variation patterns in educational buildings (where the presence of higher numbers of occupants in subspaces, such as classrooms, results in the accumulation of CO?) and the feasibility of scalable modeling practices were continued. Different variation patterns were explored using our compiled dataset to provide insights into indoor CO? dynamics. Additionally, various combinations of features and modeling paradigms were tested to evaluate the potential for scalable modeling. Scalable modeling focuses on methods that reduce the need to develop separate models for individual subspaces. The basic concepts and implementations of time series foundation models (models trained on large and diverse data) have been also studied to see whether they can be adopted for CO2 modeling in educational buildings. A journal paper is in preparation from the results. For Aim 3. Dr. Boris Vinatzer continued our study on how human touch affects the microbiome of lettuce plants grown hydroponically inside buildings with heavy human activity. Specifically, we want to determine the risk to plant health and human health (due to consumption of plants grown in the presence of people). Therefore, we repeated the study from year 1 with slight variations and also cultured bacteria from the lettuce plants that were either touched or not touched by students. All results so far do not reveal the presence of any human pathogens on the touched plants and no plant pathogens. We only found known skin-dwelling bacteria and known plant-associated bacteria. Dr. Blake Johnson's grouphas established a high-throughput platform for optimizing the working electrode surface functionalization approach for biosensor-based pathogen detection using electrochemical and electromechanical biosensors. We have also established physics-guided machine learning models capable of improving biosensor performance in practical pathogen detection-based biosensing applications, such as improving the measurement confidence and reducing false negative and false positive alarms/results. In particular, this reporting year we have focused on generating biosensing datasets that will allow us to potentially improve the detection limit of biosensors for pathogen detection using machine learning. Dr. Juhong Chen's group has been working on novel pathogen detection methods.Current pathogen detection methods, including culture?based plate counting,enzyme-linked immunosorbent assay (ELISA), and polymerase chain reaction (PCR), suffer from slow turnaround, destructive sampling, and reliance on centralized laboratory infrastructure. To overcome these limitations, we continued to develop a portable platform that couples a swelling?driven polyvinyl alcohol microneedle patch and a CRISPR fluorescence assay. By enabling on site and real time monitoring ofpathogens, our microneedle-CRISPR platform empowers growers to implement targeted interventions earlier, enhancing crop health and productivity. For aim 4, graduate student Lee and Dr. Bovay conducted an extensive review of the literature on agricultural technology adoption and identified affecting factors in the context of hydroponics. Based on this review, we finalized potential determinants of hydroponics adoption. We then collected data from multiple sources on producer characteristics, environmental and climate-related variables, and demographic factors at the state level for the analysis. After evaluating various estimation approaches, we determined that a random forest model would be appropriate for identifying the key predictors of hydroponics adoption. We also completed data cleaning and preparation for the analysis. In addition, graduate student Qi and Dr. Bovay developed a research plan for analysis of CEA feasibility at a more refined spatial level and the effects of local CEA production on food availability and affordability. For aim 5. Dr. South worked with Aims 1 - 3 in identifying their target audiences and created extension article and online resource tool plan to share impact of the I2GROW project including the use of CO2 in CEA facilities and factsheets on the use of the community gardens and applications of the production practices in commercial facilities. Dr. South gave a talk about community gardening using CEA technologies to a group of Master Gardeners who work with adult and K12 communities. Dr. South also started a project with an industry collaborator on the impact of community CEA garden systems on room CO2 levels. Dr. South and graduate student Fite have completed an extensive literature review to build a thesis project to investigate the production of basil in indoor vertical farms with varying production practices including varying CO2 levels. This project will be completed in reporting period 3. Dr. Li is the PI of this project, and during this reporting period, he has organized the advisory board meeting by inviting seven experts in the domain of CEA to evaluate the progress of the I2GROW project and provide guidance for the future direction of this project. Dr. Li also organized the review of I2GROW mini grants and released the first phase of six mini-grants. These grants support the collaboration of external investigators with the I2GROW team to address questions that related to CEA production.?
Publications
- Type:
Other Journal Articles
Status:
Other
Year Published:
2025
Citation:
Predictive CO2 concentration and supplementation in indoor agriculture
- Type:
Other Journal Articles
Status:
Under Review
Year Published:
2025
Citation:
The Role of Wireless Sensor Networks in Precision Agricultural Production: A Systematic Review
- Type:
Theses/Dissertations
Status:
Published
Year Published:
2024
Citation:
Meimand, Mohammad, Soft Sensing?Driven CO? Predictive Models in Educational Buildings, Masters thesis, Virginia Polytechnic Institute and State University, 14 Oct. 2024.
- Type:
Peer Reviewed Journal Articles
Status:
Published
Year Published:
2025
Citation:
Alaviani, Seyyed Shaho, Marc W. van Iersel, and Javad Mohammadpour Velni. Optimal Supplemental Lighting Control Using LED Lamps in Greenhouses Considering Multicrops and Spatial Supplemental Light Distribution, IEEE Transactions on Control Systems Technology (2025). Link: https://doi.org/10.1109/TCST.2025.3553000
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2024
Citation:
Kaur, S., Karuppasamy, T., Liu, H., Li, S., Vinatzer BA. (2024) An Investigation of the Effect of Human Touch on the
Taxonomic Composition of the Lettuce Leaf Microbiome. Controlled Environment Agriculture Summit East. Danville, VA.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2025
Citation:
Karuppasamy, T., Kaur, S., Gerken, M., Liu, H., Li, S., Vinatzer BA. (2025) Hands to Plants: An Investigation of the Effect of Human Touch on the Lettuce Leaf Microbiome. Dennis Dean Undergraduate Research and Creative Scholarship Conference. Blacksburg, VA.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2025
Citation:
Karuppasamy, T., Kaur, S., Gerken, M., Liu, H., Li, S., Vinatzer BA. (2025) Hands to Plants: An Investigation Into Food Safety Risks of Growing Edible Plants in Classrooms. ACC Meeting of the Minds Conference. Pittsburgh, PA
|
Progress 05/01/23 to 04/30/24
Outputs
Target Audience:Faculty, scientists, extension agents, graduate and undergraduate students, producers, industry partners, and technical staff are engaged in various aspects of the CEA/indoor agriculture industry. The audience also includes agricultural economics, plant nutrition and physiology, automation, robotics, technology development and adoption, beneficial microbial applications, and workforce/education needs. We also target faculty and students involved in optimization, controls theory, lighting control, and smart building monitoring and control for indoor environments. Other audiences outside academia include current and future industry members (suppliers and producers) to K-12 and extension agents with an interest in CEA. Changes/Problems:Aim 1. Dr. Song Li: (1) we did not host an advisory board meeting in the first year because participating teams have not yet produced enough research output for the advisory board to evaluate. One advisory board member passed away and one industry member disengaged from our activities. Therefore, we plan to recruit a new advisory board in fall 2024 and share the project report with this new advisory board. (2) we originally plan to host all information about this project as a facebook page, however, this approach does not seem to reach enough audience. In the 2nd year of this project, we plan to focus on updating a google sitefor the I2GROW project to disseminate the research results to the public and to engage more collaborators in the research community. (3) we plan to start the first phase of external collaborative grants in Fall 2024. The goal is to find additional partners in academia and industry to extend the impact of I2GROW project. The whole team will draft an RFP before Fall 2024 and release to the public in Fall 2024. The first phase of collaborative projects is expected to start in Spring 2025 and end by Fall 2025. These new projects will serve as preliminary results for new grant applications to be submitted in Fall 2025 or later to local and federal agencies. Dr. Tie Liu and Song Li will be the main lead on this aim. Aim 3. Dr. Blake Johnson: While we did not face any major challenges, we did have to replace a workstation in our lab used for modeling of biosensor response that is planned to be used for validation of experimental measurements made in year 2, which led to a slight delay. Dr. Juhong Chen: Due to Juhong's transfer from Virginia Tech to UC Riverside, a new graduate student (Chulin Cao) needed to be recruited to work on the delaying the development of the CRISPR-based biosensor to detect plant pathogens. Aim 4. Dr. John Bovey: Aim 4.1 outlined a plan for surveying Eastern U.S. hydroponic vegetable growers about the cost of production and practices undertaken to reduce emissions. However, after extensively considering the logistics of implementing such a survey, we have decided not to pursue this Aim. With coauthors (Goodrich et al., 2023), Bovay documents experiences with fraudulent responses to surveys eliciting information from farmers and the small number of valid responses obtained in each of two surveys. Given that the target population of growers under study in this aim would have been so small, we would have been unlikely to obtain a sufficient number of valid responses for statistical inference. Other co-PIs have nothing to report in this period. What opportunities for training and professional development has the project provided?Aim 1. Dr. Eva Colla'kova': We tailor training and professional development of our students and postdocs to their interests and needs. The GRA recruited for this project has medical bioengineering training, so this was a great opportunity for him to learn about plant growth and metabolism as well as about different analytical methods used in plant research. His engineering experience and newly acquired plant knowledge both proved useful in helping with the construction of the gantry/hydroponic system in Li's lab. He also attended the Cold Spring Harbor Laboratory "Frontiers & Techniques in Plant Science" course in summer 2024 to expand his knowledge on the current plant-related methods useful to this project and his planned future career as a NASA scientist. Dr. Song Li: One graduate student (together with Eva's student) was trained to build I2GROW Oasis system. Both students visited CMU robotic lab, and learned the system setup of the I2GROW Oasis at the CMU site. Both students participated in constructing the Oasis testbed at Li Lab in Virginia Tech Cooperate Research Center. The student was trained to assemble the farmbot on top of a NFT hydroponic system and control the robot with an online interface for autonomous image collection. Dr. Jactone Ogejo: One Ph.D. student is working on the project, and the work is expected to be part of their dissertation research for their graduate studies. Dr. Abhisesh Silwal and George Kantor: Under Aim 1, a graduate student was hired and is currently working on his Master's degree in Robotics, building the robot and perception system to perceive and measure plant's physical parameters. One other master's and a undergraduate student are also involved at a robot systems integration level and have received hands on robotics experience through internships. Aim 2. Dr. Javad Velni: One graduate student (working as hourly) is partially working on the project and another one was recruited and will start (pending US visa clearance) in Fall 2024. Dr. Hasan Seyyedhasani: A graduate student was hired and is currently working on her Master's degree in Plant and Environmental Sciences, developing a digital twin to predict and evaluate CO2 supply for indoor agriculture. Another graduate student is also involved at sensing system development level and has received hands-on experience through internships. Dr. Farrokh Jazizadeh: One graduate student was involved with this project that was trained on understanding indoor air quality, specifically CO2 variations in buildings. The student was trained to compile relevant literature, extract, and organize data, and use different machine learning and visualization tools to explore the patterns and opportunities for modeling as well as interpreting model outcomes. Aim 3. Dr. Boris Vinatzer, Blake Johnson, and Juhong Chen: As part of the Smart health objective, two undergraduate students and five graduate students were trained in the PI's respective departments. Aim 4. Dr. John Bovey: nothing to report. Aim 5. Dr. Kaylee South: 1 GRA was recruited to begin his master's degree in Horticulture in the second reporting period. How have the results been disseminated to communities of interest?Aim 1. Dr. Abhisesh Silwal and George Kantor: Seminars: Co-PI Silwa and Kantor shared research methodologies and findings through seminars at multiple institutions including Oregon State University, Iowa State University, Virginia Tech etc.Class lectures: Co-PI Abhisesh Silwal included the robot design and AI concepts as a case study in one of the lectures on the course 16765 Robotics and AI in Agriculture from 2021 through 2024.Demonstrations: We have demonstrated the system design and automated data collection pipeline to several aspiring high school students, undergraduate freshmen seeking degrees in robotics, faculties and researchers from various academic institutions across the country, government officials from the USDA. Dr. Song Li: We have participated in CEA-east conference in Danville in Fall 2023. I presented an overview presentation about the I2GROW project as a short talk to an audience include academic researchers, producers and industry partners. I also presented poster presentation in CAIA big event in Fall 2023 to show case the I2GROW project to researchers in Virginia Tech. I also presented our research as an oral presentation at the PAG 2024 conference in the post harvesting session. I also described our I2GROW project to ~20 high school students participated in summer governor's school of agriculture in summer 2023. Aim 2: Dr. Hasan Seyyedhasani: We presented the initial set-up of this work in an i2GROW-wide meeting (Jiaye Liu & Hasan Seyyedhasani) and are planning to present the generated results in the annual international meeting of American Society of Agricultural and Biological Engineers. Furthermore, another conference paper was submitted and accepted to be presented at ASCE International Conference on Computing in Civil Engineering which has been scheduled to take place July 28th -31st, 2024 in Pittsburgh, PA hosted by Carnegie Mellon University. Although the paper is accepted and the presentation will be in July 2024, the official publication will be later potentially in 2025. Aim 3. Dr. Boris Vinatzer: For objective 3, the undergraduate student shared the experimental design in poster presentations at university-level research symposia and results were shared at the ICPPB & Biocontrol conference. Other co-PIs have nothing to report in this period. What do you plan to do during the next reporting period to accomplish the goals?Aim 1. Dr. Eva Colla'kova': We will keep evaluating plant growth and nutritional quality grown on different types food waste and will initiate growth experiments in the automated gantry. Dr. Song Li: we plan to recruit a new advisory board and report our progress to the advisory board and seek feedback from the advisory board. We plan to update the project website that can better reach to a broader audience. We plan to host the 2nd I2GROW meeting including all participating PI, co-PI, students, postdocs, potential collaborators, and advisory board members. Dr. Abhisesh Silwal and George Kantor: We plan to use the developed robot system and validate its capability over repeated experiments. The experiment would include cross comparison between closeness of variables measured between the robot system and manual measurements. Dr. Jactone Ogejo. Refine the methods of making biofertilizers from food waste using the ongoing entomocomposting and Bokashi fermentation processes. We also expect to derive fertilizer using the anaerobic digestion process. We are currently setting up the digesters for the work. Dr. Hasan Seyyedhasani's team: Our study will continue by further examining CO2 recycling systems. Also, we aim at developing a low-cost IoT-based sensing system to collect data from various indoor environments and control them remotely. Furthermore, our study will be extended to conduct season-based data acquisition in order to enable us to generate longer prediction horizons for CO2 concentration and the need for the supplements. Dr. Farrokh Jazizadeh's team: We will continue our efforts to further investigate the generalizability of CO2 concentration models and further show the feasibility of modeling for different environments. This may include collecting data locally to augment the available datasets. We are also planning further dissemination. Dr. Javad Velni's team: Our team will examine the interrelation between supplemental CO2 and supplemental lighting and how the two can be co-optimized under the assumption that other variables are regulated. Aim 3. Dr. Boris Vinatzer: For the Smart health objective, we will repeat the experiment done in year 1 to assess the impact of human touch on the leaf microbiome of edible plants grown hydroponically adding additional comparisons to plants grown in soil outdoors, plants grown in commercial CEA settings, and plants sold at the grocery store. Dr. Blake Johnson: We plan to move forward with validation of several biosensing modalities for detection of target food and water-borne pathogens as well as creating a biosensor-integrated hydroponics testbed that can enable us to both assess the performance of biosensors for pathogen detection and the performance of a digital twin to support 'smart' biosensing for hydroponics applications. Dr. Juhong Chen: In the next year, we will catch up to finish the development of CRISPR-based biosensors to detect plant pathogens. Likely, we will have one manuscript submitted and plan to present our discoveries in a national conference. In addition, we will investigate the performance of microneedles to extract pathogenic DNA from plants. Aim 4. Dr. John Bovey: By the end of the next reporting period, Bovay's Ph.D. student should have nearly completed her dissertation, which will include one chapter described in section 5.1 above. Bovay is also planning to hire a scientist with technical expertise and the experience necessary to complete aim 4.2, on the monetary costs and benefits of our production system. Depending on timing, this scientist may also lead work on the life-cycle analysis described in aim 4.3. Aim 5. Dr. Kaylee South: A GRA will begin in the lab and start working on research related to crop production in CEA systems and assist with outreach efforts related to K12 and workforce training. His work will likely result in a poster presentation at a conference. After the testbeds are developed, industry partnerships will be established and workshops will be developed for training and feedback from stakeholders who will utilize the technology/systems developed in this project.
Impacts
What was accomplished under these goals?
?Objective 1: Smart Growth Dr. Song Li:As the PI of the project, I primarily organized the I2GROW project. I sent weekly or bi-weekly updates to the entire team to ensure everyone was informed about the project's activities. I organized sub-aim meetings to discuss project progress, including student recruitment, research design, updates, and coordination between different sub-aims. I managed the budget, set up the I2GROW Facebook page, and presented our project at several local and national conferences. Additionally, I participated in the GROW conference as a panelist and reviewer for urban CEA projects. Lead on Robotics and Automation--Dr. Abhisesh Silwal and George Kantor: We've developed a fully integrated prototype of the I2GROW-Oasis at CMU. This prototype consists of a 3-axis gantry robot system with a growth bed for crops, LED lighting to facilitate growth, an NFT-based irrigation system, an RGB camera to image the plants, and sensors such as CO2 and humidity sensors to measure environmental variables. The robot system, along with the sensors, is housed inside a large growth tent to isolate it from external environmental influences. The current level of autonomy includes a data acquisition pipeline where the robot can navigate to known plant locations and capture images from multiple views. Environmental sensory data are captured simultaneously with the images for precise data alignment. To aid in data collection, processing, and visualization, we've also developed robust data acquisition and storage protocols, along with a graphical user interface to monitor and display processed spatiotemporal information. Lead on Plant Science--Dr. Eva Collakova: We developed a small-scale hydroponic system to test the growth and nutritional quality of lettuce and tomato plants under different nutrient conditions. Using chromatography-mass spectrometry-based methods, we assessed levels of nutritionals and antinutritionals, including mineral macro and micronutrients, antioxidants, and other specialized metabolites through 10 different assays. Our initial experiments with lettuce and processed liquid food waste revealed that sugar content needs to be reduced by exploring different methods of preparing food waste, such as compost tea. In parallel, a functional automated gantry system, mirroring the one at CMU, was established in Dr. Li's lab with an NFT hydroponic system. Dr. Jactone Ogejo: We tested methods to produce nutrient solutions for hydroponic growth systems from food waste. Three potential pathways were explored: hydrothermal carbonization, compost tea from entomocomposted frass, and solid-state fermented food waste using a commercially available process called Bokashi fermentation. The hydrothermal approach seems to generate solutions high in sugars, posing challenges for growing systems. The fermentation-based processes reduced the sugar levels in the solutions and provided suitable nutrient profiles. Objective 2: Smart Building Lead--Dr. Javad Velni: Leveraging our preliminary results from another NIFA grant, we developed an optimal supplemental lighting control method using LED lamps in greenhouses. Our past work, which provides the only available optimal lighting tools in the literature, did not account for supplemental light distribution and multi-crop cultivation in greenhouses. To address these challenges, we proposed sectioning LED lamp control sources and formulated an optimization problem to minimize electricity costs, regulate light intensity, and minimize the peak of supplemental lights. This was achieved under both supplemental light distribution and multi-crop conditions, using a learning-based algorithm with sunlight prediction. The next step is to deploy and demonstrate the benefits of these tools in commercial greenhouses, beyond laboratory settings. Dr. Hasan Seyyedhasani: To control CO2 levels within a sealed chamber, we explored CO2 Capture & Release through recycling the CO2 delivered to outdoor air handling units.We focused on chemical absorption and identified a collaborator to adapt this method for indoor agriculture. We developed a simulator/digital twin to predict and evaluate CO2 supply. We have started data collection from two indoor agriculture environments--a greenhouse at Virginia Tech and vertical farming facilities at the IALR in Danville, VA. Relative humidity, temperature, and atmospheric pressure are being collected to predict CO2 concentrations. Learning-based models were developed to predict CO2 levels. Time series models, like long short-term memory (LSTM), were developed to predict CO2 concentrations for the next hour and day based on the previous six hours of data. Dr. Farrokh Jazizadeh: We investigated methods to monitor and infer CO2 variations in educational buildings, where higher occupant numbers result in CO2 accumulation. We focused on scalable, generalizable models that do not require instrumenting each subspace. Using publicly available datasets and literature, we studied CO2 concentration variations under different conditions, including buildings with natural and mechanical ventilation. Our machine learning models estimated and predicted CO2 concentrations for various subspaces, showing the feasibility of developing generalizable models despite some challenging cases. Objective 3: Smart Health Lead--Dr. Boris Vinatzer: We assessed the plant and human health aspects of growing edible plants in public spaces. In the first year, we focused on the impact on the plant leaf microbiome by varying degrees of student interaction with hydroponically grown lettuce in a classroom. The experimental design included four sets of lettuce plants: (1) plants grown in a rarely occupied room, and plants grown in a frequently occupied classroom,(2) without student contact, (3)with students invited to approach but not touch the plants, and (4)with students encouraged to touch the plants frequently. No significant differences were found between the microbiomes of the four plant sets, suggesting that growing plants in public spaces may not significantly increase the risk of contamination with human-associated bacteria. Dr. Blake Johnson: Our group established a high-throughput platform for optimizing the working electrode surface functionalization approach for biosensor-based pathogen detection using electrochemical and electromechanical biosensors. We developed physics-guided machine learning models to improve biosensor performance in practical applications, enhancing measurement confidence and reducing false negatives and positives. We also created a digital twin for the hydroponics-integrated biosensing system to interface with surface-based biosensors for pathogen detection. This digital twin model helps guide the identification of pathogen entry points and optimization of hydroponic system parameters. Dr. Juhong Chen: We are developing a CRISPR-based biosensor to detect plant pathogens. So far, we have optimized the CRISPR system to detect a biomarker gene (avr) for plant pathogens, demonstrating high specificity for the detection of the avr gene. Objective 4: Analysis of Economic Feasibility and Social Benefits Lead--Dr. John Bovey: We developed a methodology to explore hydroponics technology adoption rates and their impacts on market supply and producer surplus. Our ongoing research simulates the adoption rates for various crops in the U.S. and measures the impacts on producer welfare. Objective 5: Outreach and Extension Lead--Dr. Kaylee South: We host the CEA Summit East to bring together various stakeholders to discuss CEA-related topics. At the 2023 event, we networked with industry members and presented the I2GROW project. We also conducted a survey to identify workforce training needs for CEA, which will aid in developing training programs. We established relationships with a Virginia Governor School program for K12 outreach, providing hands-on experience with our technologies.
Publications
- Type:
Websites
Status:
Published
Year Published:
2023
Citation:
https://www.facebook.com/profile.php?id=61550880821913
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
The I2GROW project. A testbed of climate-smart technologies for indoor and controlled environment agriculture. 2023. CEA-EAST conference, Danville, VA, Invited talk. by Song Li
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2024
Citation:
Application of AI and hyperspectral imaging in quality assessment of agricultural products. PAG 2024, by Song Li, Invited Talk.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
I2GROW: Intelligent and Integrated Greenhouse gas RemOval With Indoor and Controlled Environment Agriculture. Poster presentation. Fall 2023, By Song Li
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2024
Citation:
Robotics Institute Special Seminar: Title: Robotics and AI for Sustainable Agriculture. May 9th, 2024. Pittsburgh, PA, invited talk, Abhisesh Silwal
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2024
Citation:
Kaur S, Karuppasamy T, Liu H, Li S, Vinatzer BA (2024) An investigation of the effect of human touch on the taxonomic composition of the lettuce leaf microbiome. Poster presentation at ICPPB & Biocontrol 2024, July 7-12, Blacksburg, VA, USA.
- Type:
Websites
Status:
Published
Year Published:
2024
Citation:
https://sites.google.com/vt.edu/i2grow/home
- Type:
Journal Articles
Status:
Under Review
Year Published:
2024
Citation:
S. S. Alaviani, M. W. van Iersel, and J. Mohammadpour Velni, Optimal Supplemental Lighting Control Using Light-Emitting Diode (LED) Lamps in Greenhouses for Multi-Crops Considering Light Distribution, IEEE Trans. on Control Systems Technology, submitted in Feb. 2024, revision submitted in May 2024 and still under review. From a previous project (also supported by NIFA), we extracted some preliminary results and submitted the following manuscript with a former postdoc. Proper acknowledgement to i2GROW will be given once the paper is accepted.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2023
Citation:
Seyyedhasani, Bio-Automation and Smart Farming. Nov 8, 2023. School of Plant and Environmental Sciences, Virginia Tech. Blacksburg, VA. Invited Talk
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2025
Citation:
Meimand, M., Jazizadeh, F., 2024 Toward Generalized CO2 Prediction in Educational Buildings , ASCE International Conference on Computing in Civil Engineering, July 28-31, 2024, Pittsburgh, PA, USA
|