Smart Agriculture using Autonomous Robots Authors and Affiliation. J. Pandya1*, T. Ford1; K. Davis2, Dr. A. Nagchaudhuri3, Dr. C Nindo1, Dr. M. Mitra4.

 Food Science and Technology1, University of Maryland Eastern Shore, Princess Anne, Maryland 21853.

Engineering Undergraduate2, University of Maryland Eastern Shore, Princess Anne, MD 21853.Advisor: Professor, Engineering and Av. Science3 University of Maryland Eastern Shore, Princess Anne 21853.

Professor, Natural Sciences4, University of Maryland Eastern Shore, Princess Anne 21853

ABSTRACT“Smart Farming” efforts at University of Maryland Eastern Shore (UMES) have been largely supported by National Institute Food and Agriculture(NIFA) and have been integrated with AIRSPACES (Autonomous Instrumented Robotic Sensory Platforms to Advance Creativity and Engage Students) project supported by Maryland Space Grant Consortium. Broad goals of the endeavors are aligned with USDA’s “environmental friendly agriculture” and NASA’s “earth science” mission objectives. As part of the AIRSPACES project the UMES team has developed and integrated autonomous platforms such as autonomous boats, autonomous ground robots, as well unmanned aerial systems with the goal of advancing sustainable agricultural practices. This poster mainly focuses on the Farmbot and the experiments conducted using the platform. Farmbot, a small autonomous farming machine, has been recently acquired by the “Smart Farming” and “AIRSPACES” project team at UMES for conducting precision farming field experiments on a small scale. The machine seeds, kills weeds, senses soil moisture levels, and irrigates plants individually over 10 feet by 20 feet area. It is, in essence, a three-axis Cartesian robot quite similar to a 3D printer and runs on Raspberry Pi 3 and Arduino like microprocessor board. It can be easily manipulated using a web application over a smartphone. The Raspberry Pi Camera (Pi-Cam) integrated with the machine can be used for weed detection and time-lapse photography. Food Science and Technology graduate students are working in concert with engineering undergraduates and UMES staff to oversee the installation and preliminary experimental layout. Some of the pitfalls and successes during the installation of both the hardware and software of the mechatronic device will be highlighted in this work. The initial plans of growing turnips with the FarmBot was unsuccessful due to severity of winter weather. The project team has now built a hoop house around the FarmBot to extend the growing season. The Farmbot has been equipped with a solar-wind turbine to generate required electricity to self-sustain its power needs; and its irrigation needs through rainwater harvesting is in progress.  

INTRODUCTION

FarmBot efforts at University of Maryland Eastern Shore(UMES) is integral to smart and precision agriculture efforts[2,3]. The project is envisioned to be a demonstration platform for small scale sustainable food production[4] using a robotic device[5], solar energy[6], and rainwater harvesting[7,8].

Farmbot is an open source automated farming machine consisting of a Cartesian coordinate robot. The Farmbot has been installed at UMES to service a 10ft by 20 ft bed. Most of the autonomous capabilities of the platform have been utilized and demonstrated in the preliminary study implemented . The bed has been prepared with fertile soil to study the impact on growth rate of selected vegetables. The hoop house, rainwater harvesting, wind turbine, and solar panels have been set up for the project for an automated sustainable food production platform.

MATERIALS AND METHODSWe provide below a pictorial overview of the FarmBot software, hardware, and the farm design layout for the preliminary study. A randomized block design ( RCB) has been implemented to study the effects of 4 different irrigation treatments on the growth of Radish and Arugula plants.

Farmbot Web App

A cloud-based web service that performs the following functions;•Farm Design•Farm Events•Controls•Device •Sequence•Regimens•Tools •Firmware•Logs

Farm DesignView and layout of the field experiment in a virtual representation of the Farmbot’s

bed

Blocks:A,B and C.

Irrigation levels: ( for both Radish and Arugula)I0 - No irrigationI1 – 0.75gal/sequence and 1.5 gal/dayI2 – 1.5gal/sequence and 3gal/day I3 – 1.75gal/sequence and 4.5gal/day

The plants were harvested at predetermined times. Radish was harvested in two different intervals. Due to unavoidable circumstances Arugula harvest had to be delayed well beyond the typical harvest time. A mini electronic scale was used to weigh a selected number plants from each irrigation zone (shown in Figure 6) and their average value (in grams) was used as the harvest data in the ANOVA analyses.

Farm Event•Event Scheduling•Choose Sequence start time and Date•Run Sequence as one-off or repeating basis

Controls

Manual Controls for the Farmbot were utilized to ensure error free operation prior to switching to the autonomous mode. Specifically manual controls were used to (i) To test the desired home location.(ii) To ensure all the peripheral devices were working properly.(iii) Check all the sequence locations and regimens are running as desired

CONCLUSIONThe FarmBot platform provides an educational and research platform to build awareness with issues related to food, energy, water and the environment.

First trial for Radish and Arugula was executed successfully despite challenges due to COVID 19 restrictions on campus. The trial was conducted according to the farm design with four levels of irrigation. The cloud computing and autonomous features of the Farmbot were effectively utilized.

The efforts to develop a semi- autonomous self-sustaining environment for the FarmBot as reported in this poster demonstrate efficient water and energy use to meet the growing demand of farm-to-table services for the urban and suburban population. It may also be worthwhile mentioning here that NASA personnel at Kennedy Space Station FL are exploring ways to grow food at the International Space Station, as well as on Mars and Moon using devices similar to FarmBot with appropriate enhancements.

LITERATURE CITED

[1] Davies, F., & Garrett, B. (2018, November 20). Technology for Sustainable Urban Food Ecosystems in the Developing World: Strengthening the Nexus of Food–Water–Energy–Nutrition. Retrieved July 21, 2020, from https://www.frontiersin.org/articles/10.3389/fsufs.2018.00084/full

[2]Nagchaudhuri, A., Mitra, M., Schwarz, J.G., Marsh, L., Daughtry, C.D., Teays, T., “UMES STEM Faculty, Students, and Staff Collaborate to address Contemporary Issues Related to Energy, Environment and Sustainable Agriculture,” Proceeding of 2012 Annual Conf. of American Society for Engineering Education, June 10-13, San Antonio, Texas

[3]Nagchaudhuri, A., Mitra M, Hartman, C., Ford, T., and Pandya, J., “Mobile Robotic Platforms to Support Smart Farming Efforts at UMES”, Proceedings of 14th IEEE/ASME Mechatronics and Embedded Systems Applications (MESA 2018), July 2-4, 2018, Oulu

[4]Chel, Kaushik. Renewable energy for sustainable agriculture. Agronomy for Sustainable Development, Springer Verlag/EDP Sciences/INRA, 2011, 31 (1), pp.91-118.

[5]Aronson, Rory L. Farmbot: Humanity's Open-Source Automated Precision Farming Machine. Tech. Farmbot, 19 Sept. 2013. https://farm.bot/blogs/news/the-farmbot-whitepaper

[6]Mekhilef S., Faramarzi S.Z., Saidur R., Salam Z. (2013) The application of solar technologies for sustainable development of agricultural sector Renewable and Sustainable Energy Reviews, 18 , pp. 583-594.

[7]Fidelia, N. and Chris, B., Environmentally friendly superabsorbent polymers for water conservation in agricultural lands. Journal of Soil Science Environmental Management., 2011, 2(7), 206–21 [8]Shakti Singh, Mukesh Singh, S. C. Kaushik. (2016) A review on optimization techniques for sizing of solar-wind hybrid energy systems. International Journal of Green Energy 13:15, pages 1564-1578. 

RESULTS In this section we will delineate the endeavors of the project team to achieve the objectives mentioned in the earlier section:

1. A raised bed was built using 2-12ftx2ft, 2-8ftx2ft, 2-10ftx2ft logs with 2ft post that goes a feet into the ground ( Figure 7). Wood warping was a major issue noticed on the FarmBot proper functionality, which lead to aligning the x-axis track on a metal frame as shown in the Figure 8.

Figure 1: Robotics at UMES

Figure 4: Farmbot Web App Login Page

Figure 5: Farmbot Farm Layout and farm design

ACKNOWLEDGEMENT

The project is partially supported by Maryland Space Grant Consortium(MDSGC) and National Institute of Food and Agriculture(NIFA/USDA). The project team ( Figure 12) would like to acknowledge the assistance of several interested engineering undergraduate students, UMES farm and physical plant personnel who have helped with the set-up efforts. Significant assistance was provided by farm manager Mr. Earl Canter and his assistant Mr. Ronald Haymaker in installation of the FarmBot and developing the tunnel house to extend the growing season.

OBJECTIVES1. Design and implement a raised bed as a platform and install Farmbot Genesis

V1.4 XL2. Building a Hoop house around the FarmBot and install a ECO LLC 800W

Wind-Solar Generator Kit3. Implement rainwater harvesting kit.4. Utilize the FarmBot capabilities (Farm design, Farm events, Seeding, water,

weeding, Sequences, Regimens, Time-lapse Photography) to conduct a field experiment to study irrigation treatments on vegetable growth.

Figure 12: UMES Farmbot engineering team .

Figure 2: Farmbot – Farmbot bed displaying Radish and Arugula Plants.

Figure 11D : Tukey HSD results for Radish harvest (Phase 1)

4. Radish and Arugula were planted according to the layout. Sequences and Regimens associated with farm design, farm events, seeding, watering, weeding, and time-lapse photography were finalized.

Radish and Arugula were seeded on April 11, 2020. The harvesting process was divided in 2 phases for Radish. Radish harvest was done on June 8th ( Phase-I) and June 20th ( Phase –II). During each phase ten of the best Radish plants from each zone were harvested and weighed for ANOVA analysis. Due to unforeseen circumstances the Arugula harvest was delayed well beyond the scheduled plan.

ANOVA analyses was done on the harvest for both Radish and Arugula. For brevity the ANOVA results and harvest data ( Figures 11A and 11B) are presented only for the Phase I of the Radish harvest.

Although ANOVA results ( shown in Figure 11A)suggest that there was not an appreciable effect on the harvest due to the irrigation treatments. Treatment I3 ( maximum irrigation) yielded better results compared to other treatments for Radish (both in Phase 1 and Phase 2) as well as for Arugula. These results confirm that the sustainability goals and costs associated with water use will be favorably impacted utilizing the rainwater harvesting capabilities.

Figure 3: Farmbot web application of manual controls

2. ECO LLC Wind-Solar power generator was installed to generate the power requirements for Farmbot. It has three solar panels each having a capacity to generate about 140W and a wind turbine which has a capacity to generate about 400W are integrated to a battery and an invertor with a power regulation capacity of 3300W ( see Figure 9)

Significance of study

Field experiments have been designed to study the impact of different levels of irrigation on the harvest. Radish and Arugula were chosen for this trial. Autonomous and cloud computing capabilities of the FarmBot device have been utilized effectively in this study. These capabilities facilitated the implementation, following all the social distancing protocols during COVID-19 related restrictions and mandates on campus. The efforts undertaken are designed to illustrate small scale efficient farming technologies at the nexus of food, water and energy.

Significance to societySmart integration of technology can help create sustainable urban food ecosystems (UFEs) for the rapidly expanding urban and suburban population in the developing world. Technology advances in digital-enabled devices based on internet connectivity, are essential for building UFEs at a time when food production is increasingly limited especially in urban areas due to space constraints, concerns with regard to carbon emissions for energy use, and water scarcity[1]. Large sections of urban and suburban population are interested in growing their own food and are supporting the farm to table practices. The FarmBot set-up described in this poster holds promise to meet these needs.

Figure 7: FarmBot Bed :

3. For the field experiment described and analyzed in this poster water supply from the campus was used. Installation of rainwater harvesting set up is currently underway according to the following plan ( see Figure 10). Two barrels will be used in this setup, one will be placed outside and the other inside the hoop house The outside barrel will be equipped with a large size funnel to collect rainwater and the inside barrel will work as a reservoir. A pipe will connect the two barrels. The outside barrel will be at a higher elevation to gravity feed the one inside. A water valve will be installed in-between the two barrels to regulate the amount of water inflowing. The collected water will be pumped using a 12volts electric pump which will be controlled by the Farmbot .

Figure 6: Mini weight scale

Figure 9: Wind-Solar power generator and thermostat and exhaust fan and the completed structure.

Figure 10: Design to implement rainwater harvesting with Farmbot in the hoop house and Wind-Solar power generator

Figure 11E : Tukey HSD and Tukey B

Figure 11C : Irrigation treatment graph.

Figure 11A: ANOVA results of radish harvest phase 1 using SPSS.

Figure 11B : Radish Harvest average data in grams.

Tukey test (shown in Figure 11D) results confirm that the harvest with the I3 treatment was better than the other treatments (P<0.05) and the blocks had some effect on the harvest as well.

Figure 8 : Wood warping issue solved using metal frame :