Evaluation of inexpensive wireless sensor networks for managing small vegetable farms

View the project final report

Commodities

Practices

• Crop Production: irrigation
• Pest Management: integrated pest management, weather monitoring
• Soil Management: soil physics

For maximum productivity managers of small vegetable farms growing a wide diversity of crops need current information on conditions in their fields, crops and facilities to help with timing of planting, frost and blight protection, detection of equipment failures, etc. Recent advances in inexpensive wireless sensor networks, Internet based data storage and display, and devices such as iPhones and iPads now make it possible for farmers to monitor such things as soil moisture and temperature, growing degree days, leaf wetness, temperature and humidity in real-time from wherever they are and be alerted when a situation warrants their attention. We propose to test and evaluate a variety of such sensors at Brookfield Farm, Amherst, MA, and Woodbridge Farm, Salem, CT during the 2012 growing season with regard to being practical, affordable, easy-to-use, and useful. We want to find out if and how such real-time information can help conserve water, reduce pollution, improve productivity, reduce stress on the farmer and make the farm more sustainable. This investigation will be innovative as were unable to locate any other projects that have used such a wide variety of wireless sensors, used the latest information technology, and focused on vegetable farms growing many different crops. During the project we will keep records on all installations and uses of sensors and related software and whether or not they affected farming decisions during the season. Our final report and outreach efforts will produce information to aid other farmers in adopting the technologies on their own farms.

The Problem

Information on crop, field and facility conditions is often
--difficult to observe physically such as soil moisture at 3, 6 and 12 inches, or
--inconvenient to observe, especially regularly, such as the temperatures in greenhouses and walk-in coolers, rainfall in widely separated fields, or
--requires observations at many locations to detect microclimates, i.e. different zones of growing conditions and monitoring requirements, or
--are not timely, such as when a water tank drops below a certain level, or a cooler or heater fails.

The Result
--lower productivity due to delayed detection of irrigation requirements, pests or diseases, etc.,
--higher costs due to over-irrigation, poor monitoring of heating and cooling apparatus, etc.,
--and higher stress and wasted time for the farmer having to check on things that can be automated.

The Solution
--place inexpensive digital wireless sensors around the farm to make observations at regular times, on demand or when certain limits are exceeded,
--automatically upload the observations to the Internet,
--use a PC browser, iPad or smart phone to display the observations,
--alert the farmer in real-time via email, cell phone, twitter or messaging if certain conditions occur, and
--archive the observations' seasons and years and provide advanced reporting capabilities to detect trends, and potential for optimizing energy and water use.

The Technologies
--sensor nodes consisting of a circuit board, radio transmitter and one or more sensors placed around the farm,
--a gateway radio receiver plugged into the farm PC or router with Internet access and associated software to forward sensor observations to the Internet,
--the Microsoft Azure "cloud" for storage and archiving of sensor data,
--RT Cirrus Web software which provides sensor monitoring and reporting in a PC browser, iPhone or iPad and real-time alerts via email, twitter or text messages if certain conditions occur.

We Propose to evaluate the newly available agriculture related sensors and software with regard to being
--practical,
--affordable,
--easy-to-use, and
--truly useful to a vegetable farmer.

We propose to test and evaluate a variety of such sensors at Brookfield Farm, Amherst, MA, and Provider Farm, Salem, CT during the 2012 growing season with regard to being practical, affordable, easy-to-use, and useful. We want to find out if and how such real-time information can help conserve water, reduce pollution, improve productivity, reduce stress on the farmer and make the farm more sustainable.

This investigation will be innovative as were unable to locate any other projects that have used such a wide variety of wireless sensors, used the latest information technology, and focused on vegetable farms growing many different crops. During the project we will keep records on all installations and uses of sensors and related software and whether or not they affected farming decisions during the season. Our final report and outreach efforts will produce information to aid other farmers in adopting the technologies on their own farms.

Our project will buy, install, test and evaluate a variety of sensors and associated software on our two farms.

During the course of the project we will experiment with many different uses of the primary sensors, experiment with, research and test the following aspects:
--best mounting and placement of sensors to avoid interference with or destruction by farm crew, machines and animals,
--best sensor network design so that observations may be relayed from widely separated fields,
--less expensive alternatives for specific sensor combinations,
--the appropriate combination of sensors for different monitoring scenarios,
--observation strategies for sensors including
--regular, at certain times each day,
--or on-demand, interactively requested,
--or exception based (report only when reading exceeds preset limit),
--battery life and solar recharging,
--purchase cost versus utility for farming,
--and ease-of-use.

Initial Scenarios--here are some sensor uses we plan to start evaluating as the project gets underway:

--irrigation management especially for leafy greens in sandy soils to determine when to start and stop irrigation; also compare soil moisture in different soil types found on the farm,
--frost protection--monitor micro-climate air temperature in strawberry fields in May, and pepper and eggplant fields in September and early October to alert farmer to frost possibility so protective measures can be taken,
--late blight prediction--monitor leaf wetness in tomato fields for early detection of late blight so that preventative measures such as spraying copper may be taken,
--planting timing--monitor soil temperature to aid in scheduling the planting of crops such as corn and beans,
--calculate growing-degree days (GDD) especially comparing GDD in various fields to detect micro-climate variations and to aid in timing release of Trichogramma ostriniae, an egg parasitoid of European corn borer.
--root cellar conditions--monitor relative humidity and alert farmer if not within the right range,
--walk-in cooler--alert farmer if temperature and humidity are not within a certain ranges,
--greenhouse and other heated outbuildings--monitor air temperature to detect heating equipment failure; alert farmer if temperature not within range,
--milking parlor and bulk tank room--monitor temperature and doors closed,
--cheese room--monitor for humidity and doors closed,
--animal tracking--experiment with attaching a GPS sensor to a cow to be able to track it if it escapes or is in a distant field.

Project Personnel

--Farmers
--Dan Kaplan of Brookfield Farm will be Project Director and make the required sensor related log entries as described in section 5 below
--Kerry and Max Taylor of Woodbridge Farm (now named Provider Farm) will make the entries for their farm.

--Consultants
--Larry Manire, Technical Manager, will order, configure, mount and install the agreed upon equipment in the test sites on both farms as an in-kind contribution to the project. Mr. Manire is a semi-retired computer programmer and consultant in Coventry, RI (and Kerry Taylor's father) has 46 years in the computer field. He has volunteered labor and expertise to Brookfield weekly for the past two years and has conducted the proof-of-concept project (Attachment B). He will continue to provide local sensor, installation, configuration and computer expertise for the duration of the project. All proof of concept equipment acquired by Mr. Manire will be donated for use in the project for the duration of the project.
--Tim Reilly, Sensor Consultant, Reilly Technology (http://www.reillytechnology.com/) will be the vendor for the software, and provide software customization as needed for the project. Mr. Reilly has 23 years experience as a software engineer, with 9 years specializing in wireless sensors including commercial projects for tree seedling farm monitoring, vineyard environmental monitoring, golf course wireless irrigation control systems and many industrial applications.

--Technical Advisors
--Ruth Hazard, Extension Educator, UMass Vegetable Program, Center for Agriculture, University of Massachusetts, Amherst, MA. See Attachment C for letter of commitment.
--Andy Cavanagh, Extension Educator, UMass Vegetable Program, Center for Agriculture. See Attachment D for letter of commitment.

Both technical advisors are currently involved in research projects that dovetail with our methods. They will be a great help in planning our sensor strategies and using the results.

Timeline

-Week 1-Meet at Brookfield Farm with our technical advisors and consultants and staff from both farms to plan sensor deployments based on our facilities and spring crop plans

-Weeks 2 to 3-Order required motes, agriculture boards, sensors and software, configure and install in the selected locations (L. Manire, T. Reilly)

-Daily or weekly during growing season-Log experience with use of sensor data, log equipment problems and issues (farm managers)

-Monthly during growing season-conduct on-line meeting (using Larry Manire's GoToMeeting.com subscription) with farm staff, technical advisors and consultants to review sensor readings on-line via RT Cirrus and discuss project progress, results and issues (L. Manire organize)

-Early December, Wrap-up meeting with technical advisors, consultants and staff to review results.

-December, 2012-Write final report, develop outreach materials (farm managers, L. Manire)

As we gain experience with the variety of sensors configuration changes will be made and additional sensors added as deemed appropriate. An important point is that we are using the Libelium products primarily because they are an easy experimentation and testing platform to use for this project. However, we expect our research during the project to locate less expensive options for certain sensor combinations. One option is for Reilly technology to design and manufacture lower cost, simplified versions of the Libelium sensor hardware. The price point for the new hardware may be significantly reduced due to simplification and lower costs using North American parts suppliers and higher volume production runs. Mr. Reilly, with extensive background in sensor technology will be conducting this research during the project.

For outreach we will
1) present interim results of our project at the Northeast Organic Farming Conference at UMass, Amherst, Aug 10-
12, 2012,
2) develop a 10 minute video posted on YouTube giving an introductory tour of the various sensors, radio gateway
and display software,
3) create a handout PDF "How to set up a Vegetable Farm Wireless Sensor Network" summarizing our results and
giving recommendations for farmers to set up their own networks,
4) add the handout material to the Brookfield and Woodbridge websites,
5) email the handout to vegetable farm related organizations in the Northeast SARE region.