Green Thumb is an Internet of Things project in the agricultural sector made for my class . I wanted to build something specifically for the developing nations, and upon my research I found out that the African countries have only 6% of the continent farmland irrigated, there is poor technology, less reliability on water management or irrigation leading to less productivity. In Zambia it was found that smallholders who were able to cultivate vegetables in the dry season earned 35% more than those who do not.
Most of the existing systems costs more than $200, which is expensive and certainly not affordable by Small farmers. Farmers in these developing nations are already taking efforts towards a small scale water management system.
Green Thumb's aim is to provide a cost-effective, individual, small-scale irrigation system to farmers in Africa that helps them with smart-irrigation and water management techniques to increase the quantity of their produce.
Step 1: Step 1: Implementing Moisture Sensors on a Plant
Choosing a Plant: I needed a plant to monitor over the course of my project, since many African countries grow eggplants, I ended up getting a small eggplant from home depot to experiment with.
Moisture Sensors: To monitor the moisture content of the plant you need to make a cost-effective sensor that could do so.
1. Galvanized Nails - 2
2. Single Strand Wires - a bunch of them
3. Particle Boron - 1
4. Resistor (220 ohm or any other value) - 1
Take 2 galvanized nails and solder them to single strand wires.
Make the following connection on your breadboard.
Connect any one of the nails to an Analog pin and the other one to a Digital Pin. Keep the nails 3 cm apart, it can be any distance as far as it is constant, since the distance between 2 nails can change the readings.
Write the following code in your Particle Boron IDE and flash the code
Insert the nails in your plant, it should be displaying readings on your serial monitor or your console.
Here's a quick guide for setting up your Boron.
Step 2: Step 2: Collecting the Moisture Sensor Readings
The next step was to collect all the readings in an Excel document for monitoring purpose through IFTTT.
1. Visit IFTTT and make an account (if you don't have already) or sign-in. IFTTT (if this then that) is a free-web based service to create chains of simple conditional statements called Applets.
2. Go to -> My Applets, Click on -> New Applets
3. for +this - choose Particle -> choose 'New Event Published' -> Write the 'PlantData' as the event name for which IFTTT should be triggered
4. for +that choose google sheets -> select 'Add row to a Spreadsheet' -> Write the name of the spreadsheet to be created -> click on 'Create Action'
5. So when you particle publishes the event 'PlantData', a new row of data will be added to a spreadsheet in your google drive.
Step 3: Step 3: Analyzing the Data
You can download the excel file and sample the data. I made line graphs of data collected for every half an hour, found that the readings did not change much over the given course of time. The nail sensors gave pretty much reliable readings.
The reading usually fluctuated between 1500-1000 whenever it needed to be watered.
So, considering the threshold to be 1500, we can say that when the reading is less than 1500, the plant is in it's wilting stage and the system can respond in about 5-10 minutes by watering the plants.
Also since the data previously was collected every millisecond, it corrodes the nails.
Once the data is monitored and we see that there is not much fluctuation in the readings, the sensor can be powered every one hour, collect the reading and check whether it is below the threshold.
This will allow the nail sensors to last longer.
Step 4: Step 4: Making Multiple Sensors and Communicating Through Mesh
The entire farm area can be divided into multiple regions and these regions can be monitored by individual sensors. All these sensors can communicate with the 'Main System' that controls the water pump.
The 'Main System' has Particle Boron - it is cellular, hence it can communicate at places without WiFi.
The individual sensors have Particle Xenon, they communicate to Boron by creating a local Mesh Network.
Here's a quick guide for adding your Xenon to an existing Mesh Network.
Here, I have made 2 sensors. Transfer the entire circuit to a protoboard.
Test the following code to see whether the Mesh communication is working.
Step 5: Step 5: Complete Physical Form of the Sensors
The electronics for the sensors need a box that can be deployed in the fields. Since the system had to be cost-effective, I envisioned spending on the electronics while saving cost on it's physical form. The physical box in which the sensor needs to be placed, can be made by a farmer or can be locally manufactured in Africa using their raw materials. The farmer can also use any material available to them and put the electronics inside.
I prototype using cardboard, which can be made water-resistant by varnishing.
Make a box with 8.5 cm in breadth, 6.5 cm in width and 5.5 cm in height. Cut these dimensions out of a cardboard. Make 2 holes at the bottom that are 3 cm apart for the sensors to go in. Stick the cardboard boxes with a glue-gun.
Make 2 layers of cardboard with 8.5 cm x 6.5 cm dimension, that would go inside the box. Cut out a hole in these layers for the wires to pass through.
The nails would go through the holes. A cardboard layer is placed on top of it that has the Protoboard . Crocodile clips are used to connect the nails to the circuit, so that these nails can be easily disconnected from the circuit.
The second layer of cardboard on top of this has LIPO battery that powers up the Xenons.
These layers can be removed by lifting them up with the help of the holes that are cut out and the nails can be replaced easily, this make the system easy to maintain and assemble.
Step 6: Step 6: Final Implementation
I divided a box full of soil, in 3 parts, one with maximum water, second with medium water content, and third was dry soil.
Each sensor when placed in one of the 3 parts of the box, communicates the reading to boron, which makes a decision whether that area needs to be watered. This is indicated by an LED, corresponding to each sensor.
The sensor would be powered up every one hour.
This is an entry in the