e-ISSN: 2548-6861
JAIC Vol. 7, No. 2, December 2023: 150 – 155
154
Figure 4. Output Panel of Re-calibration Application
Figure
4 shows that polynomial regression order 3 is the
best model with coefficients b
0
=2.43, b
1
=-0.0166, b
2
=4.1 x
10
-5
dan b
3
=3.4 x 10
-8
. The results of the Re-Calibration
application are the same as the previous manual
modelling
process. In the scatter plot there is also a gray background
color which is the 95% confidence interval in the actual soil
volumetric water content (VWC) estimation results.
Coefficient model configuration on IoT application.
double soilMoistureValue = 0;
double soilMoisturePredict = 0;
double b0=2.43;
double b1=0.0166;
double b2=0.000041;
double b3=0.000000034;
void loop()
{
Blynk.run();
soilMoistureValue = analogRead(A0); //put Sensor
insert into soil
soilMoisturePredict = b0-
b1*soilMoistureValue+b2*soilMoistureValue*soilMoistur
eValue-
b3*soilMoistureValue*soilMoistureValue*soilMoistureVa
lue;
}
The calibration results are then programmed into a
Capacitive Soil Moisture Sensor as in Figure 5 so that a
model-based capacitive soil moisture sensor is obtained. After
that, the prediction results are displayed on the Blynk
application so that soil moisture can be monitored on a mobile
phone. The results of the soil moisture prediction simulation
and monitoring display on the mobile phone are as follows:
Figure 5. Model-based capacitive soil moisture prediction
Figure 6 shows that the model-based capacitive sensor that
has been recalibrated can follow the pattern of gravimetric test
results (actual) better than the capacitive soil moisture sensor
with factory settings. Figure 6 shows the relationship pattern
between predictions and raw sensor data following an
exponential pattern. The predicted value of soil moisture is in
the range 0 – 1.2 for raw sensor data values of 100 – 530. This
range is then used to configure value limits on the mobile
phone display.
The model selection process for re-calibration capacitive
soil moisture sensors becomes easier with a web-based
application. The application will automatically carry out
calculations related to the gravimetric test. The results of the
modelling can be used to make predictions of soil moisture so
that soil moisture measurement can be done via mobile phone
in real time.
This application can be used on any soil types and sensors
other than capacitive sensors. It's just that there are a few
things to consider. First, the re-calibration application was set
up based on gravimetric experiments at 4 levels of increasing
water volume with 100 ml intervals. Second, because of this,
the polynomial order that can best be used is 3. Lastly,
measurements are made using grams unit.