Implementation of a Kalman Filter for Noise Reduction on the INA219 Current Sensor

The existence of a sensor is very essential and is needed by a digital system to produce the right decision. But every use of a sensor will have an error and noise process that cannot be avoided, and greatly affects the accuracy of the measurement results. One of the popular sensors in the market place is the INA219 sensor made by Texas Instrument which is often used to measure dc current. But the results of the experiment show that the sensor cannot be used directly because its output is relatively unstable or fluctuates. This paper implements the Standard Kalman Filter to reduce noise on the INA219 sensor to produce more accurate dc current measurements. The experimental results show that the variance value of the KF output is much smaller than the sensor output, so that the Kalman filter algorithm has worked optimally to produce accurate DC current measurements. Besides that, Kalman Filer is also very suitable for use in DC motor control as a dynamic system that implements duty cycle (D) changes in the PWM method on its DC input voltage. The Output of Standard Kalman Filter which has been implemented with the ESP32 are influenced by the values assigned to the noise sensor covariance matrix (R) and process noise covariance (Q). With a value of R=100, for resistive loads the values of Q=0.5, while for inductive load (DC motor), the values of Q=0.05 have effectively reduced the amount of noise to enhance the accuracy and precision in using the INA219 current sensor.