In an unmanned aircraft vehicle, a navigation system is needed to calculate its orientation and translation. The navigation system can utilize data from the accelerometer, gyroscope, magnetometer, and GPS. The orientation can be precisely calculated from the accelerometer and magnetometer data when the sensor is in a static state. Meanwhile, under dynamic conditions, the orientation can be more precisely calculated from the gyroscope data. In order to obtain the robust navigation system, a data fusion based on Kalman filter is built to calculate the orientation from the accelerometer, gyroscope, and magnetometer. The Kalman filter trusts more in the data from the accelerometer and magnetometer when the UAV is static and trusts more in to the gyroscope data when the UAV is in dynamic conditions. Meanwhile, the UAV translation is obtained by performing data fusion of the accelerometer data with location data from the GPS sensor. The Kalman filter combines data from the accelerometer and GPS when available, otherwise trusts in data from the accelerometer only. The trust level shifting is done by changing the measurement noise covariance. The data fusion based on Kalman filter provides more accurately the orientation and translation data. The orientation as a result of the calculation from the gyroscope has an average error of 18.12%, while the orientation as a result of the accelerometer and magnetometer has an error of 1.3%. By using Kalman filter-based data fusion, the error of the orientation decreases to 0.87%

This study aims to develop a Kalman filter algorithm in order to reduce the accelerometer sensor noise as effectively as possible. The accelerometer sensor is one part of the Inertial Measurement Unit (IMU) used to find the displacement distance of an object. The method used is modeling the system to model the accelerometer system to form mathematical equations. Then the state space method is used to change the system modeling to the form of matrix operations so that the process of the data calculating to the Kalman Filter algorithm is not too difficult. It also uses the threshold algorithm to detect the sensor's condition at rest. The present study had good results, which of the four experiments obtained with an average accuracy of 93%. The threshold algorithm successfully reduces measurement errors when the sensor is at rest or static so that the measurement results more accurate. The developed algorithm can also detect the sensor to move forward or backward.

The analog AC-voltmeter usually can only measure the ideal-sinusoid voltage with narrow frequency range. Meanwhile, in fact the grid voltage is often not in the form of an ideal sinusoidal.  To be able to measure a non-sinusoidal AC voltage with a wide range of frequency, a true-RMS voltmeter is needed. The research designed a true RMS measuring system using an ATmega 328P microcontroller. The input voltage is converted to pulse using Schmit triger and fed to the microcontroller’s external interrupt pin to calculate the input signal frequency. Meanwhile the microcontroller’s ADC sampled the input signal with a frequency of 128 times the signal’s frequency. RMS voltage calculations are performed using arithmetic operations for 16 and 32 bit integer variables. The test results show that the system can measure voltages with zero errors from 100 to 275 volts with a frequency of 50 Hz. The system can also measure voltages with zero errors at 220 volt with frequencies from 40 Hz to 150 Hz. However, this system can still be used to measure voltages ranging from 25 volts to 300 volts at frequencies from 35 Hz to 195 Hz with an average error of 0.21%. During RMS voltage calculation, the microcontroller’s CPU usage was 13.35%, so that this system can be further developed.

Abstract

Rotation angle estimates are often required and applied to the dynamics of spacecraft, UAVs, robots, underwater vehicles, and other systems before control. IMU is an electronic module that is used as an angle estimation tool but has noise that can reduce the accuracy of the estimation. This study aims to develop an estimation model for the angle of rotation of a rigid body based on the IMU-gyroscope sensor on a smartphone using a Kalman filter. The estimation model is developed in a simple dynamic equation of motion in state-space. Kalman filters are designed based on system dynamics models to reduce noise in sensor data and improve measurement estimation results. Simulations are carried out with software to investigate the accuracy of the developed estimation algorithm. Experiments were carried out on several smartphone rotations during the roll, pitch, and yaw. Then, the experimental data obtained is analyzed for accuracy by comparing the built-in algorithms on smartphones. Based on the experimental results, the accuracy rate of estimation angle is 94% before going through the Kalman filter and an accuracy level of above 98% after going through the Kalman filter for every rotation on the x-axis, y-axis, and z-axis.