Integration of Kalman Filter in Digital Compass as an Improvement of Navigation Accuracy for Blind People Ridwan Solihin1, a), Sudrajat 1, b) , TB Utomo 1, c) , Febi Ariefka Septian Putra 1, d) dan Tata Supriyadi1, e)
Politeknik Negeri Bandung
Abstract
Blind people for navigation in carrying out independent activities moving from one location to another cannot rely on a combination of visual senses and memory as a map. However, blind people rely on a combination of other senses that are still functioning including touching, hearing, smelling, and using standard or electronic canes. The activity of touching by tracing walls is very risky for health safety. The activity of remembering through hearing and smelling is very likely to change because it depends on the situation, conditions, time, and interferes with concentration. The activity of using a cane is limited to providing feedback in the form of tactile, sound and vibration signals to the distance of the object in front of it. Without the help of position navigation to reach the target location, it is likely to be further away due to inappropriate route selection. The proposed research design is a digital compass with a signal output in the form of a north direction that can be felt. In this paper, the Kalman Filter (KF) process is described in the Arduino Nano microcontroller to reduce the noise of compass sensor data input due to the impact of environmental electromagnetic field interference. The KF process, error prediction by comparing the difference in current compass reading data with the previous one and the difference is used as a correction. Furthermore, the data from the prediction and correction process is issued in the form of PWM to the servo actuator as a north direction indicator. The test scenario was carried out by comparing the system without KF and using KF against the compass application on the smartphone, as well as the accuracy and distance to reach the target location. The test results show that the KF algorithm can function to eliminate noise in the compass sensor system with a better error difference of 0.78%. While the usefulness of reaching the target location functions and produces a shorter distance.
Keywords: compas sensor, Kalman Filter, north direction navigation, servo, blind people
