Application of Kalman Filter in the Design of Inverted Camera-Based AGV Maneuver Control

A mobile robot is a type of robot that can move. This type of robot is also often identified as a type of robot that has wheels. One type of mobile robot is the AGV (Automatic Guided Vehicle) robot. In this research, a prototype of an AGV robot that is remotely controlled using vision sensors has been made. Sensors are installed in reverse with the aim of getting a broad visual. The inputs used are direction and degree of position (turn). The Kalman filter method was chosen to correct the data due to noise that occurs when there is a change in lighting that changes the turn value. Tests were carried out to check the intensity value of light and the degree of position after and before filtering. The intensity level when the robot marker color training is used is 85 lux. The Kalman filter algorithm is very important in this study, with the results of the data obtained becoming more stable. In testing the target detection system, the X coordinate value is obtained between points 225 and 244, while the Y coordinate value is obtained between points 20 and 34. After using the Kalman filter algorithm, a stable value is obtained at points 233 and 235 for the X coordinate and 26 and 27 for the Y coordinate.