Diseño de un sistema de control visual para navegación de interiores usando fusión sensorial para una plataforma móvil diferencial.

Abstract

The aim of this work was to design a visual control system for interior navigation using sensorial fusion on a Differential Mobile Platform. The research was developed about the Raspberry Pi technology and its use with the Matlab software. The computer controls the position of the servomotor by locating the camera from 0 to 180 degrees with 45° advances to capture the environment instantly. The operations of erosion and dilatation of captured images and of the database were developed, the purpose of these operations is to clearly obtain the contours of the captures in order to verify whether or not there is an obstacle, in the visual system the correlation of images were used. To the verification of the system, an Arduino Nano card with ultrasonic sensors was used to detect if there is an object within the defined work area. If the two systems agree on the sensed information, the robot makes a decision on its displacement, activating the motors from the scrip. To verify the hypothesis, the experiment was designed in such a way that it determines the level of efficiency of the control system in various tracks that vary in the number of obstacles and the form of displacement. The ANOVA statistic was considered obtaining a result of 15.62 in the value calculated from Fisher that compared with the critical point of the statistic is 9.55 the Null hypothesis is rejected, accepting the original hypothesis of the investigation. Finally, it is recommended to carry out an investigation to solve the errors found in the visual control system due to the changes that may exist in the ambient light, that is, a lighting system that adapts to the conditions of the system.