Implementación de un robot móvil para desplazamientos en ambientes no estructurados empleando visión artificial.

Abstract

The aim of this research is to implement a mobile robot to move in unstructured environments using artificial vision. The system consists of stages such as: Hardware design, Software, power and the communication stage. A prefabricated crawler chassis was used on the design of Hardware, its performance and strength allows it to move on uneven surfaces, inside it carries an Arduino Mega 2560, Raspberry P13B, ultrasonic sensors, Kinect sensor, direct current (DC) motors and a FPV camera. The next stage was the software design. For the processing, the Python software was used, its programming lines analyze if there is an obstacle thanks to the camera incorporated in one of its USB ports. For the controller, the Arduino software was used, the same software that receives the signals from the microprocessor, so that this device sends commands to the actuators (motors) to avoid the obstacle. An Arduino Nano was incorporated to avoid saturation in the communication stage. Finally, two 11.1V and 2200mAh batteries were used in the power stage, as the whole system consumes a current of 2883.3mA. Through the implementation of the robot, the aim is to provide assistance in the search for evacuation routes, optimizing the rescue time. The results obtained from the implementation of the mobile robot determined the fulfillment of the requirements raised as: the autonomy time of the system, being approximately 1 hour and 10 minutes, the optimal distance of recognition and evasion of obstacles is 40-120cm and 20-70cm respectively, with an efficiency of 96,67% in a mixed stage with distractors. In addition, the final cost of $959.60 is accessible to people interested in the device.