Implementación de un prototipo de bastón inteligente basado en redes neuronales y visión artificial para apoyo a personas con discapacidad visual.

Abstract

This research aimed to develop a prototype of a smart cane based on neural networks and artificial vision to support people with visual disabilities. The research methodology included documentary, field, and experimental research, using various information collection techniques such as surveys and interviews; the data collected was analyzed, processed, and interpreted through statistical processes, obtaining data that were used for the design of software and hardware, and implementation of the smart cane. The analysis of the necessary tools that people with visual impairments require as support was carried out, and several alternatives for the selection of electronic elements determining operation aspects, size, weight, cost, and integration with other parts, from which the design and implementation of the device were carried out, integrating elements such as a camera, ultrasonic sensors, alarms (sound and vibrator) for the detection of obstacles, and GPS to provide information about the user's location. According to functional testing carried out in block modules, the parameters for the optimal functioning of the cane were established. It is concluded that implementing the smart cane prototype was necessary to carry out several functional tests in which the appropriate elements and error correction were determined for a correct operation. It is recommended to adjust the smart cane considering the anthropometric measurements of the users; id est. the device can be modified and adapted to the height of each person with visual impairment.