Diseño e implementación de un prototipo de silla de ruedas para personas parapléjicas, comandada mediante reconocimiento de gestos dinámicos de la cabeza.

Abstract

The objective of this research work was to design and implement a prototype wheelchair for paraplegics, commanded by recognition od dynamic gestures of the head. This prototype consists of three parts for its operation: artificial vision part, mechanical part and electronic part. The artificial vision system consists of software and hardware. The hardware consists of a LED lamp, a camera, the Raspberry PI3B card and a 5’ LCD screen. For the software, the dynamic head gesture recognition algorithm was performed using the techniques of color conversion, contour detection and midpoint detection. In the mechanical part, the motor coupling was made to the driving wheels and an articulated support was designed to hold the artificial vision hardware. In the electronic part an integrating board was designed that contains: The Monster Moto Shield driver to control the DC motors, terminals for the battery connections and the artificial vision system. Three tests were performed on 5 people, 3 men and 2 women, aged between 10 and 54 years. The results showed that the average current consumption is 8.29A, also that the average delay time between performing dynamic head gesture and activating the motors is 2.32s. In addition, it was determined that at 25% of the total speed of the engines, the linear speed of the chair is: without load 0.361sm/s and with load 0.241m/s, this is the most suitable for the operation of the prototype, if its speed is increased, recognition is lost. A deeper study is recommended for the realization of a wheelchair lighting system commanded by artificial vision.