Application of artificial vision as measurement validation tests on a robotic hand driven by shape-memory alloys

In this work, a robotic hand has been developed. This hand is driven by SMA (Shape Memory Alloy) wires. Some researches has showed the high rejection rate by the amputees using prosthetic upper limbs due to several problems, such as weight, high noise and need of anthropomorphism. In this context, it is presented the development of a robotic hand which is not driven by conventional actuators, in this case, SMA actuators. The mechanical structure of the robotic hand was designed using a computer program (CAD) and subsequently manufactured in ABS (acrylonitrile butadiene styrene) polymer by a rapid prototyping machine (3D printer). The robotic hand was instrumented with a bend sensor (also known as Flexsensor) to monitor the angles formed by each phalanx. The data presented, as the obtained angles, were quite satisfactory for this application. In order to corroborate the data presented by the bend sensor, a system was developed using a camera to procedure a non-contact measurement with image processing.