Manipulation of Bio-Inspired Robot with Gesture Recognition through Fractional Calculus

This paper describes the implementation of a simulated bioinspired 3D model, which was moved through segmented gestures using fractional calculation segmentation methods. It was observed that for the increase in the detection of such gestures from this segmentation an improvement in the recognition and standardization of gesture images using the FODPSO method would be obtained. We used the AlexNet neural network for training and recognition of the processed gestures, which resulted in a simulated 3D model of a spider moved by gestures recognized through a webcam, which had an accuracy of approximately 98% recognition through the neural network.