Neural Network-Based Shape Analysis and Control of Continuum Objects

Soft robots are gaining increasing attention in current robotics research due to their continuum structure. However, accurately recognizing and reproducing the shape of such continuum robots remains a challenge. In this paper, we propose a novel approach that combines contour extraction with camera reconstruction to obtain shape features. Neural networks are employed to model the relationship between motor inputs and the resulting shape output. A simulation environment is established to verify the shape estimation and shape control of the flexible continuum. The outcomes demonstrate that this approach effectively predicts and reproduces the shape of flexible continuum robots, providing a promising solution for continuum shape control.