Dynamic Modeling of a Soft Laparoscope: A Deep Neural Network Approach

Abstract

Soft robotics is a research area with a diverse number of designs and shapes depending on the application. It is due to this variety that the modeling of soft robots is reduced to a few methods. However, a model based on neural networks simplifies and connects the necessary variables to perform tasks in real-time. This model relates the coordinates of the end effector in the Cartesian plane with the inputs of the soft actuator, which are the internal pressures of each chamber. In addition, a model based on neural networks considers the limitations of the system since the base data for learning is similar to real conditions data. With this approach and the piecewise constant curvature kinematic modeling, the position and orientation in the workspace of the soft laparoscope can be accurately identified.