On the kinematic modeling of a class of continuum manipulators

This paper addresses the forward kinematic model (FKM) of a continuum manipulator, namely the Compact Bionic Handling Assistant (CBHA). This model is an essential step in the design of an autonomous robot control. The CBHA is considered as a concatenation of three degrees of freedom (DoF) parallel robot with 3UPS-1UP (Universal-Prismatic-Spherical) joint structure along the backbone. A mathematical solution of the inverse kinematic problem of such manipulators is easy to derive; however, the forward kinematic problem is mathematically intractable. Therefore, a FKM of the CBHA based on the IKM of each parallel robot is proposed. A neural-network-based hybrid approach that solves the FKM with a good degree of accuracy is developed in this paper. The proposed approach is validated by real measurements obtained by using an industrial manipulator (Kuka) as an external sensor. A comparison with a quantitative geometric approach, according to the model accuracy is also presented.