Finite Element Analysis and Error Compensation for Wrinkled Bellow-Like Soft Robotic Manipulator Kinematics Modeling

In this paper, a finite element analysis (FEA) and error compensation method are conducted for the wrinkled bellow-like soft robotic manipulator (WBSRM) kinematics modeling. Firstly, WBSRM is designed, which has good axial expansibility and radial bending deformation. The link and restriction of the interference plate make the WBSRM have better motion effect and can be better used implemented for FEA. The wrinkled bellow actuator has good axial scalability. Then, the model of WBSRM based on the constant small curvature (CSC) methodology is established. To verify the accuracy of the established kinematics model, the FEA approach is employed, in which the local coordinate system is established to obtain the local coordinate value and the curve length is calculated by curve fitting and integration, such that the idealized constraint limit of the common actuator length can be avoided. Finally, the FEA simulation result and the established kinematics model are compared and analyzed. Concerning on the modeling error between the established mathematical model and the simulation result, the back propagation neural network (BPNN) technique is introduced to realize the error compensation. The obtained method results show that the compensation method can effectively eliminate the modeling error and then dedicates to characterize the detailed kinematics model of the WBSRM.