Design and Nonlinear Error Compensation of a Multi-Segment Soft Continuum Robot for Pulmonary Intervention

Interventional lung biopsy is an effective way to diagnose lung diseases. However, traditional bronchoscopy lacks maneuverability in complex bronchi. A multi-segment soft continuum robot with full-dimensional bending capability is proposed in this study. The manipulator features a 3.5 mm outer diameter and a 2.6 mm internal working channel. Mechanical assembly, hysteresis, friction, and drive rope extension nonlinear errors lead to deviations in the kinematic model of the continuum robot. The proposed hybrid kinematic model includes a multi-segment decoupling constant curvature model coupled with a particle swarm optimization-based back propagation neural network error compensation model. The dual-joystick master-slave heterogeneous operation mode is based on configuration space incremental mapping. The operation mode is used to complete position accuracy comparison experiments of the continuum robot with the compensation model. With the error compensation model, the mean Euclidean position error accuracy of the single-segment continuum manipulator motion experiment is improved by about 60%. The mean Euclidean position error accuracy of the double-segment continuum manipulator motion experiment is improved by about 70%. Pulmonary intervention target sampling experiments are performed in a bronchial model of the lung. The experimental results indicate that the proposed multi-segment soft continuum robot has full-dimensional dexterous guidance and potential utility.