Tendon-Driven Parallel Continuum Robots: Modeling, Characterization and State Estimation

Continuum robots are slender and flexible manipulators, that are mainly characterized by their inherent compliance, easy miniaturization and ability to bend into highly non-linear shapes. Thanks to their inherent structural properties, continuum robots are predestined for applications in confined environments and hard-to-reach locations with limited access along narrow insertion paths, such as minimal-invasive surgery or industrial in-situ inspection and maintenance. One recent development in continuum robotics is the introduction of parallel continuum robots. Through the parallel arrangement of multiple continuously bending links, these structures aim to overcome some of the inherent drawbacks of continuum robots, such as their low accuracies and strength when compared to conventional rigid-link robots. In this thesis, a novel type of parallel continuum robot is studied, which features actively bending continuum segments that are driven by tendons. Planar and spatial designs of such tendon-driven parallel continuum robots are introduced and modeling frameworks are derived to compute their forward, inverse and differential kinematics and statics. Based on these models, the kinematic properties, such as the reachable workspace, manipulability, compliance and singularity occurrences are evaluated and studied, characterizing tendon-driven parallel continuum robots on a general level. Lastly, this work focuses on an efficient state estimation approach for both conventional and parallel continuum robots that allows to estimate their state given noisy measurements.