Performing intricate minimally invasive surgeries with continuum robots demands that the robot be highly dexterous within the specific anatomical context of the procedure, which in turn requires careful tuning of its geometric design parameters. This paper introduces a design optimization framework that simultaneously accounts for dexterity and anatomical constraints. We propose the Reachable Volumetric Dexterous Solid Angle (RVDSA) as the objective metric, which quantifies the robot end‑effector’s ability to reach every point in a target volume from multiple directions via collision‑free paths. Formally, $$\text{RVDSA}=\int_{V}\Omega(p)\,dp$$ where $\Omega(p)$ denotes the solid angle of reachable directions at point $p$. To evaluate RVDSA efficiently, we develop a sampling‑based motion planner that rapidly checks collision‑free feasibility for each configuration and accumulates the solid angle contribution. An asymptotically optimal simulated annealing optimizer then searches the design space to maximize RVDSA. We applied the method to a bimanual dexterous sheath robot intended for polyp removal in colon anatomy. Compared with conventional optimization that only maximizes 3‑D voxel coverage, our approach achieved an average RVDSA increase of roughly 78%.
Review