Selecting stopping locations (sites) and chassis headings for mobile manipulators requires costly inverse-kinematics and collision checks. We propose critical-gap-guided verification, which checks alternatives for covered tasks that prevent site removal. A capability atlas proposes poses and joint seeds; the planner obtains complete support, then verifies replacements and reselects layouts, seeking fewer sites before fewer headings. On a building information modeling (BIM)-derived benchmark with 1252 height-resolved normal-approach samples, three nominal 60 s runs yield 28, 30 and 28 sites. Relative to low-load transfer, which prioritizes low-workload sites for removal, median sites decrease from 34 to 28 (17.65%), while median headings increase from 51 to 61 (19.61%). Controls favor targeted replacement acquisition, without consistent gains from the priority function or failure weighting. Independent triangle-mesh checks of 1341 saved configurations from three predefined static layouts find no contacts or external-clearance violations at the thirteen checked lift heights. Under a separate whole-body criterion (WB5), the task assignments of five saved plans admit connected model paths with local base motion, without enforcing the original six-dimensional arm-Jacobian screen. Critical-gap conditional model times are 12.67–18.70% lower than transfer under shared assumptions. Clinic and office tests remain incomplete. Broad building-level applicability, tool-equipped execution and actual operating times remain unvalidated.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: