结构仿生六杆张拉整体机器人折叠控制的形态智能方法OA

Morphological intelligence refers to leveraging a robot's physical body——Its physical properties,geo-metric structure,and dynamic characteristics——To offload computation(e.g.,controller design)and enhance en-vironmental adaptability;It is a core mechanism of embodied intelligence.This paper targets complete folding of a six-bar tensegrity robot and develops a morphology-driven simplified control method that achieves whole-body equi-valent folding under partial cable actuation.A folding objective based on endpoint aggregation is first formulated;symmetry analysis then enumerates four folding patterns together with their associated cable-length variations.A graph-theoretic cycle-space analysis is employed to identify redundancy in length changes induced by geometric coupling,from which the actuated-cable set during folding is determined.Within a static framework,the mapping from motor inputs to cable-length variations is established and a reachability criterion is provided,yielding a simpli-fied control strategy for each pattern.Quasi-static MATLAB simulations and hardware experiments validate the approach:Across all four patterns,complete folding is achieved while the number of actively actuated cables is re-duced from 24 to 9.The results highlight the potential of morphological intelligence to simplify controller design for tensegrity robots.