结构仿生六杆张拉整体机器人折叠控制的形态智能方法

Morphological intelligence refers to leveraging a robot＇s physical body——Its physical properties, geometric structure, and dynamic characteristics——To offload computation (e.g., controller design) and enhance environmental 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 equivalent 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 simplified 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 reduced from 24 to 9. The results highlight the potential of morphological intelligence to simplify controller design for tensegrity robots.