A Versatile-Designable Framework for Active and Programmable Shape-Morphing Soft Matter Systems: From Inverse Design...

Developing soft matter systems with programmability, multifunctional integration, and environmental adaptability represents a critical challenge in soft robotics and smart materials. Herein, we propose a versatile framework for constructing active and programmable shape-morphing soft matter systems based on addressable actuation and strain-constraint mechanisms. Using liquid crystal elastomers (LCEs) and conductive constraint strips serving simultaneously as geometric constraints and localized Joule heaters, this framework circumvents complex microstructural manipulation, enabling addressable electrothermal actuation and deterministic 2D-to-3D morphological transformation. Combined with the established analytical model, we propose an inverse design strategy capable of reconstructing complex target surfaces featuring spatially non-uniform curvatures. To demonstrate its integration capability, we incorporate shape memory polymers (SMPs) and crack-based sensors via a thermally decoupled design to construct a proprioceptive lockable soft robotic system (PLSRS), exhibiting zero-energy shape retention and real-time proprioception. Finally, we validate this system by deploying the PLSRS in a flapping-wing robot, where a 1D convolutional neural network (1D-CNN) optimized via the grey wolf optimizer (GWO) deciphers aeroelastic signals to estimate wind speed and trigger autonomous adaptive wing regulation. This work successfully fuses physical intelligence with computational intelligence, providing a versatile platform for next-generation adaptive soft robots.