Multistability towards reprogrammable mechanical devices

Recently, there have been advances in achieving more intricate control over structures by harnessing instabilities to transition between states with distinct programmed behaviors. For example, arrays of multiple bistable units enable the encoding of path-dependent switching between various deformation modes in response to a single input. Further, geometric nonlinearities in the kinematic motion of linkages can be exploited to create tunable multi-welled energy landscapes with non-abelian minimum energy pathways. Notably, the multi-welled energy landscape of the linkages can be easily modified through geometry adjustments triggered by environmental energy inputs, enabling autonomous adaptation. As a practical demonstration of this platform, we create a four-degree-of-freedom robot capable of navigating mazes and avoiding obstacles — all without the need for computational intelligence and utilizing only a single actuator.