Biomimetic metamaterial robotic systems with predator-prey interactions

Public summary

* A biomimetic metamaterial robotic system comprises interacting predatorbots and preybots.

* The lizard-like predatorbot crawls, turns, and visually recognizes preybots and environmental targets.

* The inchworm-like preybots perceive predatorbots through hybrid triboelectric tactile-visual sensing.

* Hybrid visual-tactile sensing enables pursuit, warning, and coordinated escape between the robots.

Abstract

Natural predator-prey interactions inspire highly adaptable soft robotics, yet biomimetic soft robots lack the real-time perception needed for mimicking biological behaviors. This study presents a biomimetic metamaterial robotic system featuring lizard and inchworm robots powered by origami metamaterial artificial muscles. Combining twistable and contractile artificial muscles generates coordinated limb and spine movements. This enables the predatorbot to achieve lizard-like locomotion and low-speed maneuverability, with a forward displacement of approximately 0.25 body lengths (BLs) per motion cycle. The preybots utilize origami tribo-epidermis and tribo-antennae to mimic crawling at 0.19 BLs per second (BL/s) while detecting individual collisions. Integrating a wireless you only look once-ecosystem (YOLO-ES) module enables active visual sensing between the predator and prey. The YOLO-ES model achieves an mean average precision (mAP)@0.5 of 93.39% on the expanded validation set, demonstrating accurate recognition of key robotic and environmental targets. The robotic system activities are demonstrated by mimicking the information and physical interactions between species, where the reported robots exhibit robust tactile-visual hybrid sensing and bioinspired locomotion. This work presents a paradigm for developing intelligent, interactive soft robots with emergent behavior, offering new insights into biomimetic robotic interaction systems.