Professor Huang Yong’an’s team from Huazhong University of Science and Technology’s School of Mechanical Science and Engineering, together with Associate Researcher Zhang Fan, Assistant Professor Zhong Junwen from the University of Macau, and Associate Professor Li Junfeng from Wuhan University of Technology, published a groundbreaking paper in Nature Communications on Oct 3, titled “AI-embodied multi-modal flexible electronic robots with programmable sensing, actuating and self-learning”.

The research introduces a modular framework for flexible electronic robots (FEbots), integrating sensing, actuation, and decision-making to endow robots with advanced environmental intelligence.

Current soft robots can crawl, jump, and roll while balancing operational safety and environmental compatibilities, yet their functionality remains limited in complex tasks such as autonomous search-and-rescue and dynamic terrain navigation. Inspired by biological systems like insects, FEbots overcome the constraints of preset movement patterns and limited real-time adaptability.

FEbots employ a distributed bristle array and oscillatory driving mechanism, simplifying control systems while maintaining environmental adaptability. Modular reconstruction through bristle density and circuit layout allows robots to switch between configurations for diverse task requirements. Coupled with flexible electronics, FEbots gain multidimensional sensing, including visual, proximity, temperature, and humidity detection, working in synergy with edge processing units.

The modular, Lego-style design enables customized configurations, giving FEbots superior environmental adaptability, multi-modal locomotion, and multi-modal perception. Low-power, high-dimensional computing enables embodied AI, granting the robots autonomy in hazard avoidance, cooperative obstacle navigation, and thermal field tracking, making them highly capable in dynamic, unstructured environments.

Source: School of Mechanical Science and Engineering of HUST