Human-robot interaction (HRI) is becoming ubiquitous where both humans and robots perform tasks, while reliable robotic sensors are the prerequisite for efficient and safe HRI, especially in unstructured or dynamic environments. A wide spectrum of robotic sensors has been developed but most of them are limited to single or dual functionality, making it challenging to perceive complex environments.
This study, led by Dr. Xiaodong Wu and Dr. Zhimeng Liu in Sichuan University of China, presents a new type of intrinsically soft robotic sensors with quadruple sensing functionalities integrated into a single device, including spatial approach sensing, thermal approach sensing, thermal touch sensing, and mechanical force sensing. Through such quadruple sensing functions, both thermal and mechanical stimulations can be well resolved in both contact or non-contact manners. More importantly, all components of the robotic sensors can be fully recycled for reuse upon the sensor's end of service, achieving superior cost-efficiency and eco-sustainability. As demonstrations, a close-loop intelligent HRI system is constructed via integrating their intrinsically soft sensors with pneumatic soft grippers and programmable robotic arms. A diversity of reliable HRI scenarios (e.g., human-robot interfacing, object perception/classification, bedside clinical care, etc.) are successfully demonstrated leveraging the quadruple sensing functionalities.
This study presents a new path to enrich robotic sensing functionality and enhance HRI reliability in complex environments.
Science Bulletin
Experimental study