NRI:FND:COLLAB: M3SoRo - Mobility and Morphing using Modular Soft Robots

点击翻译按钮获取中文摘要

The main objective of this National Robotics Initiative (NRI) award is to develop collaborative Modular Soft Robots (MSoRos) that can move in complex terrestrial and climbing environments and change size and shape. A swarm of MSoRos could be used in disaster relief (search and rescue operations), space exploration and precision agriculture. For example, search and rescue scenarios require small robots to autonomously navigate holes and to crawl through narrow cracks/spaces. The collaborative MSoRos will be composed of soft individual units that can deform to penetrate these spaces without prior programming. In agriculture, where the environment is complex, unstructured (soil) and adverse (changes include heat-cold and rain), these robotic modular devices will be capable of multiple behaviors to match their tasks. For example, individual modules could crawl around locally to monitor soil-health and then re-configure as a three-dimensional ball to roll to a centralized station after the task is complete. The ability to form different structures in this way can minimize locomotion costs. Furthermore, this research is easy to disseminate among high-school and undergraduate students as soft robots are cheap, safe to operate and intriguing. The MSoRos will excite young minds by connecting popular robot icons such as Transformers or Big Hero 6, with real-life morphing soft robots. Simultaneously, it will introduce them to futuristic robotics and mechatronics technologies with applications to wearable robotics, collaborative robotics and robots-in-homes, and encourage them to pursue career in STEM and robotics.This combination of morphing and modularity can dramatically increase the adaptability of a robot. The proposed research will: a) learn principles for robot mobility in complex environments. This is analogous to building reduced-order models (ROMs). Environment-specific ROMs for a highly deformable, soft, continuum robot will be reverse-engineered by learning factors that dominate robot-environment interactions; b) design open-source, untethered MSoRos that will increase the versatility, robustness and cost effectiveness of traditional modular robots; c) establish that environment awareness is a powerful strategy for controlling deformable robots. MSoRos will use their interaction with the environment to learn and deploy appropriate behaviors. This will lead to the development of hybrid robots that will combine environment-centric exploratory learning (this research) with existing model-centric strategies, to carryout complex autonomous tasks.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.