查看文章

Soft robots represent an emerging class of bio-inspired machines mainly composed of soft matter, such as elastomers, fabrics, and fluids. Due to their inherent hyperelasticity and compliance, soft robots can adapt to uneven surfaces and distribute external forces. Therefore, they have potential advantages over traditional piecewise rigid robots in maneuvering through complicated environments. As the cornerstone of a soft robot, soft robotic artificial muscles have been developed for actuation with various mechanisms, including fluidic actuation, electrical actuation, magnetic field actuation, and light actuation. Despite dramatic advancements in soft robotic artificial muscles in the past decades, many challenges remain in the development of a soft artificial muscle for robots that is capable of swift and precise actuation, produces sufficient forces for dynamic and versatile locomotion, can be controlled with high fidelity, and can be modeled efficiently and accurately. Progress has been limited by the following critical gaps: 1) advanced functional materials in combination with material architectures that can deliver the required actuation performance; 2) mechanical designs that lower the requirement of bulky hardware for untethered operation; 3) physical understanding of the underlying nonlinear mechanics and dynamics of soft artificial muscles for control purposes; and 4) feedback control systems that connect the materials, designs, and models for practical operation of soft robots with artificial muscles.