Department of Aerospace and Mechanical Engineering, University of Southern California (USC), Los Angeles, CA 90089-1453, United States of America.
Bioinspir Biomim. 2019 Feb 15;14(3):036004. doi: 10.1088/1748-3190/aae7bb.
We present the design, fabrication, modeling and feedback control of an earthworm-inspired soft robot capable of bidirectional locomotion on both horizontal and inclined flat platforms. In this approach, the locomotion patterns are controlled by actively varying the coefficients of friction between the contacting surfaces of the robot and the supporting platform, thus emulating the limbless locomotion of earthworms at a conceptual level. Earthworms are characterized by segmented body structures, known as metameres, composed of longitudinal and circular muscles which during locomotion are contracted and relaxed periodically in order to generate a peristaltic wave that propagates backwards with respect to the worm's traveling direction; simultaneously, microscopic bristle-like structures (setae) on each metamere coordinately protrude or retract to provide varying traction with the ground, thus enabling the worm to burrow or crawl. The proposed soft robot replicates the muscle functionalities and setae mechanisms of earthworms employing pneumatically-driven actuators and 3D-printed casings. Using the notion of controllable subspace, we show that friction plays an indispensable role in the generation and control of locomotion in robots of this type. Based on this analysis, we introduce a simulation-based method for synthesizing and implementing feedback control schemes that enable the robot to generate forward and backward locomotion. From the set of feasible control strategies studied in simulation, we adopt a friction-modulation-based feedback control algorithm which is implementable in real time and compatible with the hardware limitations of the robotic system. Through experiments, the robot is demonstrated to be capable of bidirectional crawling on surfaces with different textures and inclinations.
我们提出了一种受蚯蚓启发的软体机器人的设计、制造、建模和反馈控制,该机器人能够在水平和倾斜平面平台上进行双向运动。在这种方法中,运动模式通过主动改变机器人与支撑平台之间的接触表面的摩擦系数来控制,从而在概念上模拟蚯蚓无肢运动。蚯蚓的身体结构是分段的,称为体节,由纵向和环状肌肉组成,在运动过程中,这些肌肉周期性地收缩和放松,以产生向后传播的蠕动波,同时,每个体节上的微观刷毛状结构(刚毛)协调地伸出或缩回,以提供与地面不同的牵引力,从而使蚯蚓能够挖掘或爬行。所提出的软体机器人采用气动驱动执行器和 3D 打印外壳,复制了蚯蚓的肌肉功能和刚毛机制。利用可控子空间的概念,我们表明摩擦在这种类型机器人的运动产生和控制中起着不可或缺的作用。基于此分析,我们引入了一种基于仿真的方法来综合和实现反馈控制方案,使机器人能够产生前进和后退运动。在所研究的可行控制策略中,我们采用了基于摩擦调制的反馈控制算法,该算法可实时实现,并且与机器人系统的硬件限制兼容。通过实验,机器人被证明能够在具有不同纹理和倾斜度的表面上进行双向爬行。