2020/12/20 by Brian Bittner, Ross L. Hatton, Bittner, Brian +3 · 2 citations
Computer Science · Engineering · #FOS: Computer and information sciences #FOS: Electrical engineering #Music Technology and Sound Studies #Robotic Locomotion and Control #Robotics (cs.RO) #Systems and Control (eess.SY) #Time Series Analysis and Forecasting #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2012.11064
openalex publication_date 2020/12/20 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Systems whose movement is highly dissipative provide an opportunity to both\nidentify models easily and quickly optimize motions. Geometric mechanics\nprovides means for reduction of the dynamics by environmental homogeneity,\nwhile the dissipative nature minimizes the role of second order (inertial)\nfeatures in the dynamics. Here we extend the tools of geometric system\nidentification to ``Shape-Underactuated Dissipative Systems (SUDS)'' -- systems\nwhose motions are more dissipative than inertial, but whose actuation is\nrestricted to a subset of the body shape coordinates.\n Many animal motions are SUDS, including micro-swimmers such as nematodes and\nflagellated bacteria, and granular locomotors such as snakes and lizards. Many\nsoft robots are also SUDS, particularly those robots using highly damped series\nelastic actuators. Whether involved in locomotion or manipulation, these robots\nare often used to interface less rigidly with the environment.\n We motivate the use of SUDS models, and validate their ability to predict\nmotion of a variety of simulated viscous swimming platforms. For a large class\nof SUDS, we show how the shape velocity actuation inputs can be directly\nconverted into torque inputs suggesting that systems with soft pneumatic\nactuators or dielectric elastomers can be modeled with the tools presented.\nBased on fundamental assumptions in the physics, we show how our model\ncomplexity scales linearly with the number of passive shape coordinates. This\noffers a large reduction on the number of trials needed to identify the system\nmodel from experimental data, and may reduce overfitting. The sample efficiency\nof our method suggests its use in modeling, control, and optimization in\nrobotics, and as a tool for the study of organismal motion in friction\ndominated regimes.\n