基于前馈神经网络的蜂巢气动软体执行器的分层控制方法

Soft compliant materials and novel actuation mechanisms ensure ?exible motions and high adaptability for soft robots, but also increase the dif?culty and complexity of constructing control systems. In this work, we provide an ef?cient control algorithm for a multi-segment extensible soft armin2Dplane.Thealgorithmseparatetheinversekinematics into two levels. The ?rst level employs gradient descent to select optimized arm's pose (from task space to con?guration space) according to designed cost functions. With consideration of viscoelasticity, the second level utilizes neural networks to ?gure out the pressures from each segment's pose (from con?guration space to actuation space). In experiments with a physical prototype, the control accuracy and effectiveness are validated, where the control algorithm is further improved by an optional feedback strategy.?????