Individual deformability compensation of soft hydraulic actuators through iterative learning-based neural network

Abstract

Robotic devices with soft actuators have been developed to realize the effective rehabilitation of patients with motor paralysis by enabling soft and safe interaction. However, the control of such robots is challenging, especially owing to the difference in the individual deformability occurring in manual fabrication of soft actuators. Furthermore, soft actuators used in wearable rehabilitation devices involve a large response delay which hinders the application of such devices for at-home rehabilitation. In this paper, a feed-forward control method for soft actuators with a large response delay, comprising a simple feed-forward neural network (FNN) and an iterative learning controller (ILC), is proposed. The proposed method facilitates the effective learning and acquisition of the inverse model (i.e. the model that can generate control input to the soft actuator from a target trajectory) of soft actuators. First, the ILC controls a soft actuator and iteratively learns the actuator deformability. Subsequently, the FNN is trained to obtain the inverse model of the soft actuator. The control results of the ILC are used as training datasets for supervised learning of the FNN to ensure that it can efficiently acquire the inverse model of the soft actuator, including the deformability and the response delay. Experiments with fiber-reinforced soft bending hydraulic actuators are conducted to evaluate the proposed method. The results show that the ILC can learn and compensate for the actuator deformability. Moreover, the iterative learning-based FNN serves to achieve a precise tracking performance on various generalized trajectories. These facts suggest that the proposed method can contribute to the development of robotic rehabilitation devices with soft actuators and the field of soft robotics.

Latent Representation-Based Learning Controller for Pneumatic and Hydraulic Dual Actuation of Pressure-Driven Soft Actuators

“…The control results include information on coping with individual characteristics to realize neat movements, and the FNN can efficiently learn each inverse model from the training data. 15 Figure 1B shows the process flow of the ILC.…”

“…The general characteristics of soft actuators (nonlinearity, hysteresis, 1 and individual deformability differences 14 ) render their control challenging with a simple model-based or model-free controller. 15–18 Therefore, a learning controller is currently a popular solution for soft actuator control, which has overcome the difficulties in soft actuator control. 19 A learning controller is also a potential solution for dual actuation control.…”

“…Many researchers have developed learning controllers to control soft actuators, 15 , 19–36 and achieved precise control performance. The existing learning controllers can be classified into two types depending on the amount of prior model information: a model-based learning controller and a model-free learning controller.…”

“…For example, we developed an iterative learning-based feedforward neural network (IL-FNN). 15 The IL-FNN obtains an inverse model from the control result of an iterative learning controller (ILC). The IL-FNN achieved precise tracking on generalized trajectories with an RMSE of 2.11°.…”

“…For example, an FNN optimizes its neural network to be equivalent to a control target's model. 15 The characteristic differences between pneumatic and hydraulic PSAs are too significant to compensate for with one specialized learning controller, 11–13 despite the generalization ability of the learning controller.…”