A soft artificial muscle driven robot with reinforcement learning

Abstract

Soft robots driven by stimuli-responsive materials have their own unique advantages over traditional rigid robots such as large actuation, light weight, good flexibility and biocompatibility. However, the large actuation of soft robots inherently co-exists with difficulty in control with high precision. This article presents a soft artificial muscle driven robot mimicking cuttlefish with a fully integrated on-board system including power supply and wireless communication system. Without any motors, the movements of the cuttlefish robot are solely actuated by dielectric elastomer which exhibits muscle-like properties including large deformation and high energy density. Reinforcement learning is used to optimize the control strategy of the cuttlefish robot instead of manual adjustment. From scratch, the swimming speed of the robot is enhanced by 91% with reinforcement learning, reaching to 21 mm/s (0.38 body length per second). The design principle behind the structure and the control of the robot can be potentially useful in guiding device designs for demanding applications such as flexible devices and soft robots.

Review of Soft Linear Actuator and the Design of a Dielectric Elastomer Linear Actuator

“…Future design of the system may include a strain sensor of flexible electronics and closed-loop feedback control to enhance the accuracy of actuators. Enhanced learning based on massive experimental data can also be induced in the modeling of the complex system [78,88]. The fabrication methods for both soft active materials and flexible electronics should also be investigated to enhance the flexibility and endurance of the composite structure and hybrid system.…”

A Reconfigurable Crawling Robot Driven by Electroactive Artificial Muscle

“…To achieve real applications of such reconfigurable robot in rescue and surveillance using either swarm vibrational crawling or assembled worm configuration crawling, on-board power supply as well as control systems must be developed. With the advancement in power electronics and battery technologies, several untethered DEA robots have been developed, such as [23] [24], proving the feasibility of developing small scale untethered DEA driven robots. The attachment and detachment of the modules in current prototype was achieved manually, but in the future work, a self-reconfigurable robot can be developed by using fast transfer of angular momentum in the body of each module to perform attachment and detachment, as has been demonstrated by [25].…”

Contactless coupling of dielectric elastomer membranes with magnetic repulsion

“…DEAs are driven by large electric fields that generate a Maxwell pressure which induces large biaxial planar expansion and transverse compression. Out-of-plane biasing mechanisms can help to maximize the linear actuation stroke output from this biaxial expansion, and these include the use of a rod [2,3], spring [4,5], deadweight [6,7] or magnetic force [8,9] to deform the DEA membrane into a protruding conical shape. Out-of-plane biasing can also be achieved using an antagonistic DEA configuration, as demonstrated by Artificial Muscle Incorporated's Universal Muscle Actuator [10], which connected two offset membranes via a co-axial central disk to form a recessed conical shape.…”