Distance-directed Target Searching for a Deep Visual Servo SMA Driven Soft Robot Using Reinforcement Learning
- PDF / 3,630,690 Bytes
- 13 Pages / 595.22 x 842 pts (A4) Page_size
- 58 Downloads / 196 Views
Journal of Bionic Engineering http://www.springer.com/journal/42235
Distance-directed Target Searching for a Deep Visual Servo SMA Driven Soft Robot Using Reinforcement Learning Wuji Liu1, Zhongliang Jing1*, Han Pan1, Lingfeng Qiao1, Henry Leung2, Wujun Chen3 1. School of Aeronautics and Astronautics, Shanghai Jiao Tong University, Shanghai 200240, China 2. Department of Electrical and Computer Engineering, University of Calgary, Calgary AB T2N 1N4, Canada 3. School of Naval Architecture, Ocean & Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
Abstract Performing complex tasks by soft robots in constrained environment remains an enormous challenge owing to the limitations of flexible mechanisms and control methods. In this paper, a novel biomimetic soft robot driven by Shape Memory Alloy (SMA) with light weight and multi-motion abilities is introduced. We adapt deep learning to perceive irregular targets in an unstructured environment. Aiming at the target searching task, an intelligent visual servo control algorithm based on Q-learning is proposed to generate distance-directed end effector locomotion. In particular, a threshold reward system for the target searching task is proposed to enable a certain degree of tolerance for pointing errors. In addition, the angular velocity and working space of the end effector with load and without load based on the established coupling kinematic model are presented. Our framework enables the trained soft robot to take actions and perform target searching. Realistic experiments under different conditions demonstrate the convergence of the learning process and effectiveness of the proposed algorithm. Keywords: biomimetic soft robot, SMA, deep visual servo, Q-learning Copyright © Jilin University 2020.
1 Introduction In many applications, robots made of articulated rigid bodies are unable to perform complex tasks. These applications include manipulation of fragile objects, locomotion in difficult terrain, interaction with humans and surgical applications[1]. Soft robots are made of soft materials, such as resin, and their movements are generated by deformation. They are generally regarded as complete or partial soft structure of infinite degrees of freedoms (DOFs), intrinsically underactuated and extremely adaptable to the environment[2]. Biological organisms exploit softness of their body for compliance to reduce the risk and complexity in interacting with an uncertain environment. This characteristic gives the potential to advance robotic systems to operate robustly and adaptively in unstructured environments[3]. Previous researches have paid considerable attention to building prototypes of soft robots by adopting various soft materials to address requirements in specific tasks. Electro-Active Polymers (EAPs) and Dielectric Elastomer Actuators (DEAs) are emerging as feasible *Corresponding author: Zhongliang Jing E-mail: [email protected]
materials to mimic muscle-like actuation in soft robots design[4]. As a result, how to realize an effective contr
Data Loading...
