Novel Mechatronic Systems and Soft Robotics enabled by 4D Printing and Machine Learning

Background

Four-dimensional printing (4DP) is a novel additive manufacturing technology that builds 3D printed structures with smart materials to enable the shape and/or property change with external stimuli after the printing process. Our contributions to 4DP include (1) improving the precision and robustness of 4DP methods by exploiting feedback control technologies and (2) exploring new applications for 4DP. For example, 4DP is applied to develop a recycle, remanufacture and reuse process to fabricate customized protective visors against COVID-19. Other major applications of 4DP and smart materials are for soft robotics. Soft robots are promising for personal care such as home services, elderly/health care.

Aims and objectives

This project addresses three research challenges. (1) Mature sensors and actuators are often incompatible with soft robots. Soft sensors and soft actuators using smart materials will be developed and seamlessly integrated in the soft robots. (2) Soft robots are often fabricated ad hoc in multiple steps with many devices and professional skills.

A unified fabrication procedure that can embed soft sensors and actuators inside the robot bodies is desired. (3) The control problem for soft robots is inherently nonlinear and time variant. A satisfactory controller must be robust against varying working conditions and adaptively improve its performance online.

Project plan

The contributions of this project are to exploit and improve 4DP methods to embed sensors and actuators inside robots, use smart materials for novel means of perception and actuation, and develop data-driven and learning based control methods for soft robots.

Applied interdisciplinarity

Cooperation of many researchers with distinct disciplinaries, such as 3DP/4DP, smart materials, design of new sensors and actuators, robotics, automatic control, production engineering, etc.

Papers

• Q. Ji, M. Chen, X. Wang, L. Wang, L. Feng, “Optimal Shape Morphing Control of 4D Printed Shape Memory Polymer based on Reinforcement Learning”, Robotics and Computer Integrated Manufacturing, 73:102209, 2022.
• Q. Ji, M. Chen, C. Zhao, X. Zhang, X. Wang, L. Wang, and L. Feng, “Feedback Control for the Precise Shape Morphing of 4D Printed Shape Memory Polymer”, IEEE Transactions on Industrial Electronics, 68(12), 2020.
• S.T. Muralidharan, R. Zhu, Q. Ji, L. Feng, X.V. Wang, L. Wang, “A Soft Quadruped Robot Enabled by Continuum actuators”, in Proc. 2021 IEEE 17th International Conference on Automation Science and Engineering (CASE), Lyon, France, 2021.
• M. Chen, Q. Ji, X. Zhang, L. Feng, X.V. Wang and L. Wang, “Study on Efficient Fused Deposition Modelling of Thermoplastic Polyurethane Inflatable Wall Features for Airtightness,” Proceedings of the 9th Swedish Production Symposium, pp.417-427, October 2020.
• Q. Ji, C. Zhao, M. Chen, X.V. Wang, Lei Feng, L. Wang, “A Flexible 4D Printing Service Platform for Smart Manufacturing”, Proceedings of the 9th Swedish Production Symposium, pp.575-585, 2020.
• Q. Ji, X. Zhang, M. Chen, X.V. Wang, L. Wang and L. Feng, “Design and Closed Loop Control of a 3D Printed Soft Actuator,” Proceedings of 16th IEEE International Conference on Automation Science and Engineering, pp.842-848, August 2020

KTH collaborations

• Production Engineering: Prof. Lihui Wang, Assistant Prof. Xi Vincent Wang
• Material Science and Engineering: Raquel Lizarraga Jurado
• School of Engineering Science, Teknisk Mekanik: Prof. Malin kermo, Prof. Lanie Gutierrez-Farewik, Assistant Prof. Ruoli Wang

Duration

January 2020 – December 2023

Project participants

Data-driven modeling and Control

Soft robot actuation is inaccurate. Qinglei Ji develops data-driven models and closed-loop control systems to increase the controllability for different soft robots and deploy them in different applications. Watch the examples below.