Tsinghua University: Latest research progress in flexible sensing technology for soft robots

New sensor products [Tsinghua University: Progress in soft flexibility research]

Recently, Assistant Professor Juntian Qu's Ocean Soft-Robot and ligent Sensing Lab (OASIS-LAB) at Tsinghua University Shenzhen International Graduate School published a frontispiece review paper in the field of soft robot flexible sensing. This work comprehensively summarizes the smart materials and advanced manufacturing methods for flexibility, and summarizes different types of sensing modes in detail. Then, the latest progress of flexible sensing technology in soft robot applications was systematically sorted out. Finally, the article deeply explores the existing challenges and development prospects in this field. This review aims to provide important guidance for the development and practical application of flexible sensing technology for soft robots in the future.

Figure 1 vanced Functional Materials: Volume 34, Issue 29

Soft robots have the advantages of good durability, flexibility and deformability. They can adapt to unstructured environments and perform various complex tasks. They show great application potential in disaster relief, health care, human-computer interaction and other fields. However, the dynamic model of soft materials is much more complicated than that of rigid joints, which brings great challenges to the shape and position control of soft robots. Perception is crucial for soft robots. In order to better understand biological systems, flexible sensors need to be integrated into soft robot systems to obtain proprioception and external perception. Although significant progress has been made, flexible sensing technology is still in its infancy. One of the main problems is the limitations of flexible sensors in practical applications. The performance of sensors may be affected under different temperature, humidity and chemical environments. Therefore, the development of flexible sensors with high stability and multimodality is a mainstream direction. In addition, the sensor must be able to stretch, bend and deform with the robot without hindering its free movement, while maintaining soft characteristics during the sensing process, which leads to the sensor obtaining nonlinear, high-dimensional, and redundancy. To solve these problems, it is necessary to carry out complex modeling and analysis of sensor data and continuously explore effective signal processing methods. By combining with flexible sensing technology, soft robots can better utilize feedback information and complete tasks safely, efficiently and accurately in complex and extreme environments.

The authors first summarized the flexible sensors made of different soft materials (Figure 2), mainly including elastomeric materials, conductive gels, bio-derived materials and composite materials. The performance of flexible sensors can be further improved by developing new smart materials and rationally designing the sensor microstructure.

Figure 2 Flexible sensors made of soft materials

This article summarizes representative flexible sensor manufacturing technologies (Figure 3). 3D printing, due to its versatility, is becoming increasingly popular in the production of flexible sensors. Through direct ink writing, multiple materials can be combined with complex geometries to manufacture functional devices in one go. In addition, 4D printing shows application prospects in the manufacture of deformable thin film structures. There are currently two main strategies. One is to use smart materials directly in the 3D printing process, and the other is to create components and structures to achieve controllable changes in shape and behavior. In addition, scalable large-scale production of flexible sensors can be achieved using methods such as fiber fabrics and screen printing.

Figure 3 Flexible sensor manufacturing technology

In order to reliably perform the predetermined tasks, soft robots need to identify the shape and position of each part to better perceive themselves and interact with the environment. The article summarizes the different perception modes of soft robots (Figure 4). In the motion control of soft robots, proprioception is the key issue for the robot to understand its own motion state. One of the most commonly used methods is to embed stretchable strain sensors in the cavity channel to convert the shape of the soft robot into a measurable signal change. At the same time, adaptive control algorithms are used to achieve precise closed-loop control of soft robots. And sensing external stimuli, effectively exploring the unknown world, and safely interacting with humans and the environment can better adapt to various complex tasks. By integrating flexible sensors, soft robots have the ability to perceive the external environment of humans, further improving their intelligence and autonomy. In addition, relevant scholars have also introduced multimodal sensing and sensor-driven integration technology into the field of soft robots to improve their ability to perform complex tasks, thereby solving some specific problems.

Figure 4 Soft robot proprioception and external environment perception

In recent years, the integration of flexible sensors and soft robots has brought the performance of task operations to a new level. This article introduces the application scenarios of soft robots based on flexible sensors from the fields of smart agriculture, exploration, and human-computer interaction (Figure 5), such as automated package sorting, fruit and vegetable growth status monitoring, and marine resource exploration.

Figure 5 Practical applications based on flexible sensing technology

The author of the article comprehensively investigated the research progress of scholars at home and abroad in the field of soft robot flexible sensing technology. Looking back on previous research, the author proposed the following three challenges that still need to be overcome:

1) Durability and stability: Soft robots usually work in unpredictable and complex environments. How to ensure the durability of flexible sensors and their ability to withstand continuous deformation while maintaining sensing accuracy is one of the main challenges.

2) Sensor integration: When integrating flexible sensors with soft structures, it is necessary to consider material compatibility, connections, and sensor layout. Ensure that the sensor is tightly integrated with the robot and can adapt to the robot's deformation and movement.

3) Adaptability and multimodal sensing: In order to improve sensing performance, flexible sensors need to adapt to different shapes, curvatures and surface features, and meet the requirements of multi-modal sensing. However, there may be signal cross-interference between different modes, which seriously affects the accuracy of measurement.

In order to solve the challenges faced by flexible sensing technology in the application of soft robots, in-depth research can be considered in the following aspects in the future:

1) Advanced material development: Considering the compatibility and interaction of different materials, exploring multifunctional composite materials with higher durability and flexibility is an important way to advance the development of flexible sensors. In addition, developing new soft materials with higher biocompatibility such as biocollagen and gel materials for the preparation of implantable flexible sensors.

2) Innovative structural design: The layered structure design is adopted to stack different functional layers together to achieve the integration of multiple sensing functions. The multi-layer structure can also provide better mechanical strength and durability. In addition, by coordinating the mechanical properties of the sensor and the main material, the seamless combination between the sensor and the software structure can be ensured, which can further improve the overall performance of the system.

3) Intelligent integration: After the technology is integrated with flexible sensors, it has powerful data analysis and intelligent decision-making capabilities. First, the algorithm can be used to extract features from massive sensor information for training and modeling, and to achieve pattern recognition and classification tasks. Second, by combining reinforcement learning algorithms with classical control methods, the soft robot has adaptive learning capabilities and can adjust its behavior and control strategies according to changes in the environment and tasks.

Statement: This article is reproduced for the purpose of conveying more information. If there is an error in the source marking or it infringes your legitimate rights and interests, please contact us and we will correct or delete it in a timely manner. Thank you.