A review of sensing technologies in soft robotics

Background

The rapid pace of meeting human needs has led to widespread adoption of automation in manufacturing, packaging, medicine, agriculture, and food industries. Equipped with the realization of "Industry 4.0" technologies such as (), big data analytics, and, for monitoring and controlling the functions of these robots. Currently, rigid robots are widely adopted with high reliability, accuracy, and durability.

However, handling more delicate and fragile objects requires robots with softness, which has led to an increase in interest in soft robots. As the name suggests, soft robots (soft robocs) are made of soft materials or have a soft interface covering a rigid skeleton/scaffold. Soft robots are flexible and can quickly adapt to the shape of objects. Compliance significantly reduces the pressure applied to grasp an object, unlike rigid robots, where the contact area is limited due to the rigid form of the gripper. Soft robots are also well suited for applications such as healthcare where safety is a priority.

In addition, due to their soft and jointless bodies, soft robots can access remote areas and perform functions that are difficult to accomplish using hard robots. Depending on the type of application, soft robots can have various shapes, forms, and drive mechanisms, such as grasping, moving, underwater exploration, flying, etc. The main focus of this review is on sensors integrated into soft robotic grippers, which are becoming increasingly popular in agricultural harvesting, warehouse management, and healthcare.

Soft robots have unique properties such as large degrees of freedom, high mechanical compliance, and the ability to deform through internal actuation and external loads. This makes it challenging to accurately determine the shape and position of each part of a robotic gripper in three-dimensional (3D) space. Unlike rigid robots that rely on precise control of joints and limbs, soft robotic control requires morphological computation, which depends on the robot's morphology and material properties. This requires the use of soft materials with adaptable material properties. However, modeling soft material dynamics is much more complex than the simple kinematics of rigid joints, making it challenging to control and monitor the shape and position of different parts of a soft robot.

To overcome this challenge, it is crucial to integrate sensors into soft robots. These sensors are able to monitor and control the shape and position of different parts of soft robots. In addition, sensors can enhance the perception of external stimuli such as temperature, pH, chemicals, pressure, light, and sound in soft robots, which greatly broadens the application range of soft robots. With the help of sensors, soft robots can perform complex tasks in different fields such as healthcare, agriculture, and warehouse management. Integrating sensors into soft robots is challenging because the sensors must be able to stretch, bend, and deform with the robot without hindering the robot's free movement while maintaining its softness during the sensing process. This leads to nonlinearities, singular configurations, and non-unique mappings associated with soft sensors. Addressing these issues requires complex modeling and analysis of sensor data to accurately map environmental stimuli to sensor data. In addition, to increase the functionality of sensors, soft robots must be equipped with sensors with high spatiotemporal resolution.

However, this generates a large amount of data that must be processed quickly for closed-loop monitoring and control. With the growing demand for soft grippers, there is an increasing need to integrate various sensors such as fruit ripeness, temperature, proximity, food spoilage, pH, gas sensors, etc. Therefore, choosing the right sensing mechanism, the number of sensors, and their integration is crucial to minimize the computational load on the robot and achieve efficient sensor integration. With the successful integration of sensors into soft robots, they can perform complex tasks in various applications with improved accuracy and efficiency.

To achieve this, it is necessary to make appropriate changes in the design and selection of manufacturing processes. An emerging field that enables the integration of multiple materials into complex shapes is multi-material additive manufacturing. This technology helps to efficiently and reliably manufacture soft robots with integrated sensors. In addition, minimizing the number of steps involved in the fabrication of smart soft robots and automating the process is essential to improve the reliability and repeatability of actuation and sensing functions.

Highlights of this article

1. This review focuses on the progress in soft robotics.

2. It first introduces the actuation techniques and material selection for soft robots, and then delves into various types of sensors and their integration methods, as well as the benefits of multimodal sensing, processing, and control strategies.

3. The current market leaders in the field of soft robotics are also briefly described in the review to illustrate the growing demand for this technology.

Graphical analysis

Figure 1. Components and considerations for building an intelligent soft robot. The figure highlights the key components necessary to build an intelligent soft robot, including actuation, sensing, operating energy/fuel, and control and devices. In addition, important considerations in fabrication methods, design, control algorithms, and applications are described.

Figure 2 Driving technology of soft robots

Figure 3 Comparison of power consumption of different actuators commonly used in soft robots

Figure 4 Material properties of typical elastomers used in soft robots. (a) Young’s modulus of various materials. (b) Elongation at break and tensile strength of typical elastomers. (c) Comparison of Shore hardness and tensile strength of various elastomers. (d) Different scenarios of sensor and substrate stiffness combinations for sensor integration in soft robots.

Figure 5: axial, piezoresistive and axial sensors for soft robots

Figure 6 for soft robots

Figure 7 Various other types of sensors used in soft robots

Figure 8 Integrated multimodal sensors on soft robots

Figure 9 Integrated multimodal sensors on soft robots

Fig. 10 Various signal processing and control strategies for soft robot control

Fig. 11 Some relevant signal processing technologies in soft robotics. (a) Workflow of signal processing and subsequent actuator control based on sensor data. (b) Integration of soft electronic skin on robot fingers and hands for multimodal detection of touch, proximity, temperature, and hazardous substances (c), and corresponding electronic circuits for data processing of data collected from sensors. (d) Electronic circuits for data processing of multi-array sensors of asynchronously encoded electronic skin (ES), and (e) corresponding integration of pressure and array on the robot hand.

Figure 12 Example of closed-loop control of soft robots

Figure 13 Various applications of soft robots with sensing capabilities

Figure 14: Examples of soft robots with integrated sensors in prosthetic limbs

Review editor: Liu Qing