Soft robots have been extensively researched due to their flexible, deformable, and adaptive characteristics. However, compared to rigid robots, soft robots have issues in modeling, calibration, and control in that the innate characteristics of the soft materials can cause complex behaviors due to non-linearity and hysteresis. To overcome these limitations, recent studies have applied various approaches based on machine learning. This paper presents existing machine learning techniques in the soft robotic fields and categorizes the implementation of machine learning approaches in different soft robotic applications, which include soft sensors, soft actuators, and applications such as soft wearable robots. An analysis of the trends of different machine learning approaches with respect to different types of soft robot applications is presented; in addition to the current limitations in the research field, followed by a summary of the existing machine learning methods for soft robots.

In this study, a soft sensor-based three-dimensional (3-D) finger motion measurement system is proposed. The sensors, made of the soft material Ecoflex, comprise embedded microchannels filled with a conductive liquid metal (EGaln). The superior elasticity, light weight, and sensitivity of soft sensors allows them to be embedded in environments in which conventional sensors cannot. Complicated finger joints, such as the carpometacarpal (CMC) joint of the thumb are modeled to specify the location of the sensors. Algorithms to decouple the signals from soft sensors are proposed to extract the pure flexion, extension, abduction, and adduction joint angles. The performance of the proposed system and algorithms are verified by comparison with a camera-based motion capture system.

Soft pneumatic actuators (SPAs) based on inherently compliant materials have been developed by integrating with a fabric, a fiber reinforcement, or applying origami patterns. However, the fabric and fiber-reinforced SPAs are limited by the relatively large volumes of air needed to achieve the desired motions, and conventional `origami-structured' SPAs based on simple paper folding may not be durable and provide limited force. In this letter, an origami-structured actuating module fabricated by embedding a rigid origami structure in a fabric, which is applicable to any origami patterns including Yoshimura, Waterbomb, and so on, was proposed to improve durability and supporting force. Specially, the Yoshimura pattern-based origami-structured actuating module was analytically modeled to determine the design parameters of the actuating module so that it can provide enough supporting force and range of motion of the upper limb. A wearable system with the optimized origami-structured actuating module, which is durable, adaptable, and capable of supporting the upper limb motion, was developed to assist the upper limb without interfering natural movements.

Soft pneumatic grippers, which are based on soft pneumatic actuators have been widely studied owing to their simple morphological structure, inherent compliance, and pliable grasp. Additionally, the integration of the soft gripper with various sensors to improve its functionality has also been extensively studied. Although the soft gripper is known to exhibit a robust grasping performance without accurate control, the grasping quality of the soft gripper has rarely been studied due to the lack of adequate embedded sensors and quality metrics of the soft gripper. Therefore, a hybrid gripper, which is a soft gripper with rigid components, was sensorized by embedding a soft force sensor and a bending sensor to evaluate the grasping quality. Furthermore, a new grasping quality metric for a soft gripper was proposed, which calculates the largest-minimum wrench of a convex hull in the wrench space. The proposed grasping quality metric was experimentally verified, and a real-time program was developed to evaluate the grasping quality.

In this study, a wearable sensor system was developed to measure finger motions, which consisted of soft and stretchable silicone filled with the electro-conductive liquid metal, called eutectic Gallium Indium (EGaIn) alloy. Due to the light weight and highly stretchable properties of the silicone, the wearable sensor system allowed natural movements for the users, and it was able to adapt various hand sizes. In the proposed system, the flexion and extension of index and middle finger, and the abduction of index finger could be measured. The accuracy in finger motion measurement was verified by a motion capture system.