Inchworm‐like robots have been increasingly investigated in recent years. However, most studies have focused on one or several locomotion modes, lacking in terms of inverted climbing, movement with a heavy load, and climbing a vertical plane. Herein, an inchworm‐like soft robot is actuated with multimodal locomotion using an acrylic stick‐constrained dielectric elastomer actuator (ASCDEA). By assembling the ASCDEA with a flexible support frame, a soft saddle‐like shape body is designed to provide the shape deformation ability for the robot. This design enables the robot to move under a heavy load. Two electroadhesion actuators are designed to implement the anchoring action. Additionally, a coordination control strategy is developed to synchronize the shape deformation and electroadhesion for multimodal locomotion. Experimental results confirm that the designed robot can move at a maximum velocity of 39.55 mm s⁻¹ (0.53 body length s⁻¹) on an inverted plane. Moreover, the robot is capable of multimodal locomotion, including inverted climbing, vertical climbing, horizontal crawling, turning locomotion, and rapid movement with a heavy load. The developed soft robot can navigate through a confined horizontal tunnel while carrying a payload, cross a gap, circumvent obstacles, and is particularly robust owing to its compliance.

The annelid, which consists of several identical segments, exploits its soft structures to move effectively in complex natural environments. Elongation and shortening of different segments produce a reverse peristaltic wave while retractable setae generate a variable friction, enabling bidirectional crawling locomotion. Although several designs have applied soft technologies towards the construction of annelid-like robots, these robots do not exhibit the homonymous segmentation, reverse peristaltic wave and variable friction. This paper reports the development of an annelid-like soft robot based on an improved dielectric elastomer (DE) minimum energy structure actuator to have these annelidan features. Each biomimetic segment of the robot is supported by a polyethylene terephthalate (PET) frame adhered to the DE actuator. The DE actuator induces segment elongation or shortening, which causes silica gel pads attached to the PET frame to contact or separate from the ground, producing a variable friction. The designed robot, whose identical segments conform to the homonymous segmentation, achieves forward or backward movement via the cooperative efforts of all the biomimetic segments. This cooperative movement, which produces the reverse peristaltic wave, strongly resembles that of natural annelidan locomotion. In addition, the Kinematic analysis of the robot is investigated. Experimental results confirm that the designed robot is capable of bidirectional and rapid locomotion. The robot can achieve a maximum velocity of 11.5 mm/s and a maximum velocity/mass ratio of 86.25 mm/(min·g). Compared to other existing annelid-like soft robots, this designed robot exhibits a superior average velocity, velocity/length ratio, body length/cycle, and velocity/mass ratio, and its performance affords the best approximation to that of the natural annelid.

The extreme learning machine (ELM) has been extensively studied in the machine learning field and has been widely implemented due to its simplified algorithm and reduced computational costs. However, it is less effective for modeling data with non-Gaussian noise or data containing outliers. Here, a probabilistic regularized ELM is proposed to improve modeling performance with data containing non-Gaussian noise and/or outliers. While traditional ELM minimizes modeling error by using a worst-case scenario principle, the proposed method constructs a new objective function to minimize both mean and variance of this modeling error. Thus, the proposed method considers the modeling error distribution. A solution method is then developed for this new objective function and the proposed method is further proved to be more robust when compared with traditional ELM, even when subject to noise or outliers. Several experimental cases demonstrate that the proposed method has better modeling performance for problems with non-Gaussian noise or outliers.

The Takagi-Sugeno fuzzy model (T-S fuzzy model) has been successfully applied to a wide range of problems due to its accurate modeling ability. However, its rule redundancy and dimension disaster remain an open and unsolved issue. In this paper, a novel fuzzy modeling method is proposed to reduce the fuzzy rule number without a loss in modeling accuracy. It includes two novel mechanism: the rule fusion mechanism and the space projection mechanism. The rule fusion mechanism is developed to merge several rules into one rule according to similarity. In this way, the number of rules will have a great decrease. After rules are merged, the relationship of data is strongly nonlinear, which causes the local linear model ineffective. In order to alleviate this difficulty, the space projection mechanism is then proposed to project the low-dimensional feature space to a high-dimension space. In this high-dimension space, the relation of data is linear and easily represented by a linear model. In this way, the traditional fuzzy inference is still applicative. With the help of these two mechanism, a fuzzy model with few rules is constructed, which enables reconstruction of the nonlinear system. A solving method is then developed to determine the optimal parameters of the fuzzy model. Additional analysis and proof show that the proposed method exhibits the uniform approximation ability for strongly nonlinear systems. Case studies not only demonstrate the effectiveness of the proposed method, but also demonstrate its superior modeling performance in comparison to several conventional fuzzy modeling methods, even with the use of fewer rules.