Optimizing the Substrate for Hypercube‐Based Neuroevolution of Augmented Topologies to Design Soft Actuators

Nowadays artificial intelligence (AI) has been applied to many areas, including digital game with various genres. One of them is the fighting game. Implementation of AI in the fighting game aims to make players who are not constrained by the limiting factors which humans possess, such as speed of reflexes and hand gestures. One platform that can be used to make fighting game AI is FightingICE. This platform is official platform used in Computational Intelligence and Games (CIG) conference. One approach that can be used in the process of building the AI in the fighting game is opponent modeling, a method that utilize machine learning to make model based on the behavior of the opponent. There is one AI that utilize k-Nearest Neighbor (kNN) algorithm called MizunoAI. This AI has outstanding performance, but it suffers from inability to adjust its parameter automatically. To solve that problem, in this paper the algorithm used for modelling is changed to Neuroevolution of Augmenting Topologies (NEAT). NEAT is an algorithm with Artificial Neural Network (ANN) as basic architecture and Genetic Algorithm (GA) as the architecture optimizer. In the implementation, NEAT will receive six inputs of distance in the x-axis and y-axis, both players' hitpoints, and both players' stamina. Output of NEAT is the probability of all movements that can be done by the opponent. The entire output will be processed by a simulator to determine the best countermeasure based on the predictions. For testing, the AI made will be faced against four AI winners of CIG 2013, namely MizunoAI, Kaiju, T, and SejongAI. The results showed that the AI is capable to do well in modeling process. Although the performance is still behind the MizunoAI, the learning process can be seen along the match.

NeuroEvolution of augmenting topologies for solving a two-stage hybrid flow shop scheduling problem: A comparison of different solution strategies

More than 130,000 people have strokes each year in the United States [1]. Of these victims, 76% are left with disabilities that cost the nation over $54 billion in lost work and medical fees. One prominent disability is upper extremity hemiplegia, which occurs among approximately 50% of stroke sufferers [2]. Robotic technology has the potential to provide an automated platform for controlled rehabilitation and assisted, task-oriented therapy. Several systems have been designed to assist in patient articulation of an impaired hand using rigid robotic components. While these products have been successful in articulating the pure bending motions of the four fingers, the limited capabilities of rigid technologies fail to reproduce the complicated motion path [5] of the thumb during opposition grasp (see Figure 1). This is the most important articulation for normal hand function and specifically for picking up everyday objects. To date, robotic systems for thumb rehabilitation have not been widely investigated [3] apart from some recent work [4] that used a multi-joint rigid robot.

The benefits of soft robotic actuation have been recognized and soft actuators are increasingly used. Soft robots are generally lightweight, inexpensive, and highly compliant. They are often used for interaction with small objects which only require relatively low force output. Nevertheless, large-scale applications can still benefit from the advantages of soft actuators if force output is sufficiently large. Some soft actuation has been employed in large-scale applications, but the area remains greatly underexplored. This article presents the Soft Torque-generating Actuation System (STAS) a lean soft actuation architecture capable of producing relatively large torque output while maintaining compliant interaction with the environment. Design and fabrication of the system are described, and a theoretical model is introduced for predicting torque output. The model is validated against measured torque output, and the system is further characterized in terms of power output. The model is shown to adequately predict torque output at low actuator angles. The system is shown to require only 30.0 kPa to achieve a large torque output of 10 N-m. The system also reaches a peak power output of 1.29 Watts. Finally, a versatile application of STAS is shown through a wearable joint-assistive device. The STAS is a promising architecture for a new type of high force actuation using soft actuators.

Soft robots are inherently compliant and adaptive, therefore they are promising candidates for interacting with humans. However robotic manipulators utilizing soft actuators are often constrained by a series of actuator performance limitations. In this work we design a novel linear soft robotic actuator with significantly improved performances over the existing products, achieving 300% deformation ratio, quasi-constant output force over a wide motion range, while maintaining passive compliance and adaptability. Moreover, the novel actuator is less prone to friction, and could be fabricated using inject molding and 3D printing, hence having high repeatability at very low cost. An analytical model was developed to characterize the actuator behavior and provide a guideline for actuator design according to performance specifications. A 6 DOF soft manipulator was designed and fabricated utilizing the novel soft actuator. The manipulator arm had a serial kinematic structure with a biomimetic wrist and was driven by 12 soft actuators mounted onto the arm links. With 1.2m workspace radius and 1kg payload, the working air pressure could be as low as 1bar. Preliminary results have shown the validity of the novel soft actuator and manipulator designs, as well as the strong potential of soft robots in human-oriented applications.

Neuroevolution is a powerful and general technique for evolving the structure and weights of artificial neural networks. Though neuroevolutionary approaches such as NeuroEvolution of Augmenting Topologies (NEAT) have been successfully applied to various problems including classification, regression, and reinforcement learning problems, little work has explored application of these techniques to larger-scale multiclass classification problems. In this paper, NEAT is evaluated in several multiclass classification problems, and then extended via two ensemble approaches: One-vs-All and One-vs-One. These approaches decompose multiclass classification problems into a set of binary classification problems, in which each binary problem is solved by an instance of NEAT. These ensemble models exhibit reduced variance and increasingly superior accuracy as the number of classes increases. Additionally, higher accuracy is achieved early in training, even when artificially constrained for the sake of fair comparison with standard NEAT. However, because the approach can be trivially distributed, it can be applied quickly at large scale to solve real problems. In fact, these approaches are incorporated into DarwinTM, an enterprise automatic machine learning solution that also incorporates various other algorithmic enhancements to NEAT. The resulting complete system has proven robust to a wide variety of client datasets.

Decision-making of online rescheduling procedures using neuroevolution of augmenting topologies

The soft robotics field investigates biological bodies’ movements to develop soft actuation with similar behavior seen in nature. The skeletal muscles’ movements are vastly mimicked in soft actuators; however, the contractions of smooth muscles are overlooked in soft robotics. Smooth muscles provide peristaltic contractions to the digestive system, such as the stomach, and their design challenges the limits of soft robot development that simulates digestive organs. In this article, we present a novel, bellows-driven soft pneumatic actuator (SPA) with a self-sensing capability that generates linear displacement to simulate a smooth muscle segment in a stomach. This SPA is proposed as a solution for applications such as stomach simulators. The SPA concept, design, fabrication, creep model, and experimental validation are presented in this article. A lumped viscoelastic model presents the actuator's modeling to predict the SPA's displacement from a known applied pressure. The model is successfully validated, which resulted in a maximum displacement of 20.0 mm with a displacement error of 3%. A sensory system embedded in the SPA measures the displacement, and it is found that a force of 0–8.0 N is produced. A polynomial equation calibrates the sensor readings, with a displacement error of 5%. The results show the SPA's capability as a linear soft actuator that can generate a deformation like a smooth muscle segment. To simulate a circular contraction from a stomach with the SPA, we install multi-SPAs in a ring frame that performs contractions. Although the proposed SPA is specifically designed to mimic the smooth muscle, the ring actuator's contractions can be used for other applications such as gripping soft objects.

Emerging classes of soft robotics with flexible actuation, intelligent sensibility, and biomimetic functionality are driving significant advances in academic researches and commercial applications. Such new generation robotics relies heavily on important breakthroughs in soft matter engineering and flexible actuation systems. Unlike conventional rigid robots, soft robots generally have unique sporting styles and manufacturing strategies. Recently, a series of newly developed soft materials—gallium‐based liquid metals (LMs) are found to display dramatical roles in making soft machines and robotics with their unusual properties in soft robotic actuation and self‐driven field. With superior merits of both high stretchability and electroconductivity over conventional soft materials, LM is increasingly innovated as flexible sensors and actuators in constructing intelligent soft robots. Typical advances in LM enabled soft machines and robotics are summarized and interpreted, herein, with special focus on the driving principles of the LM objects, the preparation methods of LM flexible electronics for making soft robots, and the applications of LM in biomedicines and intelligent robots, together with the consideration of key challenges facing LM‐based soft robotics and its future prospects.

Conventional robots are constructed with rigid parts and materials, restricting their ability to deform freely and adapt to the external environment. Performing tasks such as handling delicate materials and navigating narrow pathways is challenging for these typical robots made of stiff materials. However, there is a growing demand for such functions in various industries. Recent advances in soft robotics have garnered considerable attention as solutions to these challenges. Soft robotic systems in various forms present a systematic solution for performing unique tasks. The goal of soft robotics research is to develop advanced technologies, incorporating innovative robotic mechanisms based on soft materials and structures to enhance adaptability. Soft robots find broad applications in diverse fields, including industry, healthcare, and field exploration. Soft component technology plays a crucial role in configuring soft robots. Soft actuators introduce new types of motion and operation mechanisms, while soft sensors measure the deformation of the actuators with a tiny stiffness change. These advancements in soft actuators and sensors encourage researchers to explore new robotic designs and applications. Additionally, the component technology enables the integration of soft and typical robots into hybrid B Youngsu Cha

Due to their light weight, flexibility, and low energy consumption, ionic electroactive polymers have become a hotspot for bionic soft robotics and are ideal materials for the preparation of soft actuators. Because the traditional ionic electroactive polymers, such as ionic polymer-metal composites (IPMCs), contain water ions, a soft actuator does not work properly upon the evaporation of water ions. An ionic liquid polymer gel is a new type of ionic electroactive polymer that does not contain water ions, and ionic liquids are more thermally and electrochemically stable than water. These liquids, with a low melting point and a high ionic conductivity, can be used in ionic electroactive polymer soft actuators. An ionic liquid gel (ILG), a new type of soft actuator material, was obtained by mixing 1-butyl-3-methylimidazolium tetrafluoroborate (BMIMBF4), hydroxyethyl methacrylate (HEMA), diethoxyacetophenone (DEAP) and ZrO2 and then polymerizing this mixture into a gel state under ultraviolet (UV) light irradiation. An ILG soft actuator was designed, the material preparation principle was expounded, and the design method of the soft robot mechanism was discussed. Based on nonlinear finite element theory, the deformation mechanism of the ILG actuator was deeply analyzed and the deformation of the soft robot when grabbing an object was also analyzed. A soft robot was designed with the soft actuator as the basic module. The experimental results show that the ILG soft robot has good driving performance, and the soft robot can grab a 105 mg object at an input voltage of 3.5 V.

Continued technological progress in robotic systems has led to more applications where robots and humans operate in close proximity and even physical contact in some cases. Soft robots, which are made of highly compliant and deformable materials, provide inherent safety features unlike conventional robots that are made of stiff and rigid components. Soft robotics is a rapidly developing field exploiting biomimetic design principles, novel sensor and actuation concepts, and advanced manufacturing techniques. In this study, we propose novel 3D printable soft vacuum actuators that are inspired by the sporangium of fern trees. These actuators that are directly manufactured using commercial and affordable fused deposition modeling 3D printers offer many advantages such as high actuation speed (5.54 Hz), long lifetime (123,000 cycles), large payload to weight ratio (∼26), and significant output forces (∼16 N). The behavior of these actuators is accurately predicted, and their performance is optimized using finite element modeling. Furthermore, diverse robotic applications such as locomotion robots (a walking robot moving with an average forward speed of vf = 3.54 cm/s, and a hopping robot called Gongaroo hopping with an average speed of vf = 3.75 cm/s), grippers, and artificial muscles have been established and activated using the new soft actuation concept. Finally, to demonstrate the modularity of the proposed actuation concept, soft actuators with multiple degrees of freedom and variable length are established using a series of 3D printable vacuum hinges.

Dielectric elastomer actuators (DEA) have gained increased attention in the design of electrically driven soft and lightweight robotic devices. They rely on Coulomb forces for their actuation and are able to undergo large displacements as a function of the applied electrical field. Fabricating DEAs with additive manufacturing opens the way to a personalized and cost‐efficient fabrication of soft active devices. In particular, extrusion‐based techniques provide a versatile strategy to multi‐material 3D print DEAs by using electrically conductive and soft filaments with high elongation. One key aspect of DEAs is the electrode's performance, which has to provide a high electrical conductivity and a good adhesion to the sandwiched membrane. Herein, selected 3D printing parameters were varied to optimize the electrode infill density and electrode infill direction. Mechanical and electrical properties, as well as actuation performance, of fully 3D‐printed DEA demonstrators were studied comprehensively. With the optimized parameter set, a maximum DEA displacement of 91% was achieved with respect to the free length of the actuator. This is comparable to the performance of non‐3D‐printed DEAs and demonstrates the potential of 3D printing for the production of DEAs undergoing large deformations while leveraging the design and manufacturing freedom of multi‐material 3D printing.

Sensing forms an integral part of soft matter based robots due to their compliance, dependence on loading conditions, and virtually infinite degrees of freedom. Previous studies have developed several extrinsic sensors and embedded them into soft actuators for displacement and force estimation. What has not been investigated is whether soft robots themselves possess intrinsic sensing capabilities, especially in the case of pneumatically powered soft robots. Such an approach, that exploits the inherent properties of a system toward sensing is called sensorless estimation. Here, we introduce sensorless estimation for the first time in pneumatically powered soft actuators. Specifically, we show that the intrinsic properties of pressure and volume can be used to estimate the output force and displacement of soft actuators. On testing this approach with a bending actuator, we observed errors under 10% and 15% for force and displacement estimation respectively, with randomized and previously unseen test conditions. We also show that combining this approach with a conventional embedded sensor improves estimation accuracy due to sensing redundancy. By modelling soft actuators additionally as sensors, this work presents a new, readily implementable sensing modality that helps us better understand the highly complex behaviour of soft matter based robots.

This paper reports on design and fabrication of a novel soft fish robot. Application of soft actuators for the fish tail will generates continuum bending motion which resembles the natural motion of the fish. However, most soft actuator mechanisms are complex and have low efficiency. Thus, to address this issue we have developed a 3D printed soft bending actuator which can be actuated with an electromotor. The basic design idea of the soft bending actuator is explained, and iteration of the design showed to create the desired motion for the soft tail. The soft actuator has been successfully integrated with fish body and it has been shown that the fish can swim.