This paper addresses a force control method that optimizes the joint-space viscoelasticity to realize a task-space viscoelasticity. In previous reports, we proposed an optimization method considering the whole-body dynamics, setting Frobenius norm of inverse of viscoelastic matrices as the evaluation function. In this paper, on the other hand, we test another evaluation function, setting Frobenius norm of viscoelastic matrices. Results of forward dynamics simulation show that the proposed method has a good performance when soft joint viscoelasticity is set as a reference. Moreover, we found that the proposed method resulted in the smallest Riemannian geodesic distance between the reference and optimized stiffness, compared with previous methods. We also validated the proposed method by an experiment using electro-hydrostatically-driven manipulator Hydracer.

Suction pads such as vacuum pads or Bernoulli pads are used for pick-and-place tasks in the industrial field. When the object is conveyed at high acceleration with the suction pad, sometimes a drop occuers due to the acceleration exceeding the limit of contact force. We made a system to detect the acceleration limit during transfer by slip detection. We have verified experimentally whether the limit of contact force can be detected by sensing. We conducted an experiment to maintain transfer by reducing acceleration when partial slip was detected.

There is a need for a robot to execute many tasks. As one of the solutions, conventional research reveals the way to save and reproduce the motion data including both position and force information. Robots therefore can execute a certain task by saving the motions demonstrated by a human using a robot, and expand the area of the task by saving many motions beforehand. However, the methodology of dealing with many saved motions, which is represented as time series of position and force, is still unclear. In this paper, we focus on writing motion and present the way to deal with many saved motions of writing those demonstrated by a human utilizing a generative model. The generative model consists of a neural network, which is known to have the ablity to approximate any non-linear function, and its parameters are learned using a variational autoencoder (VAE). We present the structure of the VAE for learning many saved time series of the writing motion and how we deal with these motions.

Until now, we have proposed a combined control of admittance and impedance controls which can expand the stable region compared with the traditional admittance, or impedance control method. In this paper, the relationship between the change in each mechanical property, such as admittance and impedance, and the stability is analyzed using the frequency-domain analysis when the proposed controller is applied to a 1DOF linear mechanical system. The effectiveness of the proposed controller is demonstrated through the qualitative analysis of the vector locus of the transfer function of the system.

In this paper, we propose a high-speed and accurate operation system for an air hockey robot by master-slave control. In this system, the position of the operator’s hand is predicted using a Kalman filter, and the robot is controlled based on the prediction. In addition, several visual presentation techniques using AR can help the operator.

In this paper, we proposed learning-based motion planner for autonomous mobile robot which outputs discrete motion commands by referring from the onboard RGB image and target position without using map. It is difficult to get policy for navigation of mobile robots using only visual sensor. In recent years, research on visual navigation using deep neural networks has been conducted. However, these methods do not have a structure to store time-series information in a network. Therefore, this paper presents the DNN motion planner with LSTM trained by using expert trajectories. As a result, we show that our method can navigate agent in simulation environment without map.

The collective behavior of animals has attracted considerable interest in various fields, such as computer science, robotics, and complex networks. When designing collective behavior using an artificial system, it is difficult to predict which individual behavior leads to the desired collective behavior. In this paper, interactive evolutionary computation is applied to generate collective behavior by automatically adjusting individual parameters based on the designer’s subjectivity. The Boids model is used to simulate the collective behavior. The results show that the interactive evolutionary computation was sufficient to generate collective behavior without the designer having specialized expertize.

In recent years, cooperative work using multiple robots, motion strategies were designed to adapt in advance to the environment. However, there is a problem in the fact that conventional method cannot function when encountering unexpected disturbance. In this research, we propose an environment adaptation algorithm based on the imitation of individuality and the maintain of diversity as human social recognition. For each individual, a local evaluation is prepared and an operation strategy is designed. The local evaluation value changes according to the speed of each individual and the distance from the goal. Based on this value, personality imitation can be realized. In this paper, we introduced fixed motion parameters and grouped of individuals, constructed a face-to-face traffic simulation, and examined its effectiveness. As a result, it was confirmed in the number of collisions between individuals decreased.

Each robots in a multi-robot system has to preserve the connectivity of the communication network structure for accomplishments of their tasks, under a risk of robot failures. This paper addresses the problem and proposes a distributed control law to solve the problem using an eigenvalue of a perturbed laplacian matrix. The definition of the perturbed laplacian, the control methodology, and some simulation results are detailed in the context.

This study tackles the task for swarm robotics where robots explore the environment to detect many targets. When a robot detects a target, the robot must be connected with a base station via intermediate relay robots for wireless communication. In our previous results, we confirmed that Lévy flight outperformed the usual random walk for exploration strategy in an indoor environment This paper investigated the performance of Lévy flight on many targets detection problem in an outdoor environment through a series of real crawler robots experiments. The result suggests that the swarm robots were not able to find all the target in a given time and there were several positions of the targets which the robots were not good at finding.

Swarm robotics is the study of a large group of autonomous robots that emerges collective behavior from local interactions. In This paper, neuroevolution is applied to design a controller for a robotics swarm. Neuroevolution is a technique to optimize neural networks with artificial evolution. The task of the robotic swarm is to achieve consensus on the color that occupies more of the two-colored environment. The robotic swarm shows the strategy to explore the environment while disseminating the one opinion, and stay in one place while supporting the other opinion.

To understand the behavior of living things, technology to measure natural behavior is necessary. In particular, measurement of long-term behavior is indispensable to elucidate global behavior. In this paper, we develop a Visual Servo that compensates for the movement of small creatures. We have shown from experiments that the response performance of the developed measuring instrument has been dramatically improved.

Recently, welfare vehicles are widely used. However, it is difficult for users to pick up luggages and to open doors using welfare vehicles. These tasks can be supported by the mobile robot with robot arm. When users operate the robot, it is desirable to give instructions as simply as possible. By using HMD, users will give instructions for the robot intuitively. However, the working coordinates pointed by users that is obtained by Mixed Reality(MR) on the HMD are unknown for the robot system. In this paper, an instruction system of running path for the arm-mounted assist robot using HMD is constructed. The running accuracy of the robot system is evaluated by experiments.

In schools and other educational institutions, there are many opportunities to carry classroom chairs according to purpose. Some chairs are on casters, but educational institutions typically use stacking chairs or folding chairs. Although effective in terms of storage, it must be lifted by hand during transportation, increasing the burden on workers. Automation of baggage transport in warehouses has improved significantly, but little attention has been given to chair transport automation. Self-propelled chairs were developed, but despite the need, automatic transport of chairs without casters was not noted. In this study, we constructed an automatic chair movement system using an omnidirectional mobile robot. Recognize the chair by image processing and bring the robot closer. After the robot dives under the chair, the chair is lifted and transported using a lifting device attached to an omnidirectional robot.

In this study, we propose a tele-operated mobile manipulator which carries out underfloor works which include not only inspection but also repairment. As one of such works, we focus on the caulking operation and we develop a mobile manipulator which finds and repairs the cracks of the foundation. When we design the mobile manipulator, we consider typical narrow underfloor environment and went through detailed calculation. As user interface, we furnish an image pointing system which enables us to operate the manipulator simply by clicking on a hand camera image. Then we evaluate its performance by running tests and trajectory tracking tests. Overall, the robot runs well. As to the manipulator, it can track the vertical line correctly, but it cannot yet track other lines precisely enough. There are some more points to be improved.

We are now facing labor crisis especially in logistics area because of the quickly shrinking population of working-age. Therefore, we must make logistics more efficient by utilizing robotics technologies. This paper proposes a method to plan local and global paths more effectively using hierarchized “maps”. By using this method, an autonomous mobile robot can plan its own local path using the information that cannot be observed just by itself. And more, the management platform of the autonomous mobile robots can prevent them from crashing or deadlocking each other, and can plan more efficient global paths as a whole.

3D jobs (dangerous, dirty, demanding) are expected to be replaced by autonomous robots. At a theme park, surveillance, cleaning, and guiding tasks can be regarded as 3D jobs. This paper aims to develop an autonomous tour guide robot system and co-experience system at a large theme park. One of the characteristics of the tour guide robot we developed is the use of QZSS (quasi-zenith satellite system) and CLAS (centimeter-level augmentation service), which realizes the centimeter-level positioning for autonomous service robots. On the other hand, co-experience realizes the sharing of experience through the robot utilizing the 5th generation mobile communication system. Experimental results at the theme park showed that the tour-guide experiment was successfully carried out and the co-experience system can provide sharing of the experience of the robot to the user.

This paper describes a new way to inspect and clean pipelines by carrying wired sensors such as cameras, and washing hoses, through long, narrow, curving pipelines. To meet this demand, we propose a drive method improvement by making the supply line slimmer, which can be expected to simplify the overall structure, save air energy, and improve responsiveness. In order to confirm the effectiveness of the proposed method, a propulsion experiment was conducted with a horizontal pipe and a vertical pipe, and it was confirmed that the response was improved even if the total length was increased, suggesting the effectiveness of the proposed method.

In recent years, with the declining working population, machines that can cooperate with humans in various industries have been introduced. Among them, there is an increasing demand for automated guided vehicles in the transport of goods in factories. In this study, we developed an omnidirectional cart with automatic and manual modes. This enables direct push-pull operation by disconnecting the electromagnetic clutches. Furthermore, a self-position estimation system was constructed. We have also developed a path following function for any route. A route recording function records the trajectory of the cart in the manual mode as route data. By integrating them, it was possible to automatically drive the route or the reverse route that the person moved by hand. Finally, experiments were performed to evaluate these. From the results of the experiments, it was confirmed that these systems worked properly.

Gymnasiums are daily inspected but small wood splinters from the floor are often overlooked because these wood splinters are very small compared to gymnasium floor. Those wood splinters occasionally cause serious accidents. Currently, gymnasiums are inspected visually but it costs a lot of people or time or both. The purpose of this research is developing equipment which inspects the gymnasium floor in a short time, with a small number of people, with low inspection omissions. In order to accomplish these goals, combination of rectangular scan and sub goals is adopted. The result of the running test through the gymnasium (range of depth 37 meters width 27 meters) is as follows. First, total inspection time was 87 minutes. Second, operation required one person when setting and no person when running. Third, inspection omission was at most 2.4%. In the future, to shorten the time, we will stabilize at high speed driving and optimize localization system.

We have proposed and verified a small mobile robot that can travel on rough terrain and can be used for research on driving control and application verification. In our application research on mobile robots, we found that we needed a mobile robot with a robot arm. In this paper, we propose a robot arm system that can be easily installed with minimal design changes for mobile robots. Then, we report the outline of the prototype of the robot arm we developed and the result of mounting the robot arm on a mobile robot.

We propose robot system for restocking task at retail stores. At the store, restocking robot needs to watch out for the floor environment and pay attention to the restocking product. Therefore, our proposed robot system adjusts robot max velocity and acceleration by the floor environment and restocking product. These Parameters are adjusted by previous task data, stability recognition result of restocking product and task success or failure. We validated the effectiveness of the proposed robot system by an experiment where a moving dual-arm robot restocked products at retail stores while adjusting robot max velocity and acceleration by the floor environment and the restocking product.

The robot JelaC-3, developed to perform various tasks by this laboratory, has an articulated dual-arm arm in an omni-directional mobile robot mechanism with mecanum wheels. This paper reports on the ROS (Robot Operating System)-based control system developed to participate in the Future Convenience Store Challenge (FCSC) display and disposal task of World Robot Summit 2020 (WRS2020). Developed System composed products recognition use realsenseD435(RGB-D camera) and ArUco marker (AR marker), the mobile robot’s Self-location estimation and moving route planning by gmapping package and navigation packages, Planning robot arm’s orbit by Moveit! (planning Framework), Human detect use Lidar (Range sensor).

In recent years, users of online shopping have been increasing. Along with this, problems such as shortage of staff of the delivery company and long working hours have occurred. Therfore, the burden on the delivery company is increasing. As a solution to the problem, there is a development of a transport robot adapted to various delivery environments. In this study, we develop a half-drone inverse pendulum transportation robot that integrates a wheel and a drone. This robot can travel efficiently in both wide and flat places and narrow places such as indoors. It also has few restrictions on the load capacity and cruising distance. In this paper, the first prototype of the half-drone inversed pendulum transportation robot for the inversed pendulum mode was implemented. The attitude and speed control experiments of the inverted pendulum were performed using the primary prototype. Consequently, the validity of the system was verified.

This paper presents a survey of the Tsukuba Challenge 2019. The Tsukuba Challenge is an experiment for autonomous navigation of mobile robots on public walkways. Pedestrian paths through the city are given as the navigation tasks. Participating teams develop their own robot hardware and software. We investigated the results of the real world experiment for all teams.

In recent years, vehicles equipped with automatic driving functions and unmanned transport robots have become popular. Along with this, running stability and speeding up of the running of the robot are required. In general, when guiding a robot, it is common to follow a line for guiding the robot, such as a white line, using an optical sensor such as an infrared phototransistor. When using an optical sensor, the distance from the line must be short, so if there is a step, the line may be lost. This time, we have created a high-speed line following robot that uses a camera to keep a distance between the line.

This paper analyzes the relation between the number of ultrasonic beams and that of required microphones to reduce the measurement time by utilizing the grating lobes. An ultrasonic phased array device for environmental recognition of human-support robot is being developed by the authors. The goal of this research is to develop the ultrasonic phased array device that can scan a wide range of environment in high resolution. We proposed a method to estimate the object position by combining a microphone array with a phased array in the previous study. Experimental results showed that it is possible to estimate the object position. However, there are situations that estimation is impossible. In this paper,the number of microphones required is discussed to resolve the situation.

When using a mobile robot in an environment where there is a possibility of submersion, it is necessary to detect the water hazard as an obstacle. Since the water hazard has different characteristics from ordinary obstacles, there is a method using polarization as its detection method. However, when a still image was used, the detection became difficult due to the movement of the water surface. In this research, we propose the water hazard detection method using the polarization video and the background subtraction method, and also utilize the polarized information to enhance the accuracy of the stereo measurement by reducing false matchings.

In this paper, we describe a viewpoint planning method by predicting observation results using deep neural networks. We use two networks for viewpoint planning. The first network predicts the size of a newly observed area, while the second one predicts an updated map after observation. We develop a look-ahead viewpoint planning using these results. We compare the proposed method with a geometrical method and a method which uses the first neural network to consider only the current state. The proposed method is the most balanced method among them.

The life span of bridges is thought to be about 50 years, and in recent years these social infrastructures have seriously deteriorated and have become a social problem. Development and introduction of existing and new robot technologies are being promoted as a policy for this social infrastructure aging problem. This research proposes a relative position measurement system that can be used with an infrastructure inspection robot. In particular, in order to solve autonomous control under non-GPS environment, which is a technical issue of UAV, position and orientation measurement using image information is performed using projection markers. The main aim of this study is to develop a system for autonomous movement by estimating the self-position and posture using projection markers.

In this research, we focus on estimating the awareness which a person have of surrounding other people and objects. We deal with the robot behavior generation considering the awareness by deep reinforcement learning to investigate the effectiveness of estimating the awareness. We compare generated behaviors with and without awareness information and find that the use of human’s awareness information reduces the number of collisions and the time to reach the goal.

The Nakanoshima Challenge is being held to develop an autonomous mobile robot that collects garbage for the Osaka Expo Pavilion in 2025. Autonomous mobile robots were developed to participate in this challege and run in public such as the actual city parks.

We developed the Navit(oo)n, which equipped two 2D LiDAR with a 1-axis gimbal mechanism. This gimbal mechanism is effective on inclined surfaces to increase the robustness of the its localization on SLAM. The gimbal mechanism corrects the angle of LiDAR and prevents erroneous detection of the ground to build a robust autonomous mobile system for environments with many disturbances such as the natural public parks. The system performances were evaluted in Nakanoshima and Ogimachi parks.

An intelligent cleaning robot for indoor environments is proposed. Traditional cleaning robots do not have access to real-time information about the room’s status. Hence, efficient dirt detection is not possible. We solve this problem through dirt detection using an external camera that can communicate with the robot. The external camera detects the dirt, and notifies the information to the cleaning robot. The robot can then start cleaning the room. The proposed method enables cleaning robots to have an access to a birds-eye view of the environment for efficient cleaning. This is very efficient compared to traditional approaches used in cleaning robots.

In this paper, the walkability of a pedestrian moving past an autonomous mobile wheelchair is evaluated based on pedestrian measurements. In the past, the mainstream method was to use a subjective impression of a questionnaire to evaluate after an experiment. By evaluating the time series of the pedestrian's ease of walking, it is possible to reflect it on the appropriate control timing of the wheelchair. From the measured trajectory of the pedestrian, the temporal variation of the moving velocity vector is used as an index of the ease of passing, and the validity of the index is evaluated. An experiment was performed under three conditions with different predictability of the behavior of a pedestrian in a wheelchair, and the evaluation method based on the walking trajectory was verified. It was confirmed that became smaller. In addition, as a result of investigating the relationship with subjective walkability, a negative correlation was found, indicating the possibility that it was useful as an index of walkability.

In recent years, in Japan, the birthrate is decreasing and aging of the population are increasing. So mobility support for elderly people who have difficulty in walking are required. We aim to the mobility support using a mobile robot and considered a real-life moving from a facility to home through indoor and outdoor environments. Robot needs to get its own position whether indoors or indoors. In this study, we propose positioning system that can use both indoor and outdoor environment. We combine ultrasonic beacon and GPS receiver to get position. And switching sensor indoor to outdoor, we got the seamless location. Installing the proposed system on a mobile robot, we evaluated the position information.

Surface robots are applied to many applications due to its hyper degrees of freedom. In this paper, we will introduce an example of applying Circle Packing Mesh, which is a mesh of connected circles, to a surface robot, and we will focus on the S-Isothremic surface that can be easily embodied as a robot in the Circle Packing Mesh. S-Isothermic surfaces are useful in the theory for surface robots because the orthogonality of the local coordinate systems can be visualized. In this paper, we will show the results of developing the robots.

An ionic polymer metal composite (IPMC) actuator is one of the electroactive polymer devices. To realize arbitrary 3D-shaped IPMC actuators, we have proposed a novel fabrication method of IPMC actuators. With the support of paper or fabric, IPMC actuators of any three-dimensional shape can be manufactured. In this paper, such IPMCs are called PF-IPMCs. Using the proposed method, a sheet-shaped PF-IPMC actuator have been created from several kinds of paper and fabric. We have compared these basic properties and operations with normal IPMC. In addition, we have selected materials suitable for making three-dimensional shape from the results of the experiment, and have produced three-dimensional PF-IPMC robots. We have conducted motion tests of these robots.

The neck of Giraffe has excellent characteristics as a long robot mechanism. The neck of giraffe can move violently represented in necking, a motion striking necks each other. Further, in a necking motion, a neck of a giraffe wards off impacts and adapts to the shape of the opponent. On the other hand, conventional robot mechanisms have hardly achieved both powerfulness and kinematic and dynamic flexibility to the opponent. In order to apply these excellent characteristics of necks of a giraffe to robot mechanisms, we developed a half-size musculoskeletal robot prototype based on anatomical knowledge of the neck of giraffes and the skeletal shape of an actual giraffe. The musculoskeletal robot protype is actuated by thin McKibben pneumatic artificial muscles, which bent flexibly, and we confirmed adaptability to an external force.

Birds with flexible feathers achieve stable flight under unpredictable disturbances. In this study, aiming at enhancing the robustness of the drones, the effect of the deflection of flexible feather models attached to the kiteplane on the robustness of the wing in disturbances is investigated experimentally. The bird-inspired kiteplane with flexible feather models on upper wing inspired by covert of bird wing is fabricated. As disturbances, the Kármán vortex street was generated by blower and cylinder at the upstream of the bird-inspired kiteplane. We measured the fluctuation of pitch angle of bird-inspired kiteplane in disturbances by using motion capture system. As a result, it is found that fluctuation of pitch angle in disturbances can be reduced by mounting flexible feather models near the leading edge. The results point out that the kiteplane can be stabilized by attaching the appropriately designed flexible feathers.

Life-size humanoid robots are expected to replace work at disaster sites, but there is little discussion about a quick introduction to the site. We will consider parachute descent as quick. In this case, we focus on landing impact, which is a big problem. A drop experiment was conducted to reduce impact by using a simple life-size one-legged robot. As a result of performing the motion after landing and examining the shock absorption, it was confirmed that changing the attitude reduced the acceleration by 50%.

We proposed a foam lost molding for fabrication of hollow rubber. First, we searched the fabrication condition using foam molds and silicone rubber. The foam is dissolved by acetone and is melted at over 70°C temperature. The heatproof temperature of silicone rubber is 300°C. Then, the curing temperature was room temperature. The molded rubber has the hollow structure nearly to the foam mold. However, the rubber has the porous structure on their inner wall, because the liquid rubber was fed into the gap between foam particles. Finally, we fabricated the pneumatic actuator using the proposed foam lost molding.

The McKibben pneumatic artificial muscle (MPAM) is flexible and can be driven like a biological muscle. On the other hand, the compressed air used for driving is released into the atmosphere after the use, so its energy efficiency is generally not very good. In this study, we modeled the MPAM based on the Port-Hamilton theory, and considered the energy efficiency and characteristics from the results of the experiments. As a result, we found that characteristics such as pressure and shrinkage ratio under the load. In terms of energy efficiency, the results showed that even if the volume of compressed air to be injected did not change, the loss decreased as the flow rate decreased. We found that controlling the flow rate can change energy efficiency.

Robot hand had used hard joint for that precise movement and strong gripping force. This research used 2 kinds of soft material for joint. The materials are Dragon Skin 30 and Elastic Resin. I made finger model by 3D printer and incorporate 3 kinds of material for joints. First one is Standard Resin , second one is Dragon Skin 30, third material is Elastic Resin. By using a flexible joint, it was able to avoid external contact flexibly and successfully protected not only the contact target but also the robot hand itself. A comparison of gripping force between a finger with a soft joint and a finger with a hard joint showed that a finger with a soft joint had 36% stronger gripping force than a finger with a hard joint. Flexible joint could keep good gripping posture.

This paper shows the simplified fabrication process of vacuum gripper with water-filled pad. In the previous fabrication, the elastomer sheet was layered and adhered. We proposed the fabrication process which the elastomer cup was folded back. The gripper was fabricated by this process. Since the adhesion area is large, the water-filled pad was not burst. The load capacity of gripper was 90 N as same as one of the previous gripper. However, the adhesion area was decreased, when the applied force is larger. Moreover, water was leaked when the applied force was 46 N, because the water was fed into pad by needle.

In recent years, soft robot grippers capable of gripping various types of objects including food materials have been studied in order to improve the efficiency of food inspection and transport operations in food factories. However, gripping finely shredded food materials with conventional soft robot grippers and gripping in environments where there is no gap between foods has been difficult. This paper presents a pneumatically driven needle gripper that can penetrate confined areas and grip shredded food materials. By controlling the positive and negative pressure of the flexible membrane, which is the drive unit of the gripper that combines a needle-shaped hard material and a film-shaped flexible material, with a pneumatic regulator, it can perform two types of operations. The gripper was gripped on the shredded food using an industrial robot arm, and the gripper mass and appearance after gripping were analyzed to evaluate the gripper performance.

This paper introduces novel soft chamber embedding stretchable conductive cloth. It measures the contacting information of the chamber in case that the soft robot equipping the chamber contacts the obstacle. The mechanism of the cloth and the chamber embedding the cloth are explained. The experimental results show that the chamber rapidly distinguishes the contacting position by measuring the capacitance of the cloth.

This study aims at developing a soft active tactile sensor that can change its sensitivity by changing sensor’s morphology. We focus on the phase change of material between solid and liquid states, which cause a stiffness change resulting in a sensitivity change. We chose gallium as a phase-changing material, because of its relatively low melting point (approx. 29°C), which means that small amount of energy is required to heat the gallium. The fabricated sensor consists of a skin and a base made from silicone rubber, gallium, strain gauges, and a Pertier element. Three strain gauges are distributed in the silicone rubber base so that the direction from which an indenter is applied can be determined. We conducted preliminary experiment and confirmed that the tactile sensor can change its sensitivity by melting the gallium and it is possible to determine the direction of the indenter.

By using a thin McKibben muscle as an active thread, fixed part of an assist suit that can automatically obtain compression. In this paper, applying the fitting mechanism with the thin McKibben muscle to the thigh fixing part of the lower limb assist suit and comparing the fitting mechanism with the conventional fixing method are reported.

In this paper, a casting mechanism using a snap motor is proposed. The proposed casting mechanism is mechanically simple, and thus can be compact and light in weight so as to be equipped with a small mobile robot. In order to evaluate the mechanism, measurement experiment of initial velocity of casting object was conducted, but it could not measure the initial velocity accurately.

The purpose of this study is to realize a robotic hand that is capable of high output force and safe contact with the environment. In this paper, we propose soft robotic fingers driven by hydraulic pressure and describe design, fabrication and evaluation of our prototype of soft robotic fingers. We perform finite element analyses to simulate a bending motion of the finger. An experimental validation of the prototype finger is also shown.