Powdered materials are often used in various industries. So, low shear and high-speed conveyance with at a low temperature (Heat is not generated) is required. In previous study, we developed a transport device based on the intestinal tract and demonstrated successful transport of the highly viscous and solid-liquid mixed fluids. In addition, it is possible to convey some powder by peristaltic device. In this paper, we developed a tubular peristaltic conveyor that improved shape of the conduit and the chamber for high-speed powder conveyance.

This paper proposes a novel sequential pattern recognition method enabling calculation of a posteriori probability for learned and unlearned classes. To discriminate undefined classes via model parameter estimations using given learning samples, probability density functions of unlearned classes are incorporated in a hidden Markov model. This method can be applied to various pattern recognition problems such as the motion classification with electromyogram (EMG) signals and the automatic discrimination for diagnosis support. In the experiments, artificial time series data generated from hidden Markov models and EMG patterns measured from forearm muscles were classified to validate classification performance of the proposed method for learned and unlearned classes. The motion classification using EMG signals were conducted with five subjects and eight forearm motions including four learned motions and four unlearned motions, were performed. Compared with previous methods, the proposed method provided a higher classification performance on learned (artificial data: 100%; EMG patterns: 91.5%) and unlearned (artificial data: 99.5%; EMG patterns: 89.6%) classes. From the results, the effectiveness of the proposed method was demonstrated.

The goal of this study was to construct a multi-legged robot to provide stable operation on the seabed. It would be difficult to reach the work destination only by walking movement, since the ocean floor is often an irregular area owing to the rocks of various sizes and also there are many obstacles such as marine resources in the shallow area. We propose a remotely operated vehicle (ROV) that avoids obstacles on the seabed by swimming locomotion. In this paper, we implemented an algorithm to control the posture control and the output direction of thrust by moving the center of gravity (COG) of a radially arranged six-legged octopus-motion-like ROV (RASOL), and demonstrated the swimming locomotion in a three-dimensional manner. As a result, RASOL could swim by flapping its legs and move toward the goal in a three-dimensional manner by controlling the position of the COG.

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 real robot experiment. This paper investigated the performance of several Lévy flight formulations on a sweeping task through a series of computer simulations. The results suggest that the approximated Lévy flight formulation outperforms the others.

The model of human's decision process by Bayesian Inference cannot represent the sudden change of the decision manner. Featuring novel Inverse Bayesian Inference, we modeled the decision process including sudden change, and evaluated the effectiveness through psychological experiments. In the experiment, the subjects progress in zigzag toward the goal in virtual 3D space by selecting the left or right. As a result, many unreasonable and sudden change decisions are observed. We show our model of decision process featuring Inverse Bayesian Inference can represent such a sudden change. We also examined examples applied to trends in stock prices and the influence of parameters of Inverse Bayesian inference.

As the issue of mental stress in severity, researchers have started to make efforts to give healing through the interaction with robots. Most of robots used in therapy developed for interacting in a state to be held. Therefore, the relative evaluation between the expected weight and the weight sense when holding the robot (hereinafter called “subjective sensory weight”) is likely to affect stress relief effects. In this paper, we examined influences of the subjective sensory weight for stress relief effects. According to the assimilation-contrast theory, in consideration of the adaptation-level varied with the users, we divided participants into the assimilation group and the contrast group and carried out the experiments. Psychological evaluation results showed that the reduction in the depression, the anger and the confusion score in the assimilation group were significantly higher than those in the contrast group. Physiological evaluation results showed that the increase in the percentage of alpha waves in the assimilation group was marginally significantly higher than that in the contrast group. These results suggest that it is necessary for improving stress relief effects by interacting robots to set the weight of the robot which was neither more nor less than users had expected.

To realize human-like dynamic manipulation, robotic juggling has been studied so far, in which simple end effectors were used. The purpose of this research is to achieve juggling with a multi-fingered hand-arm like a human. In the case that a multi-fingered hand arm is used, it becomes a serious problem that the weight of the multi-fingered hand is heavy, and it is necessary to optimize the movement of the arm in order to compensate for the hand's heavy weight. In this paper, we propose a method to optimize motion trajectory using a smoothing spline. Parameters of the smoothing spline are optimized offline using a genetic algorithm. The robot hand-arm performs an optimum throw-up operation online according to the falling position of a ball using the optimized parameters. We show the effectiveness of the proposed method by experiments, in which the hand-arm realized throw-up and catch motions in seven consecutive times. Moreover, we show that the proposed method reduces the magnitude of the joint torques and reduces the tracking error on the target trajectory, as compared with conventional methods.

In this paper, we describe an eccentric oscillating speed reducer equipped with a shock absorbing mechanism and a humanoid robot to which the mechanism is applied. One problem with humanoid robots is that the reducer is sometimes broken down by the impact force at the time of falling down. In order to solve this problem, we propose a novel eccentric oscillating speed reducer with a shock absorbing mechanism, and verify its effectiveness by an experiment. In addition, we develop a humanoid robot of the same size as an adult human using the speed reducer and verify the possibility of walking. We also measure the behavior of the joint angle and the duty ratio of the PWM control during walking and discuss the influence of the shock absorbing mechanism from the obtained data.

In this paper, we propose a novel method for identifying hand poses by emitting vibrations actively, toward intuitive user interfaces. The proposed method is an active acoustic sensing method, which can simultaneously identify hand poses and provide haptic feedback. We used the power spectrum of the propagated vibration as a feature to identify when the hand pose changed, and data compression was applied to improve the precision of hand pose identification, using a support vector machine. We evaluated the recognition rates, which depend on the number of estimation targets and the compression ratio. Further, we confirmed the feasibility of our proposed method through evaluation experiments. It was found that the recognition rate of seven hand poses was around 90%, and there is little reduction in the recognition rate when haptic feedback is provided.

Rehabilitation in convalescent stage is important to recovery of paralysis. The purpose of this study is to develop a wearable assist device like normal clothes to assist rehabilitation for paralyzed patient. This device called ‘Power Assist Wear’ is constructed with body and lower limb supporters, McKibben type artificial rubber muscles. This wear has many advantages such as flexible structure, light weight, and so on even though assistance force is not so high. In this paper, the structure and principle of this wear are described, and then control method to swing up a lower limb based on lower limb model is proposed. Proposed method is verified by using lower limb mannequin. Finally, assistance effect is confirmed from experimental results in which subject wears this device actually.

The purpose of this study is to develop a wearable power assist robot to decrease a shoulder burden in shipbuilding work. The developed device is constructed with exoskeleton, disk brake mechanism and pneumatic artificial rubber muscles, which drive disk brake mechanism. This exoskeleton is constructed with multi-link mechanism and rotational joint for shoulder flexion / extension. The constitution of exoskeleton is different from a human joint. Disk brake mechanism is installed on rotational joint of exoskeleton. Pneumatic artificial rubber muscles apply driving force for this brake mechanism. By driving brake mechanism, this device can apply shoulder flexion torque to human. In this paper, the structure and principle of this device are discussed, and then frictional torque model of proposed brake mechanism is described, and is verified experimentally. Finally, effectiveness of this device is verified experimentally.

In recent years, precision positioning devices containing piezoelectric elements (PZT) with displacement amplitude mechanisms have been used in various fields. Numerous positioning devices have been developed thus far, however, the aim of these was to increase the displacement amplitude. We have developed highly precision positioning devices that use PZT and lever mechanisms. In contrast to existing devices, our devices are characterized by high speed positioning owing to a high resonance frequency. However, because this device uses the elastic deformation of a hinge structure, the resonance peak is considerably high. Therefore, we attached a damper composed of a viscoelastic material (VEM) and a restriction plate. As a result, the resonance peak became very low, but the gain property was reduced by several decibels in the low frequency range and this was a factor of the overshoot in the detailed analysis conducted in our previous study. Moreover, we proposed the modeling of a gain reduction area by the high-order phase lag elements, however, the modeling error was large, occasionally, because of the slope of the gain characteristic. In this research, we propose a method to model the gain reduction part using fractional-order transfer functions and demonstrate the effectiveness of the model by control experiments. Consequently, it will be possible the highly precision positioning to improve the transient response.

This paper is aiming at a new water saving technology in agricultural crop production, in which minimum water required for crop growth exist only in root zone. In this study, we confirmed the zone of water retention during the growth of plant through observations of plant growth and soil water retention. This paper reported the basic concepts and methodology of this technology which controls moisture only in root zone, and also reported the method to observe the soil water retention and plant growth while conducting this water saving technology. The results and knowledge obtained from the experiment were then discussed.

This paper describes recognition of pedestrian traffic light at crosswalk for a mobile robot using deep learning. In order for a robot to cross a crosswalk, the robot needs to recognize the color of the traffic light like a human. A recognition of traffic light by camera images based on manually designed image features is possible. However, it requires significant amount of labor to adjust parameters under changing lighting condition. Therefore, in this paper we tried to recognize traffic lights using deep learning. The proposed method consists of two processes: a detection of traffic light and a classification of a traffic light color using deep learning. These processes can be processed in an allowable time by a small computer mountable on a mobile robot. Through a series of experiments, the proposed method successfully recognizes traffic signal in real environments.

This study aims to develop a precise sensor to realize a real-time measurement system for ground motion or dynamic behaviors of large structures. Our proposed sensor requires no electric power supply with using optical interferometry, which has an advantage of long-term measurement. In our previous study, we have confirmed the principle of the sensor. However, the sensor detects the motion of the inside mass part, not the motion of the sensor package. This paper describes an estimation method of the sensor package acceleration for practical use. A mathematical model of the sensor is built with spring-mass-damper system for the estimation of the sensor package acceleration. We compared the estimated acceleration with a result of servo accelerometer, and they showed good agreement. This means the possibility of practical use of the proposed sensor as a real-time accelerometer.

Portability is required to compressed air supply system in wearable device driven with pneumatic actuator. Reducing weight of air valve is effective to improve portability of air supply system. It is not easy to balance portability with performance. In general, commercial valves, which can pass large compressed air, are heavy and have large size. Therefore, the purpose of this study is to develop a lightweight valve for wearable device, which can pass large compressed air. By using pneumatic soft actuator to drive this valve, lightweight air operated valve, which has large effective sectional area, can be realized. In this paper, structure and principle of operation are described, and then developed valve is evaluated based on estimated effective sectional area. Finally, pressure in actuator, which is actually used for power assist device, is controlled using developed valve.

This paper describes a component-based tabletop robot system controlled by a self-contained micro controller. The system has reliability and reproducibility for construction of the prototype system. The implementation of example RT systems and tools shows feasibility of the proposed system.

Servo systems are widely used in various position control applications from general industry machine to semiconductor machine. In these systems, it is usually difficult to adjust the controller to high response, because there are mechanical vibrations caused by mechanical resonances and phase delay in high frequency domain. This paper proposes an auto-tuning method of vibration suppression FIR filter using adaptive FIR filter. The auto-tuning method is performed by setting coefficients of adaptive FIR filter outside of control loop to vibration suppression FIR filter in turn. Simulation and experimental results show that the proposed method can suppress the vibration, and can improve stability and response of servo system. Thus, the proposed method can provide auto-tuning of vibration suppression filter under various operating conditions.

This paper proposes a new framework to design LQ optimal control laws robust to parameter variation. By constructing the variational system of the original state space model with respect to a model parameter, we evaluate the sensitivity of the system behavior with respect to parameter variation. Some properties of the variational system, such as stability and controllability, are examined. A sufficient condition to realize robust LQ optimal control is obtained by investigating the controllability of the variational system. A numerical example exhibits how the proposed control method works with controllability analysis.

In this paper, we propose a filter to satisfy velocity and acceleration limits for arbitrary input signals. The shapes of input signals are sometimes converted into trapezoidal wave in order to prevent an overload on the robot in FA field. Generally, physical protection of equipments, safety and ride quality are achieved with limitations for input signals about some properties. For achieving such limitation, the filters which limit some signal properties are widely used. A filter structure to satisfy velocity and acceleration limits was designed only simple feedback control with the saturation function. However, this structure had a problem that time-lag between the input signal and modified signal is occurred. To solve this problem, we propose a filter structure which is based on the model error compensator. Modified signal which is generated by the proposed filter satisfies the desired signal constraints for any input signals. By the proposed structure, time-lag can be smaller than the signals obtained by the previous method. The effectiveness of the proposed filter is shown by the numerical examples.