Text-to-query research remains dominated by SQL, even though many applications require graph query languages such as SPARQL and Cypher. We argue that query language and data model choice should be tested rather than assumed. To study this, we introduce Spider4SSC, a multi-domain benchmark with semantically equivalent SQL, SPARQL, and Cypher queries, and fine-tune unbiased-T5 (uT5) models pre-trained on explicitly filtered plain text. We evaluate execution accuracy ( EX ) across query language, estimated query complexity, pre-training composition, schema representation, input length, and joint multi-language fine-tuning. In clean uT5 setting, the unseen test set shows a clear performance hierarchy. SQL reaches 29.7% EX , Cypher reaches 23.9%, and SPARQL reaches 17.2%. This hierarchy depends on query hardness. SQL outperforms Cypher on easy validation queries at 58% vs 50% EX , while Cypher performs better on hard queries at 14% vs 6%. Across languages, execution accuracy generally decreases as estimated query complexity increases. Pre-training composition has a smaller effect than query language choice and data scale. For SPARQL, injecting raw SPARQL into pre-training data slightly lowers unseen-test EX from 17.2% to ca. 16.0%. For Cypher Full Compact schemas, increasing context from 512 to 2048 tokens (reducing truncation from 14.52% to 0.52%) improves validation EX from 35.20% to 39.47%. Finally, joint fine-tuning over SQL, SPARQL, and Cypher gives a small regularising benefit, improving all three languages and raising average unseen-test EX from 18.19% to 18.63%. Overall, our results suggest that query language should be selected empirically for the target task rather than assumed in advance.

This paper describes a design and control of a compact, backdrivable knee exoskeleton using an actuator modules with bilateral drive gear. The actuator module exhibits a reduction ratio of 153.2:1 and a reverse drive initiation torque of 0.020 Nm, enabling safe body interaction and torque senseless control. To generate assist torque that adapts to the wearer’s intention, four motion classifications (STOP, WALK, STAND-TO-SIT, and SIT-TO-STAND) and gait classification were performed by using knee joint angular velocity, IMU data, and estimated external torque as input data to the LSTM Neural Network. An accuracy of 92.16% was achieved for the four motion classifications. For the five motion categories (RIGHT-LEG-SWING, LEFT-LEG-SWING, STOP, STAND-TO-SIT, and SIT-TO-STAND), including the gait categories, an accuracy rate of 87.78% was achieved.

This study explores methods for improving energy efficiency in wastewater treatment facilities to address the pressure of rising electricity tariffs. As the energy consumption by water treatment facilities accounting for about 50% of the total electricity consumption in wastewater treatment facilities, particular attention is paid to the energy utilization of blowers within water treatment facilities [1], [2]. The aim is to achieve appropriate control of aeration equipment based on the facility’s conditions and environment, without relying on plant models but using data-driven control. This approach addresses the problem of excessive electricity consumption in current scenarios that rely on manual control or simple control loops. By employing VRFT for the PID controller parameter tuning, a successful improvement in the target response performance was achieved. The proposed method achieved an error in the DO response comparable to that of conventional methods and reduced the operating costs of aeration at wastewater treatment plants by about 12%.

This paper presents a simple single-carrier PWM (SC-PWM) method based on the multi-modulation technique and using the redundant switching states of the 5-level hybrid multilevel dc-link inverter to balance the dc-link voltage. The proposed SC-PWM reduces the dc-link capacitor-voltage ripple, thereby reducing the voltage stress on the switches and improving the reliability of the inverter by eliminating all the low-fundamental frequency voltage oscillations in the capacitor voltage in the full range of power factors. Additionally, the proposed modulation strategy achieves a 2-fold reduction in the size of the output filter while reducing the total harmonic distortion of the output voltage and current. The feasibility of the proposed scheme over other modulation techniques is confirmed in a comparative analysis. Finally, the steady-state performances of the SC-PWM strategy are demonstrated in simulations and experiments.

In autonomous movement using Simultaneous Localization and Mapping(SLAM), it is necessary to build a detailed map of the environment in advance to set destinations and estimate the robot’s self-position. This requirement poses an obstacle to performing autonomous movement in locations the robot is visiting for the first time. In this study, we reduce the effort by using a simplified map built beforehand. Because the pre-built map is not detailed, transformations are performed by matching with the environment. In addition, destinations given in the simplified map are converted to the actual environment. Even when movement was deemed impossible because of obstacles during the transformation, it was confirmed that movement could continue by adding or changing the position of the destination. Simulations and experiments were conducted to confirm that the robot passed through all destinations. Thus, using this method, the robot can navigate the environment it is visiting for the first time.

T In this paper, a linear-rotary Two-Degree-of-Freedom (2-DOF) position control system was proposed for the Magnetic Lead Screw (MLS) based 2-DOF actuator. An MLS converts torque into linear force through the helical constructed magnetic field providing by permanent magnets. As the mechanical transmissions are eliminated and there is no physical contact between the rotary and linear parts, the friction of MLS-based actuators is low and the backdrivability is high. Conventional MLS-based actuators can only realize linear drive. To meet the requirement in applications where both rotary and linear drive are needed, the 2-DOF MLS-based actuator was designed. In this research, principle of the 2-DOF actuator was analyzed and a 3-mass system model was built. Based on the model, angle transmission and load disturbance observer was designed. Additionally, position control system which controls the linear and rotary position independently while compensating the load disturbances was proposed. Finally, feasibility of the control method and system robustness against load-side disturbance torque/force were evaluated with simulation.

A powered exoskeleton is attracting attention to tackle the declining birthrate and aging population. One of the important factors of a powered exoskeleton is back-drivability. Safety can be ensured by improving back-drivability, and we proposed a powered exoskeleton that applies highly back-drivable reduction gearboxes at a high reduction ratio, called bilateral drive gear. The total gear ratio of the proposed exoskeleton is 155.23:1, and the wearer's torque can be estimated without a torque sensor. This paper proposed motion intention prediction using LSTM and examined each model's performance by changing the number of LSTM layers, unit size, and input data. Finally, the experiment was conducted, and 96.12 % prediction accuracy was realized, and the worst delay time for prediction was 305 ms.

In the age of Internet of Things (IoT), exploitation of security vulnerabilities is increasing, including self-propagating IoT malware. As an answer, specific research on IoT malware is being developed. Many studies use Markov chain models of malware propagation to predict the behavior of epidemics qualitatively and quantitatively. However, most studies approximate random propagation as a simple multiplicative term and no exact derivation of the Markov chain for random propagation was done so far. Moreover, systems of malware mitigation operating at the network level are rare and the majority of proposals focus on local networks like wireless sensor networks. In this article, we present a simple derivation of the exact Markov chain for random propagation of malware. Our model assumes a binomial form, compatible with binomial distributions in stochastic studies. To validate this derivation we implemented a stochastic simulation for the simplest compartmental epidemic model, susceptible–infected–susceptible (SIS). Predictions of the proposed Markov chain match simulation results with less than 0.2% error, well within stochastic variability and much smaller than the error of literature models. To complement our model of propagation, we developed and derived the Markov chain of a new system of malware mitigation, based on grouping random devices with identified infections during malware cleaning. Our mitigation system works at the network level and counteracts the vulnerability of mass deployment of IoT devices with aggressive but calculated mass disconnection. The system is able to artificially reduce$R_{0}$(the basic reproduction number) below 1 and prevent malware taking over the network—all without changing the rate of detection.

In this paper, a novel direct-drivable RotLin (Rotary to Linear) linear serious elastic actuator is studied. The RotLin actuator converts rotary motion into linear motion through helical shaped permanent magnets on this rotor and translator. Because of removing mechanical transmissions of traditional electrical actuators, force density of the RotLin machine is high. In previous research, a load force sensor based back-driving system was implemented on the RotLin machine. In this research, a force-sensor-less back-drive control system is achieved. Firstly, magnetic force of the RotLin machine is analyzed based on linearized spring force model; then, mover side friction is measured and modeled; finally, a load force estimator is designed based on the spring force model and the friction model. Besides, the cogging torque is measured and compensated. In the end, a force-sensor-less force control system is designed based on rotor position control, and by setting the reference force to zero, a back-driving control was implemented on the prototype of the RotLin actuator. The experiment results verified accuracy of the designed load force estimator and back-drivability of the proposed system.

One of the most important tasks regarding autonomous mobile robots is to plan a path for movement. Given a pre-built map, a robot is able to follow a path that is determined manually by setting way points on the map prior to exploration. However, for cases without pre-built maps, a robot has to derive its path while traveling autonomously and analyzing the surrouding environment. In this paper, we propose a method of path planning for autonomous movement under unknown environments in the case where a pre-built map does not exist. This system successfully determines way points on a road following the skeleton extracted by combining multiple data processing algorithms.

Recently, several studies have been conducted to improve and optimize the stability of magnetic levitation (maglev) train for their development. This study investigates the advantages and disadvantages of various types of maglev trains and recent research on maglev trains, and proposes a new maglev train's model, which focuses on low power consumption and stability that are realized using two levitation methods together. To verify this model, we derive theoretical equations for the levitation, propulsion, and guiding forces from the theoretical model, and compare them with the results of electromagnetic field analysis.

In recent years, research on robots has been conducted all around the world to address the reduction of the working population and the shortages of care workers needed for the aging population and disaster-relief activities. Currently, harmonic and cycloid gearboxes are widely used as reduction gearboxes for the joints of robots. Harmonic gearboxes have a high reduction ratio and a small backlash and are compact and lightweight, but their efficiency and stiffness are low. Cycloid gearboxes have a high reduction ratio, small backlash, and high efficiency, but their size is too large. Furthermore, both gearboxes are very difficult to backdrive with a reduction ratio of 1/100. Thus, we focused upon planetary gearboxes and made a 3K compound planetary reduction gearbox, made more efficient by optimizing the numbers of teeth and the profile-shift coefficients. A reduction gearbox with a reduction ratio of 1/97 and an efficiency of 93% was realized and back driving was possible. Building on this, we aim to develop a highly efficient 2K-H compound planetary reduction gearbox to achieve higher efficiency than the 3K model. However, the 2K-H compound planetary reduction gearbox has high efficiency when there is only one planet gear, so the eccentricity of the planet gear must be considered. In this paper, we show the method and experimental result of canceling the eccentricity of the planet gear using a balancer after deriving the efficiency equation in the 2K-H compound planetary reduction gearbox and showing the efficiency result by optimizing the profile-shift coefficient.

Nowadays, the elderly and the disabled are rapidly aging worldwide; therefore, the use of wheelchairs by the elderly and disabled continues to rise. However, many wheelchair users have difficulties adequately adjusting their sitting posture and often suffer from injurious falls, particularly, while traversing hilly roads or slowing down. In this study, a novel robotic seat is proposed to overcome these challenges, and numerical simulations are performed to verify its feasibility. The bottom of the seat is a spherical section, supported and driven by three omnidirectional wheels. An aluminum bar flanked with dumbbells is connected to the wheelchair through a universal joint as the test subject. The motion model is decoupled into two 2-D inverted pendulum models in the vertical plane and one rotation model in the horizontal plane. Two linear state feedback controllers based on the linear quadratic regulator method are used to stabilize the test subject in the vertical plane. The simulation results demonstrate the feasibility of the proposed robotic seat.

For human motion support, there is an active focus on research and development of powered exoskeletons. A powered exoskeleton is directly fitted to the human body; thus, it is necessary to consider the safety of the human-robot interaction, including improving the back-drivability. Toward this problem, a powered exoskeleton that uses a highly back-drivable reduction gearbox even at a high reduction ratio is proposed, and a user's external force can be estimated without using a torque sensor even though the system has very high total gear ratio of 408.57:1. This paper investigated the apparent mechanical impedance of an exoskeleton during assist control execution and examined the performance and operability of the assist control. Finally, a questionnaire of participants was conducted and the mechanical impedance that resulted in a high operability was examined by implementing admittance control.

Most autonomous mobile systems require pre- generated detailed maps that are expensive to prepare. In this study, we proposed an autonomous mobile system that does not rely on a detailed map; instead, it uses a simple map for robot navigation. The robot has two maps; one is a detailed map created by sensor observation during movement, and the other is a simple map provided as pre-information. The robot performs the matching between detailed and simple maps and converts waypoints on the simple map to waypoints on the detailed map. Matching is performed based on straight line matching and is optimized by genetic algorithm. Experiments were conducted in buildings. Our method is compared with linear transformations and conditions that our method works effectively or not are confirmed.

This paper proposes a method of rainfall forecasting with high accuracy by combining two different types of information, cloud image data and humidity numerical data. The cloud image data is processed by using Residual network. By performing rainfall prediction using cloud images acquired from cameras instead of highly specialized weather data, it is easy to use the system for the end users. In addition to the cloud image data, this study introduces the humidity information that can be easily acquired to study, which brings more accurate rainfall forecasting. The combination of cloud images and humidity information realizes short-time rainfall prediction with high accuracy of 95%.

This paper proposes a robotic cane that helps users to stand and maintain their balance while standing and walking. The design is based on an inverted pendulum model, uses a smart omnidirectional wheel, and is linearized by the Lie algebra method (LAM) combined with a nonlinear disturbance observer that estimates the external force applied by the user. The hardware and mathematical model of the robotic cane are analyzed in detail. The hardware design is complete, and considers the weight and battery capacity. The performance of the controller is confirmed in three-dimensional simulations and in physical experiments. The robotic cane tolerates a large fluctuation angle from its equilibrium point. It also has a fall prevention function, which is among the most desirable functions of any walking assistance device, and responds quickly to a human force. The balance maintenance functions are as follows: self-balance, standing and walking assistance, fall prevention, and a mechatronic safety mode. In experimental measurements of the cane in use, the gyroscope sensor insignificant vibrations on the user's bodies, verifying the effectiveness of the robotic cane in daily life activities. The performance of LAM is also compared with that of the linear quadratic regulator.

This work proposes to compare the energetic efficiency of two models of leg: human-like model, with leg bent backward, and bird-like mode, with leg bent forward. The models are represented by a manipulator with 2 degrees of freedom, which simulates the two phases of gait: swing phase (leg swinging) and support phase (foot in the ground, supporting the dislocation of the hip forward). After generating the gait trajectory and simulating the swing and support phases separately in the multibody simulator SimMechanics of the software Matlab, an energetic analysis is performed calculating the energy consumption through the power applied on the joints. In order to focus the performance comparison on the physical dynamics of the models and eliminate inneficiency from the trajectory generation itself, the control solution is generated using a pseudospectral method of optimal control implemented in the software DIDO. The results point to the advantage of the human morphology in bending more freely in lift-off, causing the leg to use better the gravity to swing, and in making use of the increase of potential energy in swing phase to slow down the leg until heel strike.

There is several implementations of techniques for solving the SLAM problem that is a process to expand autonomous mobile robots field. In this context, we will be comparing swarm algorithms for solving the simultaneous localization and mapping problem such as Firefly Algorithm, Particle Swarm Optimization and Glowworm Swarm Optimization, we directed a group of experiments using a wireless system composed of 2D laser range finder UTM-30LX, iRobot Roomba 600 and Raspberry Pi 3 as an experimental environment.

In recent years, most countries have experienced a shift toward aging population, leading to a decline in the working population; thus, an increased physical burden on workers results. The use of powered exoskeletons is one approach to reduce physical stresses on the bodies of workers and elderly people. However, conventional powered exoskeletons have many sensors and actuators mounted on them, resulting in problems associated with their high price, high weight, etc. In this paper, we introduce a new high-efficiency reduction gearbox, which is being researched and developed in our laboratory as a substitute for conventional actuators, and present the results of its attempted assist control without a torque sensor. It was found that the user's torque can be estimated with high accuracy without using a torque sensor by using a highly efficient gearbox with 1/102.14 reduction ratio. In addition, we performed motion identification using a nonlinear support vector machine with the aim of assist control with reduced discomfort on the body by suppressing vibrations.

Applying a nonlinear disturbance observer to a two-wheeled cane for walking assistance is proposed in this paper. The hardware design and controller for the two-wheeled cane are also presented. The two-wheeled cane is designed to help users, especially the elderly walk and balance. The design of the two-wheeled cane is based on an inverted pendulum model, and to validate the performance of system multiple simulation and experimental tests were performed with the designed hardware model. The nonlinear disturbance observer presented good results, which were the same as the force sensor measurement signals. The experimental results show that the two-wheeled cane based on the nonlinear disturbance observer is useful to help elderly persons maintain their balance.

Mobile robots path planning is a central problem in every situation where human intervention is not desired or not possible to accept: full automated industrial warehouses or general stocking areas and every domestic application that involves a mobile robot and special cases where environment is prohibited for human accessing like toxic wastes and bombs defusing [1]. Currently, neural networks are applied to problems related to mobile robot navigation. However, they are not as popular as in applications like image processing, speech recognition or machine translation, where they are commercially relevant. In this paper we propose a Long Short-Term Memory (LSTM) neural network as an online search agent to tackle the problem of mobile robots path planning in unknown environments, meaning that the agent relies only on local map awareness realized with a LRF sensor and relative information between robot and goal position. Specifically, a final structure of LSTM network is analyzed and its performance is compared with the A* algorithm, a widely known method that follows the best-first search approach. Subsequently, an analysis of the method developed on a real robot is described.

Since its beginning around 50 years ago [1], the research on biped robotics developed many successful prototypes, however the conventional strategies of gait generation and control have diverse performance issues. Between the promising alternatives of strategies is Optimal Control, which offers a clear definition of performance criteria and precise methods for achieving it. To highligh the possibilities of Optimal Control, this work solves the minimization of energy expenditure in the swing trajectory of a leg with 2 degrees of freedom, using an optimization method based on Pontryagin's Minimum Principle. The method applied is an indirect method that solves the optimal control of an exact non-linear model. The energy expenditure is compared with the expenditure of an intuitive Inverse Kinematics solution for the same model and task, demonstrating the minimization of energy expenditure. Unlike common optimization methods, this method relies on the use of an exact model of a non-linear system, eliminating modeling errors and generating a precise open loop control solution.

Since the population of elderly gradually increases recently in developed country, and it causes an increase of burdens to care workers, there exists a strong demand for walking assist devices. The falling down of elderly may cause serious injuries which results in significant decrease of quality of life. Conventional wearable walking support devices are not easy for elderly to wear. The authors have developed a wearable device using Wearable Control Moment Gyroscopes (Wearable-CMG) to prevent falling down. In this paper, motion equation of CMGs and steering law for driving multiple CMGs, postural control method for an inverted pendulum model are presented. Especially, some estimation method to control human posture only utilized by gyroscopes are applied, and compare them in preliminary experiments.

There are many methods for 3D simultaneous localization and mapping(SLAM) such like ORB-SLAM, LSD-SLAM and so on when we use camera as sensor. However for laser range finder, there are few algorithms for SLAM, especially for 3DSLAM. Besides, the accuracy and robustness of 3D SLAM is still not enough by using conventional methods for laser range finder. A very famous algorithm for using laser range finder and grid map, Iterative Closest Point (ICP), has been usually used in 3DSLAM. In our previous research, a method of using the Particle Swarm Optimization(PSO) algorithm is proposed to increase the accuracy of 3D-SLAM. And the proposed method that using PSO algorithm in grid map makes a better performance than using the ICP algorithm in our previous research. However, we find a way to combine the ICP algorithm with the PSO algorithm for the 3D-SLAM by using 3D laser range finder and grid map instead of comparing them. By combining these two algorithms, we can reduce the computation consumption and improve the performance. In this paper we show and analyze the result of experiments of SLAM.

For various assistive devices, safe and compact high output actuators are desired. The authors have proposed a helical motor that consists of a helical mover and stator. In the latest study, a motion controller of a helical motor is based on PD-control law and disturbance observer. However, the performance of the conventional controller is not enough to achieve more high preciesion motion and robustness. Therefore, this paper presents a comparison between new controllers based on optimal-control theorem and the conventional controller. Through simulations and experiments, an optimal servo with feedfoward for a helical motor has the best performance compared with the others.

This paper stated a problem of dilemma among accuracy, robustness, and noisiness in an impedance control. Additionally, applying Three-degree-of-freedom (Three-DoF) control for the problem is proposed as a solution. By using Three-DoF control, the performance is improved in terms of the noisiness compared with a conventional framework. The effectiveness is verified by preliminary simulations and experiments. As a result, the proposed method improved the accuracy/robustness and noisiness simultaneously.

This paper presents a basic model of an electro-mechanical energy conversion system using the resonant inductive coupling as well as the power transfer system between the coils. The proposed system utilizes two sets of two-phase air-core coils both for the stator and the mover. There exists parasitic capacitance of the coils which is connected in series. Thus the whole system has certain resonant frequencies. Two-phase voltage source connected to the primary coils is driven at the resonant frequency. Unlike conventional wireless power transfer systems using resonant inductive coupling, the proposed system can generate moving flux field. The proposed system is similar to the induction machine but it is driven at very high resonant frequency and the slip ratio is also quite high. Tendency of the generated force of the mover obtained by the FEA simulation shows that can be explained by the model derived.

In this paper, design and model of new structure of Spiral Motor is proposed. This motor consists of a mover and stator with helical shape. In addition, the stator which has claw shape, where makes winding coils become more easily to manufacture than in convention helical motors. The equivalent magnetic circuit, voltage equations and motion equations are described. Finally, by using Finite Element Analysis, model parameters of the controlled plant are identified.

Recently robots working in various situations as workplace and nursing home autonomously are expected. In those situations, it is important for such robots to recognize surrounding environment and to avoid the risk to people. We propose a method to realize an autonomous robot moving smoothly and safely with low calculation cost by performing motion prediction and human detection in dynamic or static environment. In this paper we use only a single 2D Laser Range Finder (LRF) for reducing total cost. Our proposed method leads to satisfactory results of motion prediction in dynamic environment and human detection in static environment.

This paper proposes a Three-degree-of-freedom (Three-DoF) control that is extended from a Two-degree-of-freedom (Two-DoF) control. In a classical Two-DoF control, the command input response and the disturbance suppression characteristics can be determined independently. However, the disturbance suppression characteristic and the noise suppression characteristic have a trade-off relation. Thus, if the disturbance suppression property sets high, the noise suppression property becomes low, and vice versa. In Three-DoF control scheme, command input response, disturbance suppression, and noise suppression property can be determined independently. To realize above capability, the controlled plant is divided by utilizing its linearity and an additional controller is used. In this paper, the control system architecture and the controller design methodology are shown and its effectiveness are validated by position control simulations of a 1DoF DC motor model.

We propose a bilateral control by design based on complementary sensitivity function. In this control method, we regard force applied by an operator at the master side and contact force from the environment at the slave side as input signals. We are able to determine the transfer function for displacement of the position of the master and slave systems by designing complementary sensitivity function. Thereby this control method is possible to transfer the force at the master side and slave side to the others by feedback only the position information. Adding velocity information in feedback is possible to alleviate the stability condition of the complementary sensitivity function. This paper proposed that the bilateral control can be established in low-order controller.

This paper presents a two-degree-of-freedom (2-DOF) direct drive motor to achieve flexible motion control. As one of the motion control methods, robust position control is significant in order to functionalize industrial robots, machine tools, medical robots, and so on. Furthermore, these requires multi-degree-of freedom motion. Then the authors have proposed 2-DOF direct drive motor, which realizes translational motion and rotary motion. In this paper, voltage and motion equations of the proposed motor are briefly introduced. According to these equations, simulations and experiments are carried out to verify achievement of precision position control in the proposed actuator.

Implementation of verification procedures is required to eliminate design errors that decrease the safety of an automation system. Design errors may vary from case to case but will certainly jeopardize the safety of manufacturing lines and operators. Therefore, checking the possibility of state transitions in the control systems from safe to unsafe states is essential. Formal verification via model checking procedures has proven to be efficient and is widely used. System finite element models are employed to automatically verify certain correctness properties. In this paper, we introduce a method of model checking technique for logic control design. A checking procedure based on Gröbner bases (GB) model is used to analyze and design a controller that meets the requirements defined by a predetermined safety function. We compare our proposed method with symbolic computation tree logic (CTL) model checking based on binary decision diagrams (BDDs). We implemented this technique to a case study by using a crane system.

Since electric power assisted bicycles have been released in Japan, their sales are increasing. However the prices are high because they are equipped with an expensive torque sensor to measure human pedaling torque to assist human by an electric motor. The purpose of this study is to develop an external power assist systems applicable for a normal bicycle that fits in the Japanese regulation. This paper proposes an instantaneous pedaling torque observer based on the disturbance observer and the fourier filtering in order to decouple the load torque and the pedaling torque. The validity of the proposed method is confirmed by numerical simulations and experiments.

A helical motor is an axial gap helical motor with a stator that consists of windings and teeth made of soft magnetic composite. The teeth of a previously developed helical motor were of low precision owing to the manufacturing method. The low-precision teeth affected the motor performance negatively. In this study, we proposed a new type of a helical motor with high-precision teeth; furthermore, we evaluated its magnetic levitation control, zero-power control, position control, and frequency response. We found that less current was required for the magnetic levitation control and the accuracy of the position control improved because of broadening of the stable band.

Cane type walking assist device and comparison of its control methods are proposed in this paper. Our robotic cane has been designed using omni-directional wheel, such that it moves according to user's small force applied to the cane, and to stably support the user when he/she impose large force. The conventional control method has been compared with the proposed one in 2D space. Moreover, the validity of the proposed control method has been confirmed using 3D simulation environment.

This paper proposes a new cross-coupled two-degree-of-freedom (2DOF) direct drive motor which can independently realize rotary and linear motions. In this proposed motor, the mover part is a shaft with segment magnets formed in checkers. On the other hand, two sets of helical windings are cross-coupled in the stator part. The configuration can generate thrust force and torque independently. In this paper, the principle and the validity of the proposed motor are shown by electromagnetic field analysis and fabrication of the parts of an experimental motor is introduced.

Papers dealing with sensors or actuators published in IEEE Transactions on Industrial Electronics (TIE) and IEEE Transactions on Industrial Informatics (TII) during the year of 2011-2013 have been surveyed in this paper. The papers are roughly categorized according to their subjects.

A variety of prosthetic legs are developed and some of them have electric motors controlled by the computer, which have drawback in terms of their weight, power consumption and so on. We are developing a high thrust force linear actuator with helical shaped mover and stator, which help generate high thrust force in spite of the compact size. Magnetic flux of the permanent magnet and windings cause the mover levitate, which can relieve loss of friction. In this paper, the prosthetic knee joint driven by a high thrust helical motor is proposed. To develop the helical motor applicable to the prosthetic knee joint, we designed a very thin lead-length helical motor. FEA simulation shows that continuous rated thrust force is approximately 101N. In addition, 662N of thrust force can be generated if the space factor of windings and the continuous rated current density can be improved. Moreover, the short-time allowable current can be flowed several times of the continuous rated current. This value shows that there is a potential to achieve the walking motion by the prosthetic knee joint driven by the high thrust force helical motor.

This paper investigates two zero power control methods for a spiral motor, which are current integral type and external force feedback type. A spiral motor is a linear motor having helical mover and stator that can realize a high thrust force density and high back-drivability. In latest researches, the magnetic levitation control, position control and force control were archived. Additionally, power-saving magnetic levitation is successful by a zero power control. However, a response time of the force control is delayed by the zero-power controller. This paper presents a comparison between zero power control methods through simulations and experiments. As a result, on the force control, external force feedback type zero power control is able to improve the transition time than the conventional method.

Due to birthrate decline and population aging, researches on walking assist for aged people have been conducted. However, in most of the researches, it is difficult for people with physical disabilities to wear such existing walking assist devices. Therefore, we are developing a device using Control Moment Gyroscope (CMG) to assist their walking. In this paper, we expound postural control of people using CMGs with three control methods and compare results of simulation using these methods.

This paper presents a novel half-circle-shaped tubular linear permanent magnet machine named as “circular shaft motor (CSM)” for direct-drive applications. A CSM is a kind of the tubular synchronous actuator. Both the stator and the mover of a CSM are bent in a circle. A CSM can realize the motion along the circumference of a circle. In this paper, the design of a circular shaft motor is described. The prototype of CSM is introduced. Then, the measurement experimental results of the thrust characteristic are demonstrated. The utility of the developed CSM is confirmed from the results.

In this paper, a new method of center of mass trajectory planning using the zero-phase low pass filter is proposed. This method is based on a table-cart model which simply describes the relationship between center of mass and zero moment point. Generally, zero moment point should be controlled to realize dynamic motion. This method can easily generate the center of mass trajectory which realizes the desired zero moment point. In our study, this method is applied to wheel-legged locomotion. We will show the result that zero moment point can be sufficiently controlled even if quadruped wheel-legged mobile robot is approximated to a table-cart model. The effectiveness of the idea is validated by simulation and experiment.

Spiral motor has both high thrust density and high backdrivability because of its helical structure. Main purpose of this paper is to design and analysis high thrust density SPM spiral motor. In this paper, helical structure has been investigated by using FEA. As the result of FEA, proposed model has almost 6.8 times higher rated thrust density than conventional model.

The exact and general formulation of optimal control for biped robots based on a numerical representation of the motion equation is proposed. We can solve exactly the minimum energy consumption trajectories for a biped running motion. Through the numerical study of a five link planar biped robot, it is found that big peak power and torque is required for the knee joints but its consumption power is small and the main work is done by the hip joints.

Supply chain management (SCM) is widely introduced in the field of manufacturing and logistics systems. The difficulty of the management basically depends on the demand uncertainties. In this paper, a multiperiod inventory problem under the existence of the demand model uncertainties is investigated using field demand data.

The Real-time Colored Petri Net (RTCPN), a modification of traditional Colored Timed Petri net, is proposed to describe virtual factory. In RTCPN, firing discipline differs from traditional way in that the concept of time-enabled transition is introduced to implement real-time simulation. In this paper, we address how to describe virtual factory using RTCPN model. Combined with block diagram, the concept of virtual place is also proposed to model large-scale factory rapidly and conveniently. At last, a Java-built-in implementation of RTCPN tool is developed and a Printed-Circuit-Board factory is given to investigate time cost of developed tool. Concurrent feature of RTCPN is also discussed to decrease simulation time and improved results shows its effectiveness.

Describes a biped walking control system based on the reactive force interaction control at the foothold: 1) robust control of reactive force/torque interaction at the foothold based on Cartesian space motion controller: 2) posture control considering the physical constraints of the reactive force/torque at the foothold by quadratic programming. The proposed approach realizes robust biped locomotion because the environmental interaction is directly controlled. The control is applied to the 20 axes simulation model, and the stable biped locomotion is realized even if an unknown small slope exists. Stable attitude control is confirmed by 14-axis biped robot experiments.

This paper describes a novel biped walking control based on a force control. The proposed control system can stabilize the contact force between the foot and the ground under the proposed restrictions. As a result, a hierarchical control is proposed by which the legged locomotion can be realized. The proposed control system is applied to the 19 axes simulation model and the results of stable walking are shown.

First, the three dimensional precise simulation method for biped robot has been proposed, which is the extension of the open link manipulator simulation method and the contact simulation method of rigid body mechanics. The proposed simulation model can be said to be mathematically exact, thus this approach enables the essential investigation for control algorithms of biped locomotion. This approach has advantages in the less cost compared with the experimental demonstration. Second, the autonomous walking control with hierarchical structure has been proposed in which three dimensional inverted pendulum model is considered as the motion of gravity center, and the tip position of nonsupport leg are controlled autonomously. Both of references can be simultaneously transformed to the joint space reference, then they are accurately realized by the robust-servo controller. Third, the proposed walking control is investigated by the proposed simulation method, and the stable walking motion is confirmed.