The motion control mechanism of the body is a major subject of neuroscience, and investigations of it are expected to contribute to an improved understanding of our bodies. An optimal feedback control (OFC) model is one of the newest control models introduced to neuroscience. In this paper, the motion control mechanism of reaching arm movements was reviewed in the framework of OFC referring to recent works carried out at the author's laboratory. The movement of the arm was simulated using an iterative linear-quadratic-Gaussian method to realize an approximate optimal controller. The reactive response to a perturbing force during reaching movements of the arm was measured experimentally, and compared with the simulated results using OFC. OFC was able to reproduce some aspects of the reaching movements of the arm when a cost function was appropriately defined. This suggests that OFC is a plausible base model to describe the internal state of the body, and a further improvement of the model is expected to contribute to the development of efficient training methods for improving quality of movement, both for healthy and disabled persons.

View full abstract

In many motion control applications, it is essential to handle a delayed measurement to estimate the unknown states correctly. However, a literature review shows that difficulties still remain if the delay time is quite large in comparison with the period of the control input. For instance, some methods need to increase the size of the original system considerably. This paper describes a novel filter for a linear system with a large-time-delay measurement. By utilizing the norm-bound scheme based on a matrix inequality, the upper-bound of the estimation error covariance is obtained, and the filter gain can be computed easily. The proposed algorithm is almost similar to the standard Kalman filter without any increase in the dimension of the system. The algorithm is developed for two cases: Case 1: considering only a delayed measurement and Case 2: considering the fusion of delayed and nondelayed measurements. Two motion control applications are performed to verify the algorithm: sideslip angle estimation for vehicle motion control and target position estimation for visual servo control.

View full abstract

For an assist device, a safe and compact high output actuator is desired. The authors have proposed a helical motor that consists of a helical mover and stator. In the gap control of this motor, zero power control is applied to save power for magnetic levitation. However, when a disturbance force is imposed upon the motor, the gap response with the zero power controller has an inverse response. To solve this problem, a new zero power control based on external force feedback is proposed, which is used to generate both the gap displacement and gap velocity references. Experimental results and a mathematical analysis show that the proposed method suppresses the inverse response well in position control and force control.

View full abstract

This paper describes stability analysis and a stabilizing control design for a system with time delay. A delay often occurs when the transmission path includes a computer network. For a time delay, the stability analysis in the frequency domain approach is not sufficient; thus, stability analysis in the time domain is required and typical method is based on the Lyapunov-Krasovskii functional (LKF). This paper describes stability analysis and the design of stabilizing controllers based on various types of LKFs. The paper also describes a model-based control for a system with time delay and it resolves the issue on the communication disturbance observer (CDOB), which generates steady state error if there exists model error. Results of experiments conducted on a DC motor are shown and they confirmed the validity of the proposed model based control method.

View full abstract

The paper proposes a Marx topology DC-DC boost converter (MTBC) with a high boot ratio, where a parallel connection is applied on the input side in order to reduce the conduction and copper losses while a series connection is applied on the output side in order to reduce the voltage stress on switching devices. With the proposed circuit configuration, the high boost ratio DC-DC converter achieves high efficiency. A three-stage MTBC with a boost ratio of 8.33 between the input and output voltages was designed and constructed. A maximum efficiency of 94.5% was achieved with the designed three-stage MTBC. The loss analysis showed that the iron and conduction losses were dominant. The volume analysis showed that the proposed three-stage MTBC requires approximately 23% less of a total volume compared to the conventional isolated DC-DC converter with a two-series-output-rectifier (ICSR) while reducing the power loss of semiconductor devices.

View full abstract

In this paper, we present a buck-boost-full-bridge (BBFB) inverter that is a novel circuit topology developed for all-metals induction heating cookers. A power device of a full bridge inverter serves as a part of the H-bridge PAM (Pulse Amplitude Modulation) converter, thus reducing the number of power devices. In iron pan heating, it consists of a buck converter stage and a full bridge inverter stage with a boost function. On the other hand, in aluminum pan heating, it completely becomes independent of the H-bridge PAM converter stage and a single-ended push-pull (SEPP) inverter stage. For iron pan heating, the power conversion efficiency showed an approximately flat characteristic in a wide power range.

View full abstract

With the increase in the aging population of developed countries, the demand for walking assisting devices has grown. Walkers are widely used because they provide high mobility and safety. In this paper, an electric wheeled walker (EWW) is used to prevent the user from falling down. This paper presents novel fall prevention systems that are composed of a human motion detecting algorithm and a braking system. The motion detecting algorithm focuses on the reaction torque of the EWW and the acceleration of the user. The reaction torque is estimated by the reaction torque observer (RTOB), and the acceleration is measured by using one IMU sensor. The braking system works by switching commands on the basis of human motion. Four motions are considered, and if the user motion is detected as dangerous, the command value switches, and braking is initiated in response to the user posture. Experiments are performed to verify the effectiveness of the proposed algorithm and controller.

View full abstract

Magneto hydro dynamics (MHD) phenomena exhibit complex behavior. In this paper, a new meshless analysis method employing the magnetic moment method and the particle method is proposed. The particle method is an analysis technique that can be used to calculate fluid motion by the dividing fluid into small particles. The magnetic moment method is an analysis technique in which the magnetic field is calculated by dividing ferromagnetic materials into small elements. In this study, the fluid particles in the particle method are treated as ferromagnetic material elements in the magnetic moment method. Therefore, the proposed coupling method does not require calculation points in the air space, i.e. remeshing of the entire space. In this paper, the proposed method is applied to analyze the behavior of a magnetic fluid under a magnetic field.

View full abstract

This paper discusses the suppression of torque ripples in motor drives. Current ripples caused by pulse-width modulation with frequencies of 10-100kHz can be reduced by using high-frequency switching. On the other hand, current ripples caused by dead-times with 6kth-order ripples of fundamental frequencies increase with the switching frequency. Therefore, there is a trade-off relationship between the two ripples. To compensate for dead-times, feedforward compensation is widely used. However, effective compensation cannot be achieved if the feedforward compensation values contain modeling errors. This paper therefore reveals that the robustness of the output torque against compensation errors can be improved by using multilevel inverters. The proposed method is validated through simulations and experimental results.

View full abstract