Fumitake Fujii

Analysis on the manual control characteristics of the human pilot during the operation of the powered wheelchair

The electrically powered wheelchair (EPW) is now recognized to be an indispensable tool for people with motor disabilities. However, using EPW in daily life might be a cause of danger because of the poor maneuverability of some commercially available EPW products. In order to ease the manipulating sensitivity and produce a pleasant and secure EPW, a model representing the operators control ability while using the EPW is necessary when designing an embedded controller. This paper proposes a model which describes the control characteristics of the human operator while she/he is using an EPW. The proposed model is constructed based on, and coincides with, the pilot model which represents the control capability of the human operator while doing her/his accustomed tasks. The model should be called the error model of the EPW, control because it contains two errors to be stabilized for pleasant and secure operation of the EPW. In addition, the proposed model is an enhanced version of the pilot model in the sense that the effects of the environmental changes acting on the pilot are taken into account. Experimental results show the validity of the error model and indicate that an inappropriately designed EPW controller strictly deteriorates the manual control performance. The results obtained will serve as a good basis for the design of the EPW controller.

A Simple Phenomenological Modeling of McKibben Pneumatic Actuator and Its Application to Control

The focus of the paper is to develop a simple phenomenological model of McKibben pneumatic actuator which is capable of capturing its inherit nonlinear and varying dynamical characteristics under different load conditions. It will be shown that physical work analysis determines the model structure and phenomenological nonlinear spring term is introduced into the model to capture the response characteristics. A simple control system for the actuator using proposed model structure to construct exogenous load observer is also presented. Load carry experiment using single arm exoskeleton unit has been conducted to validate the performance of the proposed control system.

An adaptive internal model control system of a piezo-ceramic actuator with two RBF neural networks

This paper presents a neural network based positioning control system of a piezo-ceramic actuator which exhibits hysteretic behavior. Proposed control system utilizes two neural networks with radial basis function (RBF) as their activation functions: one is used for modeling hysteretic behavior of the actuator and the other is assigned the role of a feedback controller for hysteresis compensation and tracking. The particle swarm optimization algorithm has been applied to the training of RBF-NN for modeling PZT dynamics to achieve high precision, whereas back propagation has been used for online controller parameters update. An internal model control (IMC) structure is employed which combines aforementioned two neural networks for positioning control of the actuator. Results of the positioning control simulation of PZT will be shown to indicate the validity of the proposed two RBF-NN internal model control system.

A simple and robust binaural sound source localization system using interaural time difference as a cue

This paper handles development of binaural robotic audition system which is capable of localizing sound source in the horizontal plane and proposes two simple and robust binaural sound localization algorithms using interaural time difference (ITD) as a cue. Presented algorithms for determination of ITD from recorded binaural signal are both based on interaural phase difference (IPD) spectrum which is easily obtained by performing Fourier analysis on recorded binaural signal. They are also fairly simple in nature, require no a-priori information like measurements of HRTF and/or physical parameters of the room, and the computational costs necessary for ITD identification are low. Results reported in this paper indicate that robust binaural sound source localization under various recording conditions is feasible with the proposed simple algorithms.

Use of angular acceleration and velocity estimates for control of pneumatic-powered wearable motion assist device

The focus of this paper is to develop a control system for wearable motion assist device giving physical support for people who are doing physically demanding tasks. Two key technical issues in developing such a device are how to generate the reference trajectory for the actuator installed to provide appropriate motion support, and how to determine the strength and duration of assist to help complete current physical task. This paper tackles the use of joint angular acceleration and velocity estimates in the control of motion assist device. How the estimates are generated, how they are used in control and how the control system is organized to achieve performance requirements will be described. Several experiments have been conducted to test basic functionality of the developed 5 d.o.f. wearable assist device as well as to gain insights on how to tune physically motivated control system parameters.

Construction of a respiratory-induced lung tumor motion model using phase oscillator

This paper discloses the development of a phase oscillator model for radiotherapy which predicts future position of the lung tumor based on the current and past phase of lung tumor, respiration waveform and heartbeat measurements. We assumed that these phases can be measured at the same time, and tried the prediction of 0.15s ahead of time of the superior-inferior direction of a lung tumor motion. The result of offline simulation shows that the proposed model provides the estimated position of the tumor with 1.5mm RMS error.

Bouc-Wen hysteresis modeling of a piezo-ceramic actuator with membrane structure genetic algorithm

Piezo-ceramic actuators are widely used in a variety of engineering applications which require high precision positioning. However, it exhibits hysteresis nonlinearity which deteriorates positioning accuracy if no proper compensation has been given. Modeling the dynamics of piezo actuator is a prime choice for hysteresis compensation. This paper proposes the membrane structure genetic algorithm (MSGA) method to identify the parameters of Bouc-Wen model to capture the dynamic behavior of a piezo-ceramic actuator which exhibits hysteresis. Results of numerical simulations have been disclosed to illustrate the performance enhancement of MSGA over classical GA when they are applied to the parameter fitting problem of a Bouc-Wen model for experimentally acquired datasets.

Attempts on the Performance Improvement of the Manual Auditory Feedback Control Systems

It has been pointed out in several foregoing literatures that the well known properties of the human pilot observed in the manual visual feedback control system can also be observed in the manual auditory feedback control system. However, performance of the manual auditory feedback control usually tends to degrade as compared to the visual feedback case. In this paper, some experimental attempts to improve the maneuverability of the manual auditory feedback control system are presented. Several differently configured auditory displays and the use of phase lead compensator are proposed and evaluated using the artificial manual auditory feedback tracking control system. Experimental result indicates that performance improvement can be achieved to some extent; however, the effectiveness of the applied auditory devices and/or lead compensator largely depend on the plant characteristics