In recent years, the number of traffic accidents related to elder drivers is increasing. To solve this issue, advanced safety technologies are in the spotlight. In this research, selective control between assistive and autonomous modes has been proposed using electroencephalogram and support vector machine (SVM). To extract mental concentration, contingent negative variation is utilized. For controllers, assistive mode using impedance control and autonomous mode using quantitative feedback technique were designed and combined. The experimental results showed that the classifier using SVM had good performance to discriminate mental concentration, and was utilized to realize the selective controller.

View full abstract

This paper describes about the problem for the sub-crawler angle restriction of the crawler robot with passive sub-crawlers. Many disaster response robots are the crawler robot with active sub-crawlers. Those robots have high mobility performance to rough terrain. However, it is complicated to control those robots. On the other hand, the crawler robot with passive sub-crawler have easy operability. However, depending on a travelling sequence and kinds of obstacles,we understand that it is necessary to change the limit of the rotary angle of the sub-crawler. This study proposed the dynamical variable angle restriction mechanism for sub-crawlers and considered the problem for the sub-crawler angle restriction of the crawler robot to each sequence by perfoming the traversal experimentation for the stairs.

View full abstract

This paper describes 5-axis centralized feedback control system for a single rotor helicopter. The controller is designed with optimal control theory using Multiple-Input Multiple-Output (MIMO) linear model that includes cross-coupling dynamics, and tuning method for evaluation function is presented. Flight test with this control system demonstrated a steady hover control performance through attitude, position, and main rotor speed. Stability analysis of closed loop system shows a greater stability margin and robustness than Single-Input Single-Output (SISO) controller is used. This approach can be apply to MIMO flight control for various rotorcraft.

View full abstract

A kind of an air cushion vehicle (ACV) is known as a hovercraft which carries out a cruise by blowing off the air from the bottom of the body. The ACV has advantages, such as amphibian vehicle, large payload, and so on. However, the ACV has drawbacks such as its drivability is bad. An improvement of the drivability is effective for the run of the ACV and the disturbance rejection performance of the controller is very significant. In this research, the position of the ACV is estimated using a marker by a turning camera. When the ACV runs on the arc, the ACV runs with a sideslip keeping the given angle against a tangent of arc. It compares with running along the tangential direction. It is confirmed through an experiment that the centripetal force from the part of the thrust by rear fans is effective for the arc running.

View full abstract

Rheumatoid arthritis (RA) patients in Japan are about eight hundred thousand. 80-90% of RA patients develop to the hands and wrist. When developing these, causing the dislocation and deformation of the finger. As treatment for subluxation due to RA, there is a rehabilitation by hand therapy, but therapist is not enough. Therefore a device for rehabilitation of thumb MP joint subluxation has been developed. To improve the device, it is necessary to clear the dynamics of hand therapy. Therefore, the authors tried to measurement four kind of rehabilitation by using motion capture and contact force sensor, and we will try develop the simulation function model for curative effect evaluations of hand therapy for RA. To develop simulation model, we think method of modeling of joint in rehabilitation and estimating the model parameters using measured data.

View full abstract

In human gait motion analysis, which is one useful method for efficient physical rehabilitation, ground reaction forces, kinetic and kinematic parameters are measured. For obtaining these data as the unrestrained gait measurement, a novel gait motion analysis system using mobile force plate and attitude sensor has been developed. However, it has been used at only specially-cleaned experimental laboratory. In this study, ground reaction forces and joint position applied on the lower limb of healthy subjects are measured by two measurement methods. One is a traditional way: a combination of force sensor based on ground and motion capture system. And the other is wearable measurement system composed of mobile force plate and attitude sensor. By compering this two measured results, we can evaluate the accuracy of the wearable measurement system. As a result of the experiments, vertical component of ground reaction force and knee joint angle are measured with a high degree of accuracy. On the other hand, antenior-postenior and medial-lateral components of the force and ankle and hip joint angles showed different results. The data measured by wearable system are mostly larger. Precision placement of attitude sensor is very difficult as the figure of lower limbs is different one by one, and this may be the reason. Future subject is structuring a new algorithm to set off a declination of attitude sensor position.

View full abstract

Magnetorheological elastomers (MREs) are class of smart materials whose rheological or viscoelastic properties can be varied by applying external magnetic fluid. MREs are the solid analogues of MRFs and expected to overcome the disadvantages of the fluid such as the particle deposition and the sealing problems. When such MREs are exposed to external magnetic field, the strength of the inter-particle force will change and thus their apparent elastic modulus is changed depending on the magnitude of the field. In this paper, the stiffness controllable elastomer composites known as MREs are first fabricated and their field- dependent dynamic properties are tested. The MREs are then applied to the vibration isolating mounts along with variable stiffness control scheme. In designing the vibration isolator, the magnetic field analysis was performed by the finite element method (FEM) in order to find an efficient way of applying the magnetic field on MRE. Experimental investigations show that the vibration of 1-dof structure can be reduced by the proposed anti-vibration mount with variable stiffness property.

View full abstract

In this study, an Energy-Regenerative MR damper was constructed and its performance was studied experimentally. Magneto-Rheological (MR) Grease is known as a kind of smart material which can change its rheological property by applying magnetic field. When this Grease is used as working oil of a damper, this damper is called MR Grease damper. This damper is classified into semi-active type of vibration control device. Usually, MR Grease damper has an electric coil. When applying current to the coil, magnetic circuit is formed and the damping property of MR Grease damper can be changed because of the rheological property change of MR Grease. This time, an energy regenerating mechanism which was composed of a ball screw system and a generator was designed and attached to the MR Grease damper. Vibration experiment was performed in three conditions; operate the damper as a passive type of damper, operate the damper as a mechatronic type of damper, and operate the damper as a regenerative type of damper. As a result, the damping force and regenerated voltage were measured. This experiment revealed that this damper could regenerate up to 30% of vibration energy.

View full abstract

This paper describes a successful application of Magnetorheological (MR) grease to a shear type controllable damper. MR fluid is known as a kind of functional fluids which changes rheological property when an external magnetic field is applied. However, there is a problem with the sedimentation of the dispersed particles. To prevent the sedimentation, MR grease was developed. In this study, we developed a shear type controllable damper using MR grease. The performance test demonstrated that the proposed damper has wide dynamic range of damping force despite its simple structure. The ratio of maximum damping force in on-state to that in off-state is approximately 65. In addition, the proposed damper was installed between first floor and second floor of a one-story model structure. The vibration control test demonstrated that the vibration response was successfully reduced by using the simplified skyhook algorithm. Our results indicate that the MR grease damper effectively works for the semi-active vibration control system.

View full abstract

In this paper, an elastomer composite with controllable stiffness, known as a magnetorheological elastomer (MRE), is used in a dynamic vibration absorber whose natural frequency is tuned adaptively to the disturbance frequency through the application of an external magnetic field. The field-dependent property test of the fabricated MRE sample shows that the stiffness changes by more than six times compared to the baseline property value at a 40% iron powder volume concentration. The MRE is then used to fabricate a frequency-tunable dynamic absorber for mitigating transient vibrations of a one degree-of-freedom system. Investigations show that the proposed absorber outperforms a conventional passive-type absorber throughout the tunable frequency range.

View full abstract

In this paper, we propose a gray-box modeling method that incorporates physical characteristics of a dynamical system. The characteristics of the system are expressed in terms of the time-moment of the transfer function. We obtain the moment information from the pre-identification experiments. Explicitly using the moment information as equality constraints, we propose a constrained least squares method. We can construct a transfer function model that includes the desired moment information at a specified frequency. Moreover, we extend the proposed identification method to an uncertain moment-constrained case relaxing the equality constraint to inequality constraints.

View full abstract

In this report, we consider a system identification input design for Multiple-Input Single-Output (MISO) systems. First of all, we derive an identifiability condition of MISO system. Then, we introduce candidates of the input signals which meet the identifiability condition. Finally, we evaluate candidates of the input signals through numerical simulations.

View full abstract

This paper presents a new identification method to obtain a continuous-time model of the mechanical systems with nonlinear friction. The method utilizes projection of measured signals onto a finite dimensional signal subspace. Its effectiveness is demonstrated through numerical example for a linear plant with nonlinear friction.

View full abstract

In these days, motion capture cameras can achieve localization of moving objects with high accuracy. However, it has some problems, for example, it may fail to measurement due to occlusion and the quantization error. In addition, the measurement data with noise or lack of data disturbs the control performance of visual feedback systems. Thus, it is necessary to estimate states suppressing the influence of the noise and occlusion. In this study, we achieve high accuracy of state estimation using Moving Horizon Estimation (MHE) which is able to consider measurement data during the finite-time. We implement MHE with the actual experimental system of 3D static cameras and show the effectiveness for visual feedback control of MHE.

View full abstract

An automatic control system for a small jet engine is designed in the study. Some PI controllers are scheduled according to a control mode in the sequence of the engine control. By tuning the control gains in the PI controllers and the timing of the gain scheduling with the experimental data, the performance of the engine control system is improved

View full abstract

This paper analyzes the stability of pleasure boats controlled by human in order to develop an electronic control steering system of those boats. At first, the equation of motion for pleasure boats in the global coordinate is derived. This equation of motion is used to compute a man-machine system model for controlled pleasure boats. As the first case, the stability of pleasure boats controlled by human proportional action is analyzed based on the man-machine system model. The stability region can be reduced when the distance to forward gazing point is long, and can be enlarged when the velocity of pleasure boat is high. As the second case, the stability of pleasure boats controlled by proportional action added derivative action is analyzed. Compared with the first case, the stability region expands when the distance to forward gazing point is short. As the third case, the stability of pleasure boats controlled by proportional action added time delay is analyzed. Compared with the first case, the stability region expands when the distance to forward gazing point is long. As the fourth case, the stability of pleasure boats controlled by proportional action added integral action is analyzed. Compared with the first case, the stability region expands when the distance to forward gazing point is short. From those analysis, it can be seen that the time delay can be expected to improvement the stability of pleasure boats since it is appropriate for implementation of the control algorithm.

View full abstract

We develop a preprocessing method to improve performance of the intelligent controller that solves competitive problems formulated in a coupled inverted pendula model. The controller outputs impulsive forces to produce desired final states based on a Look-up table (LUT) that stores dynamical correspondence from the initial to the final states. It is numerically clarified that the performance decrement can be caused by the LUT misclassification that occurs near the boundary points of the reachable sets of the model. The proposed algorism removes the boundary points from the LUT. The resulting controller actually reduces the misclassifications and improves the control performance.

View full abstract

This paper considers formation control of multi-agent systems based on leader-follower consensus. The developed multi-agent systems consist of multi-robots with four DC motors and an I/O port using Arduino. The effectiveness of the formation control technique based on leader-follower consensus is verified by using the developed multi-robots in a numerical example and an expetiment.

View full abstract

This paper proposes the Fictitious Reference Iterative Tuning-Particle Swarm Optimization (PSO-FRIT) method to design PID controllers for feedback control systems. The proposed method is an offline PID parameter tuning method. Moreover it is not necessary to derive any mathematical models of objected control systems. The proposed method is demonstrated by comparing with the FRIT method in numerical examples and an experiment.

View full abstract

Recently, many mechanical systems become complicated and various, so the system's control process and design become hard tasks. Therefore, the method controlling and designing these systems easily is needed. We focus on modularization. Most of machinery consists of modularized elements. Therefore, we can express it by using modularized common factors. In previous research, to control and plan complex machines easily, we make one modularized element's dynamics models and corresponding machines and link these models manually to make the whole model of system composed of these machines. In this paper, to adapt this system to real epic machines, we propose the method to recognize real modularized machine's model and control it based on it.

View full abstract

In this paper, we propose an optimization method of joint angles for Leg/wheel robot equipped with 3dof SCARA leg mechanism. The joint angle limit by the leg mechanism is considered in constraints of optimization and the velocity vector field is considered by optimization. The joint angle nearest to the wheel is calculated using other joint angles so that we can remove this joint angle from optimal variables. We optimize an index function using trigonometric function, however, this index function is nonlinear. So the optimal solution is calculated by sequential quadratic programming. To deal with the non-positive definite Hessian matrix, we introduce BFGS algorithm. The efficacy of this method is verified via the numerical simulations.

View full abstract

Omni-directional vehicles are good at narrow space because these can move toward arbitrary direction. Especially, these robots that steer and drive independently can work on uneven ground better than the robot using special wheels which is omni-wheel and mecanum wheel. In this study, we use the coaxial steering vehicle. This robot is equipped with four coaxial steering mechanisms. There are two steering actuators on one axis in the coaxial steering mechanisms. The two steering mechanisms have different range of movement and maximum angular velocity. In the previous study, the coaxial steering mechanism was controlled by model predictive control which was subject to these features as the constraints. However, tracking performance was not verified in the previous study. In this paper, we verified availability of the model predictive steering control for the coaxial steering vehicle by tracking control in numerical simulation.

View full abstract

Currently, many control studies of the powered suit are done by using a myoelectric sensor. By reading the electrical signals which are sent from the brain to the muscles, it is easy to synchronize the control system to the motion of the body. However, the myoelectric control is forced to change the parameters in accordance with the user. In this study, the acceleration sensor is used to perform in acceleration control of the 1-link-robot-arm. The acceleration is a factor of the force. Therefore, it is possible to be treated as the force control and simplify control system. Besides, the acceleration control can estimate the load and realize the trajectory tracking control that is difficult by the myoelectric control. In this study, we examine the trajectory tracking control and the estimation method of the load.

View full abstract

An automatic spinach harvester that can control the position and the angle of the root cutting blade has been developing. Position control of the blade is performed by a mechanism to control the arm length, the angle control is performed by the mechanism to control the angle of the blade. In harvesting experiments, the control command which was provided by trial and error is adopted. On the other hand, improvement of performance is expected by setting the control target value of each mechanism based on the desired orbit of the root cutting blade. This paper propose a method for obtaining the control target value of the cutting blade by using a geometric model of the mechanism and by solving an inverse problem.

View full abstract