This paper covers a realization of the half sampling time deadbeat control applied for a two-phase interleaved DC-DC buck converter. The rising, falling and settling times for high and low pulses voltage output become shorter than those of one sampling time deadbeat control. As a result, the performance of the output response can be drastically improved, while the stability of the system remains.

This study aims to develop a tactile display which can represent 2D haptic information. In our previous works, sensorless position control for a compact dual solenoid actuator (CDSA) has been proposed to realize a compact actuation system for the tactile display. To represent the haptic information, contact force control is also required. This paper proposes position/force sensorless force control using the position estimation and a reaction force observer (RFOB). The RFOB is constructed based on the motion equation of the CDSA. Experiments were conducted to validate the performance of the force control. The results show that the contact force is accurately controlled with about 5 mN error which is 6 % of the force command offset. The gain of the proposed force control becomes more than -2.6 dB when the command frequency was less than 20 Hz. The control phase delay was 14.0 ms which came from the cut-off frequency of the low pass filter in the RFOB. The experimental results with change of the mover position showed that the contact force was accurately controlled when the position was at 0.0 ~ 0.2 [mm]. This proposal is useful because a compact actuation system whose position and force are controllable without position and force sensors is realized.

This study aims to realize a tactile display which can represent 2D haptic information on the display surface. The display consists of many compact actuation systems whose position and force can be controlled. In our previous works, sensorless motion control for a compact solenoid actuator has been proposed to realize the compact actuation system. Additionally, we have proposed sensorless control method for a dual solenoid actuator (DSA) to improve the control performance. However, the proposed DSA system has a few problems such as large body, large mover mass, and large friction force. In order to solve these problems, this paper proposes a new compact dual solenoid actuator (CDSA). The proposed actuator has less number of coil turns, therefore much current is required to generate large magnetic force. On the other hand, the large current generates large thermal energy which causes change of the inductance characteristics of the dual coils. The sensorless position estimation is conducted using the relation between the mover position and the inductance. Therefore, thermal compensation method for the new CDSA is also proposed in this paper. Experiments were conducted to validate the proposal. As a result, the sensorless position estimation with less than 60μm estimation error was achieved.

This paper proposes intermittent voltage control of two battery HEECS chopper, which has a very high efficiency such as 99.3% and suitable for power train of EV. However, due to the nature of this chopper topology one of the demerits is that the output voltage control has a little problem when the output voltage changes from the duty ratio zero to a minimum duty ratio, and also the high output voltage changes from maximum duty ratio to unity. This is a result of the finite dead time delay of the chopper switching devices. To solve this problem, intermittent pulse density modulation is proposed, and the pulse is intermittent in the very low duty ratio and also in the very high duty ratio ranges.

3.3 kV-100 kW, 99.5% SiC chopper using gate circuit insulated by optical fiber was designed for renewable energy generation system. 8 kW Dual Active Bridge (DAB) converter was fabricated and tested. 2 phase HEECS chopper of 3.3 kV-100 kW-20 kHz which loaded it was completed, and performed a BTB (Back to Back) experiment. 99.71% (target efficiency:99.5%) was obtained.

The paper proposes a solution of State of Charge (SOC) unbalancing problem, which occurs to the two battery High-Efficiency Energy Conversion System (HEECS) chopper used for a power train of Electric Vehicle. This two battery HEECS de/de chopper shows a very high efficiency over 99%, based on the principle of a partial power conversion [1]. However, if SOC's of two batteries in this chopper do not become zero at the same time, the chopper may not work well and as a result, the vehicle can no longer run. As a solution of this problem, the soft SOC balancing method for two batteries on HEECS is proposed, in which any hard ware is required, and by the control algorithm the SOC's will be balanced. Simulation results showed that SOC's unbalancing can be solved by changing de/de chopper duty ratio from PWM to fixed duty depending on SOC's of two batteries. Based on this proposal, the driving distance can be extended 7.6 % under 10-15 driving mode test cycles. The validity of the proposal soft balancing technique was verified under another three types of driving mode cycles such as JCOS, Worldwide harmonized Light vehicles Test Cycle (WLTC) mode Low speed phase, and WLTC mode Medium speed phase.

In-circuit tests play a key role in manufacturing electronic devices with a high level of reliability. One of the most flexible technology for the tests is based on flying probes, in which a set of conductive needles is automatically positioned on multiple test points of the device under test. The probes are controlled at high speed within the working space, to increase the production rate, Position and contact force of probing system should be accurately controlled, to prevent damages to the parts. A force sensor is usually used in the contact detection and for controlling the contact force. However, the force sensor is weak against impact force and expensive. A Reaction Force Observer (RFOB) is often used in force control system without force sensors. This paper proposes a fast force controller for probing systems without force sensor. In this study, a MEMS accelerometer is used to improve the performance of the RFOB. Acceleration-aided Kalman Filter (aaKF) is combined with the RFOB to derive accurate estimates of the velocity and acceleration signals. The contact detection is conducted using jerk signal, which can be obtained by differentiating the acceleration. After the contact detection, a negative constant force is applied for a short period, to decelerate the motion of the probe. Finally, a force control is applied. Experiments were conducted to validate the effectiveness of the proposal. As a result, the proposed method can quickly control the contact force without overshoot, with a settling time of 10 ms. The proposed control is effective for circuit probing systems, because the force is controlled in short time, so making the in-circuit test time short.

A new high efficient energy conversion system is proposed for the automotive application and the renewable energy generation. It provides a partial boost switching technique which realizes extremely high efficiency chopper for battery voltage fluctuation. First prototype unit of dual active bridge converter using silicon carbide metal-oxide-semiconductor field-effect transistor was fabricated and evaluated by its system test. The experimental results show the verification of highest of over 99.19 % at 8.1 kW under the frequency of 20 kHz.

Conventional methods for storing and reproducing task are based on only the information of the actuators in the robot. However, in these methods, reproducing the stored task is difficult when the environment differs from that in storing phase. This is because these methods consider a meaningful task as just a motion. In this paper, to reproduce a stored task even if the environment differs from that in storing phase, storage and reproduction of a task based on the controlled variable of a task is proposed. To verify the principle of this proposal, the water pouring task is considered. The proposed method can reproduce the water pouring task even if the initial weight of the water in the bottle differs from that in storing phase. The validity of the proposal is confirmed by experiments using one degree-of-freedom robots.

Screws are widely used in various types of manufactured product. For screw-tightening, the pushing force and rotary torque are required. However, usually only the rotary torque is controlled in order to control the screw-tightening process and the pushing direction is important. If the pushing force becomes small, the driver-head separates from a screw. On the other hand, if the pushing force becomes large, the friction also becomes large and increase of the pushing force is needed. Therefore, the pushing force has relationship to the rotary torque. This paper focuses on the pushing direction and the aim of the research is the fast screw-tightening. This paper shows the modified screw-tightening model. Moreover, the pushing positon is controlled based on the rotary angle to reduce the reaction force for the fast screw-tightening. The availability of proposed method is verified by the simulation and experiment.

This paper discusses visual servo system for ball dribbling by using bipedal robot "Nao". Countless papers have been written by researchers since the establishment of RoboCup in 1993, which contributes to the progress in robotics and artificial intelligent. However, only few papers discuss dribbling even though ball dribbling is one of major techniques to master to win a game. To achieve ball dribbling, rigid visual servo is needed in order to kick the ball continuously. Because of the noise caused by the oscillation of walking, it is difficult to make a visual servo system that can achieve ball dribbling. In this paper, visual servo system which is able to control both stepping control and walking pattern is proposed. The proposed visual servo system can offer more rigid visual servo with stability than the one that commonly used by applying foot step planner, walking pattern generator and image processing.

Recently, devices for manufacturing semiconductor products require plasma. Plasma is generated by two types of high frequency power, a High/Low and an ON/OFF output. A power supply, which can output the high frequency power is required to be more efficient and more compact. In order to achieve high frequency output with the High/Low pulse, the output capacitance in a power supply should be reduced. From this reduction, voltage vibration and voltage overshoot are increased. These problems can be solved by a suitable control, which keeps the output voltage stable. This paper proposes a deadbeat control which uses a reference current and replacement parameter of a load variation. The proposed method is capable of controlling output voltage without taking into account voltage vibration. The theory of the proposed method is evaluated by simulations. The availability of that is verified through experiments. The results obtained show that the proposed deadbeat control reduced the vibration range of the output voltage by half. In addition, the voltage overshoot is almost suppressed. The proposed deadbeat control realizes stable output of the ON/OFF pulse and enables high frequency output of the High/Low pulse.

This paper proposes a system and its control method to boost part of the battery voltage for electric vehicle. This system consists of the powertrain of electric vehicles and a bi-directional isolated DC-DC converter called dual active bridge (DAB). The powertrain of electric vehicles in the market has the problem that its performance degrades due to fluctuations in battery voltage. In order to solve this problem, a DC-DC converter is often inserted serially between the battery and the inverter. However, in this solution, the conduction loss of the DC-DC converter always occurs. In this paper, as a solution to the above problems, a voltage boosting system using DAB converter is shown.

This paper proposes driving range extension by series chopper power train for Electric Vehicles (EV for abbreviation). The proposed method was experimentally verified and compared with the previous publication [1]. Further measurement on motor test bench, to compare this drive with the conventional power train, showed a 2.1 % improvement.

This paper aims to present a technique to generate feasible and dynamically equilibrated ZMP-based center of mass trajectories that can be applied to bipedal robots. In this regard, we utilize the ZMP concept in the spherical coordinate frame so that we can fully exploit its properties as this strategy enables us to efficiently combine intrinsic angular momentum rate change terms with inertial force terms. That being said, this strategy has certain advantages: 1) In the case of bipedal walking, undesired torso angle fluctuations are more restrainable compared with other methods, in which angular momentum information is omitted or zero-referenced. 2) Composite rigid body inertia, which is a multibody property of robot dynamics, can be characterized during the trajectory generation task. Thus, relatively more dynamically consistent trajectories may be obtained. 3) The motion interference between the sagittal and lateral planes is naturally included. In this paper, we mainly investigate the first two advantages. Applying the method described above, we conducted bipedal walking experiments on our actual bipedal robot MARI-3. As the result, we obtained repetitive, continuous, and dynamically equilibrated walking cycles, in which undesired torso angles were well suppressed. Furthermore, ZMP error tends to decrease since inertial parameters of the robot are characterized. In conclusion, the method is validated to be efficient in inducing less ZMP error and in suppressing undesired torso angle variations, compared with both flywheel-superimposed and conventional ZMP-based trajectory generation methods.

This paper is aimed at implementing Eulerian ZMP Resolution method to bipedal walking pattern generation. The main strategy in this method is to ensure the dynamic balance by generating feasible ZMP-based CoM trajectories. For this purpose, we employ ZMP equations in spherical coordinates, so that the intrinsic angular momentum rate change about center of mass is included explicitly in a natural way. This fact results in two merits: 1) Undesired torso angle fluctuation and body twists are expected to be more restrainable comparing to other methods in which intrinsic angular momentum information is ignored or zero-referenced. 2) The interference between motions in sagittal and lateral planes can be extracted. In this article, we mainly investigate the first merit and briefly discuss about the second merit. Applying the aforementioned technique, Eulerian ZMP Resolution, we simulated bipedal walking on a 3-D dynamic simulator. Secondarily, we conducted bipedal walking experiments on the actual bipedal robot. In conclusion, we obtained dynamically equilibrated bipedal walking cycles, which satisfactorily verify the efficiency of Eulerian ZMP Resolution technique over conventional methods.

This paper is presenting a method to generate real-time running and jumping trajectories that can be applied to bipedal humanoid robots. The proposed method is based on maintaining the overall dynamic balance by using the ZMP stability criterion throughout support phases. To be able to reach this goal, we utilize ZMP equations in spherical coordinates, so that the rate change of angular momentum terms in ZMP equations are included naturally by using Eulerian equations of motion for unsymmetrical bodies. Thus, undesired torso angle fluctuation is expected to be more restrainable comparing to other methods in which angular momentum information is ignored or zero-referenced. Applying the aforementioned technique, we primarily simulated running motion on a dynamic 3-D simulator. Secondarily, one-legged jumping experiments were conducted on the actual bipedal robot. In conclusion, we obtained repetitive and successful running and jumping cycles which satisfactorily verify the proposed method.

In this paper, a new considering approach for frictional constraint is proposed. It is pointed out that the conventional method only considering translational friction is not appropriate for biped walking, and the frictional rotation constraint should be taken into account. It is stressed that the translational motion and rotational motion will occur simultaneously when slip starts in biped walking, and the frictional rotation constraint plays a more important role than the frictional translation constraint. Furthermore, a method to calculate the slipping frictional force and torque, to predict the most possible slipping direction of a support foot in single support phase is developed. With this method, the quantitative compensation for the yaw-axis moment gets possible. Finally, a new frictional constraint rule applicable for any frictional motions is suggested

In this paper, a new design approach of bipedal walking pattern based on the synchronization of the motions in sagittal and lateral planes are presented and two walking patterns of ZMP fixed and ZMP variable cases are developed. Based on our previous work, bipedal walking is separated into the initial acceleration, double support, deceleration, and acceleration phases; consequently, the nature that bipedal walking is in fact a continuous acceleration and deceleration motion is revealed. With the discusses on the motions both in the sagittal and lateral planes, the fact that the motions in these two planes are tightly coupled together is clarified. The motion parameters such as the walking velocity, walking time, and phase stride can be easily changed simply by altering the swinging amplitude in lateral plane that is determined by the double support phase. The constraint conditions of the phase stride, velocity and swinging amplitude are investigated. Therefore, an approach for adjusting walking velocity by controlling the swinging amplitude is naturally developed. The motion planning is also presented and a numerical example is given out.

Walking principle of biped robot is clarified with ZMP (zero moment point) concept, friction constraint, and inverted pendulum model in this paper. The stable walking condition is derived out with ZMP constraint, friction constraint, and inverted pendulum model. With this result, the nature that the biped walking is in fact a continually acceleration and deceleration motion is presented and its walking velocity can be adjusted by controlling landing point. The desired ZMP trajectory is given out based on derived stable walking condition, and the motion parameters of biped robot, such as walking stride, period, and their restriction are investigated. The approach used in this paper is expected to extend to analyze the motion of running and jump robot.

This paper mainly discusses the characteristics of free sliding motion in releasing manipulation. Releasing manipulation is such a manipulation, in which an object is given initial velocities by striking and slides on a surface to stop at the pre-assigned destination under the action of friction. After the properties of releasing manipulation are given out based on our previous work, the trajectory, the relationship of the sliding motion and the object shape, and the motion monotonicity of any one object with arbitrary shape are intensively discussed. Several important results are originally obtained. Finally, a simple and effectual approach of determining the necessary initial velocity of the object is presented with the above obtained results.

We analyse the mechanisms of human walking patterns and propose two design methods of neural oscillatory networks for generating various desired biped walking patterns of an 8 joints model in 3D plane. First, we discuss how the construction and the parameters of the network determine the inner state variable of each neuron in the network. Then we try to use two different methods to build the suitable network which can generate the desired walking patterns. The first method is to lower the system dimension to 3/spl times/n by considering some mechanisms in human walking and solve the connection weights of the network by the harmonic balance method. The second method is to calculate all kinds of connections of the network by genetic algorithms. From the results of simulations the methods were proved to be effective.

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.

In this paper a method to solve the trajectory generation problem in redundant degree of freedom manipulator has been proposed, and the variational approach within the B-spline function are introduced for minimization of the consumed electrical energy of a robot manipulator system. In the proposed method, the inverse matrix of Jacobian calculation is not required, and the minimal trajectory error is guaranteed while minimizing the consumed electrical energy. The theory, simulation, and experiments are presented in the paper.

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.