This paper presents a unique method to cancel the backlash present in geared drive systems and thrust-wire systems. Large-sized robots need larger actuators with a high inertia to obtain the desired amount of torque. However, the use of large actuators will increase the effect of gravity. To compensate for the gravitational effect, motors with a high torque are again necessary. Therefore, geared drive systems are used to obtain higher torques and decrease the actuator size. However, adding gears will create a backlash problem. Moreover, when it is necessary to make lighter robots, thrust wires can be used to transmit the motion from the motor to the end effector while keeping them separated. However, the mechanical backlash inside the thrust wire is again a matter of concern. If geared systems and thrust-wire systems are used for the transmission of haptic information in bilaterally controlled systems, the backlash will greatly deteriorate the performance of the system and necessitate compensation. Therefore, a novel method is proposed to uniquely cancel the backlash in the gears and thrust wires used in bilateral teleoperation with twist control. This method uses two identical counter-operating drives per joint to compensate for the backlash effect with a unique control structure that can be used in both geared systems and thrust-wire systems. Experiments have been carried out for the proposed method to validate its accuracy.
Bilateral control is one of the control methods of teleoperation system. The presence of network delays between a master robot and a slave robot makes the design of bilateral control system challengeable. Because, it seriously deteriorates the performance and possibly makes the system unstable. The communication disturbance observer (CDOB) has been proposed to compensate time delay effect without a time delay model. The effectiveness of CDOB with acceleration based four-channel bilateral control system has been confirmed experimentally so far. The acceleration based four-channel bilateral control system utilizes the information of the master/slave position and force. Position and force control are realized in two decomposed virtual modal spaces, i.e. differential mode and common mode. In the acceleration based four-channel bilateral control system with CDOB, position control is stable, but force control is not stable. Thus, the control system is stable during free motion, but the system can be destabilized when the slave robot contacts with the environment. In this paper, a structural approach is applied to improve the stability of force control of bilateral control system. The proposed structure is derived from modal space analysis in differential mode and common mode. By using the proposed structure, position control and force control of the bilateral control system are stabilized. The validity is confirmed by experimental results in the case of constant delay and time varying delay.
Delicate movement with high-precision performance is required in industrial areas such as medical care and semiconductor processing. Taking the expansion of the robotic movement area into account, attainment of force control is the one of the fundamental techniques for working in a microspace. To focus on the force information in a microspace, the transmission of microforce information is important to avoid destruction of the delicate object. However, the quality of force information is deteriorated by the influence of a stochastic disturbance such as quantization noise and sensor noise. In other words, high-precision force control is closely related to a reduction in the influence of the stochastic disturbance. Therefore, this paper focuses on the stochastic disturbance as a quantum fluctuation that is derived as the Schrödinger equation. In order to reduce the stochastic disturbance, a novel modeling method that is focused on quantum fluctuation is proposed. In this paper, a resonant filter is introduced by a complex equivalent electrical circuit of the Schrödinger equation, and a reaction force observer is implemented. The viability of the proposed method and the reduction in the oscillation are confirmed by the displacement of the poles and experiments.
In the field of industrial manufacturing, there are still many parts processed manually. To overcome this problem, a motion-copying system is proposed to make a robot execute the work of a human. However, processing motion data and reproducing new motion data that can be applied to various scenes are required to extend the versatility of the motion-copying system. In this paper, a temporal compensation method is proposed in order to maintain the reproducibility of a saved human motion because loaded human motion data are different from saved human motion data when motion data is processed by spatial scaling. When saved human motion data is processed by spatial scaling, reaction force of saved human motion data and that of loaded human motion data are different, and the reproducibility decreases. Here, temporal compensation, which is based on temporal scaling, is proposed in order to equalize the reaction force of saved human motion data and that of loaded human motion data. Validity was confirmed by the using experiments that save and reproduce writing motion.
This paper describes an optimized design of secondary coils in the inductive power transfer (IPT) system for railway vehicles. The gap between the primary and secondary coils is large, and the weight of the on-board secondary coil needs to be reduced. Thus, we propose a design method of dimensions of the secondary coils and unique cross-section of the secondary coils to improve the induced voltage and to reduce the self-inductance of the secondary coil. Further, practical issues that arise when applying the design to the railways, for example, human exposure to time-varying electric and magnetic fields, are examined.
This paper proposes two types of force control systems based only on measurements of an acceleration sensor attached on the tip position of a three-link planar redundant manipulator. To estimate the acceleration with high accuracy, a Kalman filter (KF) based on a dynamic model of the manipulator and an extended Kalman filter (EKF) based on a kinematic model of the manipulator are introduced in each system. Further, a disturbance force observer (DFOB) and a reaction force observer (RFOB) are also implemented in the force control system, and the acceleration estimated by the KF/EKF is directly utilized in those observers. Simulations and experiments are conducted to examine the characteristics of the estimation performance and compare the force control performance of the two types of systems based on the KF and the EKF.
This paper describes a novel offline parameter identification technique for an interior permanent magnet synchronous motor. In general, the motor controller requires its parameter information to start the motor. Therefore, offline parameter identification is indispensable to avoid degradation in the control performance. The proposed technique changes the configuration of the current regulator and seeks the minimum point of the current norm where the parameter mismatch is eliminated. The unique feature of the method is the use of only the current information and its insensitiveness to the winding resistance variation due to a temperature change. The paper presents an identification process for the q-axis inductance Lq, the winding resistance R, the magnetic flux linkage Ψ and the d-axis inductance Ld. In addition, some simulation and experimental tests are conducted, and their results are presented to demonstrate the validity of the proposed technique. In the final part of the paper, the sensitivity to the winding resistance is also checked by the experimental test.
Bilateral control is one of the control methods for teleoperation system. The presence of network delays between a master robot and a slave robot makes the design of bilateral control system challengeable. It seriously deteriorates the performance and possibly makes the system unstable. In this paper, to overcome the destabilization from network delay, the model-free time delay compensator is proposed. The proposed compensator does not utilize time delay model and plant model, but the bilateral control system is stabilized. The stability of the proposed control system is shown by using two flow expression method. Two flow expression method models the master system and the slave system of four-channel bilateral control system as single input and single output system. This method makes the analysis for the stability and the performance of the four-channel bilateral control system simple. The validity of the proposed model-free time delay compensator is confirmed by numerical and experimental results.
We developed a quick-response technique for simplified vector control without using position sensors for permanent magnet synchronous motor drives. Our original simplified vector control has neither current control nor speed control, and the speed estimation response is limited by motor constants. We developed a highly stable method that equivalently increases the damping ratio, which is related to motor constants by the control side. Time domain simulations and experiments were perfomed to demonstrate the effectiveness of the proposed method.
Robust motion control against dynamic torque is required for rapid and precise motion control of industrial robots. In this regard, a disturbance observer (DOB) is widely used to achieve robust motion control. In general, it is difficult to achieve robust motion control against a step load torque because the DOB exhibits an estimation delay. To overcome this problem, this paper proposes a new method involving the use of a Kalman-filter-based instantaneous state observer for load torque compensation. The proposed method achieves the instantaneous load torque estimation of a two-inertia system using a load-side acceleration sensor. Torque compensation based on instantaneous torque estimation is highly robust against the insertion of a step load torque. The effectiveness of the proposed method is confirmed by performing both a numerical simulation and experiments using an industrial robot arm.
In this paper, we propose a novel method to reduce the rotor vibration of a stepping motor driven by a microstep driver. A low-pass-filter-type pre-compensator with the structure of a type-2 closed-loop control system is used to make it possible to eliminate the control lag in constant-speed tracking. The proposed approach enables the systematic design of a pre-compensator in the frequency domain. An experiment was performed to prove the effectiveness of the present method.
The paper presents a novel modeling approach of the power flow in a bidirectional dual active bridge DC-DC converter. By using basic superposition principles, the mathematical distinction of cases is avoided in the modeling process of the high-frequency transformer currents for different types of modulation. The generalized model is used in the optimization of the converter losses of a 3.3kW electric vehicle battery charger with an input voltage of 400V and a battery voltage range from 280V to 420V. Besides the commonly used control variables such as the phase-shift and the clamping intervals, the variation in the switching frequency is also considered in the optimization process. The optimal modulation including the frequency variation leads to an increase in the converter efficiency up to 8.6% using IGBTs and 17.8% using MOSFETs at the most critical point compared to phase-shift modulation at a fixed switching frequency.
Recently, increased attention is being paid to power supply networks using energy storage devices such as batteries. Network topologies using bi-directional isolated DC-DC converters of low or medium capacity are required for the diversification of power supply networks. The dual active bridge (DAB) DC-DC converter is one of the most effective bi-directional isolated DC-DC converters. However, the circuit has some inherent problems such as degradation of power efficiency and the occurrence of surges during light-load operation. In this paper, we propose a control technique to solve these problems. From the experimental results, it is confirmed that the maximum power efficiency improvement was 16% for a light load. Applying two operation modes, the proposed operation in light load and the conventional operation in heavy load, the circuit can be operated across a full range of road. To switch between the two modes seamlessly, the precise boundary point of the two modes is needed for feedback control. Therefore, a precise static characteristic analysis with loss was carried out. From the results, the loss included simple equivalent circuit model was obtained. The root mean square error between the proposed analysis and the experimental results is within 4%.
This paper proposes a new type of Modular Multilevel Converter (MMC) using a three-winding transformer. In general, MMCs require a buffer reactor in each arm, which increases the number of components required in the converter and the converter footprint. The proposed MMC with a three-winding transformer does not require buffer reactors. The mathematical representation of the proposed converter was first described. The operational performance of the proposed MMC was found to be identical to the typical MMC (with the buffer reactor topology) by making comparisons in the experimental tests on two physical prototype converters rated at 10kVA.
This paper proposes a novel control method in the flux-weakening region. The proposed method utilizes a maximum torque per flux (MTPF) control at full load, and an efficient flux-weakening control at light load. In order to simplify the MTPF control, the approximate method of MTPF curve is proposed. In the flux-weakening control, a more efficient control obtained by minimizing the armature current at light load is applied. In the flux-weakening region, it is possible to operate at the maximum torque by applying the MTPF control. In addition, if the load torque is light, operation along a constant voltage ellipse is more efficient compared to the operation of the MTPF control along the MTPF curve owing to the decrease in copper loss with decreasing armature current. In the proposed efficient flux-weakening control, a current command generation method in which the armature current ampere is adopted for operation along the constant voltage ellipse is also proposed. The validity of the proposed control method is verified by simulation and experimental results.
A laminated core has advantages such as low iron loss, high permeability, and high saturation flux density. For inductor applications, an air gap has to be inserted into the magnetic circuit to avoid saturation at lower magnetic field strengths. However, when an air gap is inserted, fringing flux leaks out of the air gap and concentrates on the surface of the core. As a result, the iron loss increases to 292% as compared with an ungapped core, and the inductance characteristic deteriorates. Although a core with chamfered edges is known as a prior art that reduces the iron loss to 147%, there is scope for improvement. In this paper, we propose a core with curved edges for loss reduction of the laminated core inductor. Consequently, the proposed shape reduces the iron loss to 113% and improves the inductance characteristic. Thus, the proposed method increases the efficiency and reduces the size of the on-board charger.
The dynamic voltage restorer (DVR) is a power electronics based solution for mitigation of voltage sags and voltage swells effects on sensitive loads that injects voltages in series with the grid. Typically, the controller structure for a DVR comprises an inner current loop and an outer voltage loop. Usually, a proportional controller or a proportional integral controller is used for the current loop, and a resonant controller is used for the voltage loop. This paper presents the design of a robust controller for the voltage tracking loop of a DVR that guarantees robust stability against load parameter variation. Moreover, the proposed controller ensures the tracking of a sinusoidal voltage waveform as well the rejection of the nonlinear load current influence, both with a prespecified error. The voltage controller design is based on the H∞ parameter specification approach. All performance and robustness requirements are specified and analyzed on the basis of the frequency response plot of the closed loop transfer function. The proposed controller performance is validated by simulation and by experiments carried out on a low-scale DVR prototype.
In this study, a PV microinverter was implemented using an interleaved flyback converter (IFC) with sinusoidal PWM (SPWM) control. A double-frequency distortion due to the use of a sawtooth-based SPWM was found in the grid current. A forward compensation circuit that eliminates the distortion on the primary side was explored to simplify the output filter design. The maximum power transfer between the PV module, IFC, and grid end was modeled. An elaborate experiment was demonstrated to verify the analysis results and predicted performance.
In this paper, a LED dimming circuit powered by the KY converter, which is named after the first capitals of two authors' names: Kuo-Ing Hwu and Yeu-Torng Yau, is presented, which is controlled based on the field-programmable gate array (FPGA). By a given dimming command and the proposed maximum gate voltage detector, the voltage across the MOSFET in the linear current regulator can be reduced so as to upgrade the efficiency of the overall system. Aside from this, each LED string takes level dimming, and is powered by the KY converter, which has an output inductor and hence upgrades the life of the output capacitor. Furthermore, via some experimental results, the efficiency based on the proposed control method is higher than that based on the traditional control one.
The Virtual Synchronous Generator (VSG) is an inverter control structure that supports power system stability by imitating a synchronous machine. Because of the restriction of inverter power and current, the VSG performance under disturbances should be evaluated and enhanced. In this paper, the response of the VSG unit to symmetrical and unsymmetrical voltage sags is assessed. A theoretical analysis that traces the trajectory of the state variable of the system during voltage sags is presented. The analysis confirms the effect of the characteristics of symmetrical and unsymmetrical voltage sags on the severity of their consequences. In addition, it is detected that two types of transients appear that must be mitigated: one is the transients during the voltage sag and the other one is the transients after voltage recovery. To prevent overcurrent during voltage sags, voltage amplitude control and output power control are implemented, and to suppress the transients after voltage recovery, virtual inertia control is implemented. The experimental results from a 10kVA VSG-controlled inverter confirm the effectiveness of the additional controllers.
This paper presents line loss minimization in a radial distribution system using multiple Static Synchronous Compensators (STATCOMs) and static capacitors. The reactive current references of the STATCOMs and the number of installed static capacitors for line loss minimization are derived. Under line loss minimization, the reactive current of each load is completely compensated by the closest STATCOMs at the left and right sides of the load. The compensating reactive currents of the STATCOMs for each load can be calculated by the line resistances between the STATCOMs and the reactive current of the load. The number of installed static capacitors for line loss minimization is selected as the closest integer to the optimal real number for achieving line loss minimization. The effectiveness of the proposed line loss minimization is verified experimentally using a laboratory prototype system.