Advancements in automation have pushed forward the development of collaborative robots, which are designed to work together with humans. The majority of actual commercial collaborative robots use geared motors with some kind of torque sensing, and their performance is directly related to how well torque is controlled. However, imperfections in the gear fabrication causes transmission error in the geared motor, producing undesired vibrations which affect the torque control performance. In this study, an integratorless 2DOF (two degrees-of-freedom) torque control scheme is proposed to improve the torque response in such case. The proposed control scheme uses first-order FPIDO (force-position-integrated disturbance observer) which nominalizes and achieves motor side acceleration control, allowing to use the natural mechanical dynamics of the motor side as an integrator instead of implementing it on the control system. This is combined with a derivative-type torsional disturbance observer (DTDOb), which uses torque sensor information to compensate all torsional disturbances and nominalize the torsion torque dynamics, achieving 2DOF (Two degrees of freedom) torque control. The effectiveness of the proposed methodology is verified by simulation and experimental results.

This paper proposes a load-side acceleration control scheme for geared motors affected by unknown backlash and friction. A backlash-deactivated disturbance observer (BDDOb) is proposed to achieve load-side acceleration control while ensuring quick traversing of backlash. The BDDOb works by executing conditional estimation-compensation of load disturbances, and is driven by a backlash detection scheme based on torsion torque sensing which does not require any information on the backlash gap width. In addition, a variable-order force-position-integrated disturbance observer (VOFPIDO) is proposed to quickly compensate nonlinear friction without requiring information of the nonlinear friction magnitude by applying velocity-dependent switching. Experimental results prove the effectiveness of the proposed method.

This paper focuses on single inertialization on the load-side parameter of two-inertia system and employs the equivalent disturbance compensator (EDC). Conventional EDC is designed in case of the plant system without friction; thus, the back-forward drivability control using the EDC is affected by friction. In this paper, we propose the back-forward drivability control using EDC and DOB for single inertialization of two-inertia system. DOB is used to compensate the friction and the EDC is applied to not only improve the back-forward drivability, but also suppress the vibration on load-side output. This proposal consists of basic actuator and force sensor and it is available to general robot actuator that uses current sensor, motor encoder, and force sensor. The effectiveness of proposed back-forward drivability control is verified through the numerical simulation.

Good forward-drivability characteristics are essential for applications such as position, velocity, and force control. Geared motors, usually employed in industrial robots, require lubrication in their gears, and consequently, some form of sealing is needed to avoid leaking. While oil seals prevent leaking, they increase nonlinear frictional effects, which result in stick-slip during velocity zero-crossings, thereby deteriorating the forward-drivability. This paper proposes a stick-slip suppression method based on a velocity-driven N-order stick compensator (VDNSC), which works as an order-variable friction compensator formed by the velocity-driven switching of the internal components of a high order Q-filter. Experiments with torque and velocity control confirm that the proposed method improves the forward-drivability, allowing for more precise motion even under large friction.

Load-side acceleration sensing based on load-side acceleration control is proposed for a two-inertia resonance system. The proposed method is divided into two stages. In the first stage, the feedback component from the load-side is canceled using the estimated value of the instantaneous state observer and the load-side acceleration. The two-inertia resonance system is converted into a simple second-order system using this method. The proposed method configures a control system of cascaded loops for the second-order system. Therefore, the load-side acceleration control is achieved by two specified bandwidths. The proposed method is easy to implement for conventional robots and yields superior cost-savings compared with other types of load-side acceleration control. The effectiveness of the proposed load-side acceleration controller is verified by theoretical analysis, numerical simulations, and experiments.

In recent years, many kinds of renewable-energy solutions, such as wind, power, and photovoltaic energy, have attracted much attention as well as the distributed power grids into which they are integrate. These systems have a highly reliable power supply. However, to connect DC buses to other devices, many isolated converters are require to ensure isolation from the DC bus. Isolation is essential for safety purposes. However, the use of many isolated converters is not preferable because of problems with the increasing system size. Triple action bridge (TAB) converters have the features of bidirectional power flow and small size, which is suitable for integration into distributed power grids. TAB converters that are based on dual active bridge (DAB) converters have useful characteristics of DABs. However, there remains a problem with respect to the interaction between two current loops with TABs, and there are few solutions to this problem when an AC supply is connected to the system. In order to enable the connection from a TAB converter to an AC system, this paper proposes a method to solve the interference problems in the circuit and control. The suitability of the propose solution is shown using simulations.

In order to realize a fine force sensorless force control, this paper proposes a new notch-type friction-free disturbance observer(DOB) and a new notch-type friction-free reaction force observer(FFRFO). Generally, a force sensorless force control system always has influence on static friction phenomenon. To overcome this problem, this paper proposes a new force sensorless control system using the notch-type friction free reaction force observer, whose inner system is the proposed acceleration control based on the notch-type friction free disturbance observer. The effectiveness of the proposed system is confirmed by experiments using an actual industrial robot.

This paper proposes a robust position control based on state feedback using double disturbance observer for two mass system. In load-side position control based on state feedback using disturbance observer, performance of control system is depend on bandwidth of disturbance observer. To solve problem, the load-side position control using I2-PI controller based on state feedback is proposed. The proposed system consists of PD position controller, motor-side acceleration controller, I2- PI controller equivalent of states feedback, velocity deviation feedback, and a zero-order load-side disturbance observer (ZLDOB) with variable noise-covariance (VNC) Kalman filter. The simulation and experimental results confirm that the load-side position control using I2-PI controller based on state feedback reduces the vibration in response load-side position. Additionally, the proposed method reduces the vibration in response load-side acceleration.

This paper proposes a new sensorless force control system based on reaction force estimation system considering static friction and kinetic friction for industrial robot. As industrial robot uses the harmonic gear and the ball screw, the proposed force control system is structured by the resonance ratio control based speed controller. Moreover, as harmonic gear has a static friction and kinetic friction, the reaction force estimation system needs to reduce the influence on static friction and kinetic friction force. The proposed friction-free force estimation system has the two technics. The first one is the dither signal based static friction free force estimation, and the second one is the Stribeck model based kinetic friction free estimation. Using these technics, the friction-free reaction force estimation system is constructed. In this paper, the validity of proposed system is verified by the experimental results using the actual industrial robot.

In the recent times, the number of amateur farmers has been increasing. Although amateur farmers use tillers, they may find it difficult to drive the tiller along a straight line. There is a need for electric tillers with independent drive wheels. This paper proposes straight control incorporating human driving input by using a disturbance observer. Numerical simulation results and experimental results confirm the validity of the proposed method.

This paper proposes a transmission torque linearization method for a leg robot using bi-articular simultaneous drive mechanism. A planetary gear is the primary mechanism of a bi-articular simultaneous drive system. Energy consumed for walking is reduced by the use of bi-articular muscles. However, the planetary gear causes a torque transmission loss in the leg robot system. The proposed method linearizes the nonlinear torque transmission by using a disturbance observer and the estimated self-weight disturbance torque. The experimental results point out that the proposed method linearizes the transmission torque.

This paper proposes a stable force control system to achieve the force control of industrial robots. The proposed stable force control system is designed by using an I-P force control system and a resonance ratio control system considering the required interaction with the environment. To design the resonance ratio control system considering the contact with the environment, this paper analyzes the stability of the proposed resonance ratio control and the bandwidth of a disturbance observer. This paper designs the resonance ratio control system using an instantaneous state observer considering the results of the stability analysis. In addition, a stable force control system is designed based on the resonance ratio control system using an instantaneous state observer. In this paper, the effectiveness of the proposed force control system is confirmed by performing numerical simulations and experiments using an industrial robot arm.

Pickup systems for optical disks operate in the horizontal and vertical directions for tracking and focusing, respectively. Conventional optical disk control systems are designed for each operated direction. On the other hand, in this paper, we consider the operation trajectory to implement a method for concurrent control of focusing and tracking of optical disks. This paper proposes a two-dimensional feedforward control system. The proposed control system consists of a feedforward controller and a two-dimensional feedforward compensator. The error vector is a combination of the focusing error and the tracking error of the compensator. The feedforward controller based on a zero phase error tracking control (ZPETC) regulates the error vector. The experimental results show the effectiveness of the proposed control system.

This paper proposes a new acceleration control method for industrial robots. To achieve high-speed and high-precision path tracking control, the motion control system must be able to control the arm acceleration. However, it is difficult to directly control the arm acceleration, because the speed reduction mechanism of industrial robots has axial torsion. To solve this problem, this paper proposes a new acceleration control method based on an inverse system. The proposed method designs an inverse transfer function system from the current command for arm acceleration. Acceleration control is achieved by the feed-forward control input designed from the inverse system. In addition, a robust feedback control system is designed to compensate for the dynamic disturbance torque and parameter variation. The effectiveness of the proposed method is confirmed by experimental results using a 2-DOF robot manipulator.