Dynamic Positioning (DP) is used to automatically maintain the position and heading of a floating structure subjected to environmental disturbances. A DP control system is composed of a motion controller to compute the desired force and moment and a thrust allocator to distribute the computed force and moment to multiple thrusters considering mechanical and operational constraints. Among various thruster configurations, azimuth thrusters or propeller/rudder pairs tend to make the allocation problem difficult to solve, because these types of propulsion systems include nonlinear constraints. In this paper, a dynamic positioning strategy for a twin-thruster ship that is propelled by two azimuthing thrusters is addressed, and a thrust allocation method which does not require a numerical optimization solver is proposed. The applicability of the proposed method is demonstrated with an experiment using an autonomous boat.

Generating motion of center of mass for biped robots is a challenging issue since biped robots can easily lose balance due to limited contact area between foot and ground. In this paper, we propose force control method to generate high-speed motion of the center of mass for horizontal direction without losing balancing condition. Contact consistent multi-body dynamics of the robot is used to calculate force for horizontal direction of the center of mass considering balance. The calculated force is applied for acceleration or deceleration of the center of mass to generate high speed motion. The linear inverted pendulum model is used to estimate motion of the center of mass and the estimated motion is used to select either maximum or minimum force to stop at goal position. The proposed method is verified by experiments using 12-DOF torque controlled human sized legged robot.

An electro-hydraulic actuator (EHA) is widely used in industrial motion systems and the increasing bandwidth of EHA position control is important issue. The model-inverse feedforward controller is known to extend the bandwidth of system. When the system has non-minimum phase (NMP) zeros, direct model inversion makes system unstable. To overcome this problem, an approximate model-inverse method is used. A representative approximate model inversion method is zero phase error tracking control (ZPETC). However, if zeros locate right half plane of z-plane, the approximate inverse model amplifies the high-frequency response. In this paper, to solve the problem of ZPETC, an adaptive model-inverse control is proposed. The adaptive algorithm updates feedforward term in real-time. The effectiveness of the proposed adaptive model-inverse position control strategy is verified by comparison with typical proportional-integral (PI) control and feedforward control by experiments. As a result, the proposed adaptive controller extends the bandwidth of EHA position control.

In this paper, a localization algorithm and an autonomous controller for PETASUS system II which is an underwater vehicle-manipulator system, are proposed. To estimate its position and to identify manipulation targets in a structured environment, a multi-rate extended Kalman filter is developed, where map information and data from inertial sensors, sonar sensors, and vision sensors are used. In addition, a three layered control structure is proposed as a controller for autonomy. By this controller, PETASUS system II is able to generate waypoints and make decisions on its own behaviors. Experiment results are provided for verifying proposed algorithms.

For indoor mobile robots, the performance of autonomous navigation is affected by a variety of factors. In this paper, we focus on the characteristics of indoor absolute positioning systems. Two commercially available sensor systems are experimentally tested under various conditions. Mobile robot navigation experiments were carried out, and the results show that resultant performance of navigation is highly dependent upon the characteristics of positioning systems. The limitations and characteristics of positioning systems are analyzed from both quantitative and qualitative point of view. On the basis of the analysis, the relationship between the positioning system characteristics and the controller design are presented.

본 연구는 능동텐던을 이용 지진을 받는 구조물의 최적 능동제어 방법에 관한 수치해법 적용 및 프로그램 개발에 목적이 있다. 능동텐던 시스템에 의한 제어이론을 적용하기 위해서 Ricatti 폐회로 알고리즘을 이용하였으며, 시간지연 문제를 고려하였다. 최적제어의 정식화를 위해서 SUMT기법의 최적화에 의해 성능지수를 최소로 하는 최적 가중치행렬을 추정토록 하였다. 구조물에서의 능동텐던의 최적 위치 선정을 위해서 가제어지수에 의한 방법을 소개하였다. 수치 예를 통해, 제어기의 최적 위치선정을 고려한 능동최적제어가 지진하중을 받는 구조물의 성능제어에 우수한 효과를 나타내는 것으로 평가되었다.

Recently, automatic parking assist systems are commercially available in some cars. In order to improve the reliability and the accuracy of parking control, pose uncertainty of a vehicle and some experimental issues should be solved. In this paper, following three schemes are proposed. (1) Odometry calibration scheme for the Car-Like Mobile Robot.(CLMR) (2) Accurate localization using Extended Kalman Filter(EKF) based redundant odometry fusion. (3) Trajectory tracking controller to compensate the tracking error of the CLMR. The proposed schemes are experimentally verified using a miniature Car-Like Mobile Robot. This paper shows that odometry accuracy and trajectory tracking performance can be dramatically improved by using the proposed schemes.

일반적으로 서보 제어 시스템에서 비선형 동적 특성을 갖는 마찰력은 제어기 성능에 악영향을 미친다. 특히, 선형으로 고려된 시스템에 제어기 이득을 잘 설계한다 하더라도 마찰 현상에 포함된 동적으로 변화하는 dead zone에 의한 정상상태 오차 및 리미트 사이클(limit Cycle) 등을 야기한다. 따라서, 본 논문에서는 비선형 동적 마찰 성분을 효과적으로 보상하고 적응적으로 제어함으로써 차세대 항만 자동화 이송시스템으로 주목받고 있는 LMTT(linear motor-based transfer technology) 시스템의 위치 정밀도를 향상시키는 것을 목적으로 하고 있다. 본 제어대상은 셔틀카(shuttle car)와 컨테이너들의 다양한 중량과, 이로 인해 발생하는 동적 마찰 특성 파라미터들의 변화가 발생하므로 마찰력 내부 파라미터들의 추정이 요구된다. 제안하는 방법은 적응 backstepping 제어 기법으로 시스템이 안정하게 제어될 수 있는 조건으로 내부 파라미터 추정기를 설계하여 비선형 동적 마찰력을 보상하도록 하였다.

Control of a robot manipulator in contact with the environment is usually conducted by the direct feedback control using a force-torque sensor or the indirect impedance control. In these methods, however, the control algorithms become complicated and the performance of position and force control cannot be improved because of the mechanical properties of the passive components. To cope with such problems, redundant actuation has been used to enhance the performance of position control and force control. In this research, a Double Actuator Unit (DAU) is proposed, with which the force control algorithm can be simplified and can make the robot ensure the safety during the external collision. The DAU is composed of two actuators; one controls the position and the other modulates the joint stiffness. Using this unit, it is possible to independently control the position and stiffness. The DAU based on the planetary gears is investigated in this paper. Performance using the DAU is also verified by various experiments. It is shown that the manipulator using this mechanism provides better safety during the impact with the environment by reducing the joint stiffness appropriately on detecting the collision of a manipulator.

본 논문에서는 항만 자동화를 위해 새로이 제안된 리니어 모터 기반 컨테이너 이송시스템에 지능제어기법을 이용하여 그 정밀도를 향상시키고자 한다. LMCTS(Linear Motor-based Container Transfer System)는 스케일의 거대함 때문에 일반 리니어 모터에서 중요시 되지 않는 정지마찰력과 디텐트럭(detent force)이 정밀제어에 큰 문제가 된다. 특히, 컨테이너 적제유무에 따라 시스템 자체가 급격히 변하므로 기존의 PID형 제어기로는 좋은 성능을 얻기 어렵다. 따라서 본 논문에서는 같은 구조를 갖는 두 개의 DR-FNN(Dynamically- constructed Recurrent Fuzzy Neural Network)를 제어기와 에뮬레이터로 구성하여 이러한 문제를 해결하고자 하였다.

This research aims to seek active control of ball-beam position stability by resorting to neural networks whose layers are given bias weights. The controller consists of an LQR (linear quadratic regulator) controller and a neural networks controller in parallel. The latter is used to improve the responses of the established LQR control system, especially when controlling the system with nonlinear factors or modelling errors. For the learning of this control system, the feedback-error learning algorithm is utilized here. While the neural networks controller learns repetitive trajectories on line, feedback errors are back-propagated through neural networks. Convergence is made when the neural networks controller reversely learns and controls the plant. The goals of teaming are to expand the working range of the adaptive control system and to bridge errors owing to nonlinearity by adjusting parameters against the external disturbances and change of the nonlinear plant. The motion equation of the ball-beam system is derived from Newton's law. As the system is strongly nonlinear, lots of researchers have depended on classical systems to control it. Its applications of position control are seen in planes, ships, automobiles and so on. However, the research based on artificial control is quite recent. The current paper compares and analyzes simulation results by way of the LQR controller and the neural network controller in order to prove the efficiency of the neural networks control algorithm against any nonlinear system.

In this paper, applications of multilayer neural networks to control of flexible robot beam are considered. The multilayer nerual networks can be used to approximate any continuous function to a desired degree of accuracy and the weights are updated by Gradient Method. When a flexible beam is rotated by a motor through the fixed end, transverse vibration may occur. The motor torque should be controlled insuch a way that the motor rotates by a specified angle, while simultaneously stabilizing vibration of the flexible manipulators so that is arrested as soon as possbile at the end of rotation. Accurate control of lightweight beam during the large changes in configuration common to robotic tasks requires dynamic models that describe both rigid body motions, as well as the flexural vibrations. Therefore, a linear dynamic state-space model of for a single link flexible robot beam is derived and PD controller, LQP controller, and inverse dynamical neural networks controller are composed. The effectiveness the proposed control system is confirmed by computer simulation.