Heikki Handroos

Technical note Sliding mode control for a class of hydraulic position servo

Abstract The present paper discusses a sliding mode control method suitable for controlling a hydraulic position servo with a flexible load. Due to the requirements of the control system, the reference model is designed for the case of a flexible load. The core of sliding mode control is switching surfaces which are specially defined for the system according to the mathematical model and control goal. A novel algorithm is developed, and the simulated and experimental results of the control system are shown. The results show the extremely good robustness of the proposed method.

Design of parallel intersector weld/cut robot for machining processes in ITER vacuum vessel

This paper presents a new parallel robot Penta-WH, which has five degrees of freedom driven by hydraulic cylinders. The manipulator has a large, singularity-free workspace and high stiffness and it acts as a transport device for welding, machining and inspection end-effectors inside the ITER vacuum vessel. The presented kinematic structure of a parallel robot is particularly suitable for the ITER environment. Analysis of the machining process for ITER, such as the machining methods and forces are given, and the kinematic analyses, such as workspace and force capacity are discussed.

Deflection compensation of a flexible hydraulic manipulator utilizing neural networks

Deflection compensation of flexible boom structures in robot positioning is usually done using tables with inverse kinematics solutions. The number of table values increases greatly if the working area of the boom is large and the required accuracy is high. On the other hand, inverse kinematics problems are very nonlinear, and if the structure is redundant, in some cases it cannot be solved in closed form. If the flexibility of the structure is taken into account, the problem is almost impossible to solve using analytical methods. Neural networks offer a possibility to approximate any linear or nonlinear function. Four different methods of using neural networks in the static deflection compensation of a flexible hydraulically driven manipulator are presented. Training information required for training neural networks is obtained by employing a simulation model that includes elasticity characteristics. The functionality of the presented methods is tested based on simulated results of positioning accuracy. The positioning accuracy is tested in 25 separate coordinate points. For each point, positioning is tested with five different mass loads. The mean positioning error of a manipulator decreases from 48 to 5.8 mm in the test points. This accuracy enables the use of flexible manipulators in the positioning of larger objects.

Computationally efficient two-regime flow orifice model for real-time simulation

In fluid power system simulation, orifice flow is, in the main, clearly in the turbulent area. Only when a valve is closed or an actuator driven against an end stopper does the flow become laminar as pressure drop over the orifice approaches zero. So, in terms of accuracy, the description of laminar flow is hardly necessary. Unfortunately, when a purely turbulent description of the orifice is used, numerical problems occur when pressure drop becomes close to zero since the first derivative of flow with respect of pressure drop approaches infinity when pressure drop approaches zero. Furthermore, the second derivative becomes discontinuous, which causes numerical noise and an infinitely small integration step when a variable step integrator is used. In this paper, a numerically efficient model for the orifice flow is proposed using a cubic spline function to describe the flow in the laminar and transition areas. Parameters for the cubic spline are selected such that its first derivative is equal to the first derivative of the pure turbulent orifice flow model in the boundary condition. The key advantage of this model comes from the fact that no geometrical data is needed in calculation of flow from the pressure drop. In real-time simulation of fluid power circuits, a trade-off exists between accuracy and calculation speed. This investigation is made for the two-regime flow orifice model. The effect of selection of transition pressure drop and integration time step on the accuracy and speed of solution is investigated.

Novel Online Re-Calibration Method for Multi-Axis Force/Torque Sensor of ITER Welding/Machining Robot

This paper proposes a new method for re-calibration of multi-axis force/torque sensors. The method makes several improvements to traditional methods. It can be used without dismantling the sensor from its application and requires a smaller number of standard loads for calibration. It is also cheaper and faster method in comparison with traditional calibration methods. The method was developed in response to re-calibration issues with the force sensor used in the welding/machining robot of the international thermonuclear experimental reactor (ITER) vacuum vessel (VV) and the approach aimed to avoid dismantling and on-site assembly of components of the ITER robot. A major complication with the ITER robot is difficult to access the robot when it operates inside the VV; especially after the first plasma, which will make the vessel radioactive. The proposed technique is based on the design of experiment methodology. It has been successfully applied to different force/torque sensors and this paper presents experimental results for the use of the calibration method with the force sensor that would be used on the ITER robot.

Challenges of Vision for Real-Time Sensor Based Control

Sensor-based robot control allows manipulation in dynamic environments with uncertainties. Vision offers a low-cost sensor modality, but low sample rate, high sensor delay and uncertain measurements limit its usability. This paper addresses three problems: uncertain visual measurements, different sampling rates and compensation of the sensor delay. To alleviate the problems above an approach for visual tracking of a moving object with end-effector mounted camera is presented. The pose of the object relative to the camera is determined with model based pose estimation method. The absolute pose and velocity of the target object are estimated by fusing visual measurements over time. A low sample rate visual measurement with sensor delay is integrated with the pose of the end-effector in an extended Kalman filter. Experiments with a 5-DOF parallel hydraulic manipulator show that integration of several measurements together with sensor delay compensation significantly reduce oscillations and phase shift in visual control.

Smooth Transition from Motion to Force Control in Robotic Manipulation Using Vision

Sensor-based robotic manipulation is becoming more and more popular as it promises increases in productivity, flexibility, and robustness of manipulation. Combining visual and force sensing is currently one of the most promising approaches for sensor-based manipulation, as vision and force are two complementary sensing modalities. One approach for multi-sensor use is the traded control where the robot is at each time controlled using one sensing modality, and the controllers are switched based on sensory input. One of the major problems with such systems is the transition between visual and force controllers. In this paper, we present a smooth transition method from motion to force control. The velocity of the end-effector is controlled by estimating the distance to the target by vision and determining an optimal velocity profile giving rapid approach and minimal force overshoot. Experiments show that the proposed control scheme is superior to earlier approaches

Robot positioning of a flexible hydraulic manipulator utilizing genetic algorithm and neural networks

Robot positioning requires that the actuator positions are calculated as a function of end effector position. This mapping is called inverse kinematics of a robot. The inverse kinematics problem is very nonlinear and in some cases it cannot be solved in closed form. Several iterative and neural network approaches are studied in solving the inverse kinematics problem. Deflection of the manipulator arms due to flexibility and mass load causes positioning error. The magnitude of the error depends on the amount of mass load and arm positions and the stiffness characteristics of arms. In this paper a method based on genetic algorithm is used to solve the inverse kinematics of a three degrees of freedom log crane. Neural networks are used to solve the correction values for deflection compensation.

ERROR MODELLING AND DIFFERENTIAL-EVOLUTION-BASED PARAMETER IDENTIFICATION METHOD FOR REDUNDANT HYBRID ROBOT

Abstract This paper focuses on the geometrical error modelling and parameter identification of a 10 degree-of-freedom (DOF) redundant serial—parallel hybrid intersector welding/cutting robot (IWR). The proposed hybrid robot consists of a kinematically redundant 4-DOF serial mechanism to enlarge workspace and a 6-DOF Stewart parallel robot to improve the end-effector accuracy. Due to its redundant degrees of freedom and the serial—parallel structure, the traditional error modelling and identification methods which tailored for pure serial robot or pure parallel robot cannot be directly used. In this paper, a hybrid error modelling method for redundant serial—parallel hybrid robot is presented by combining both the traditional forward calibration and inverse calibration method. Furthermore, because of the high nonlinear and multi-modal characteristics of the derived hybrid error model, the traditional iterative linear least-square algorithm cannot be utilized to identify the error parameters. In this paper, an easy-to-use and powerful evolutionary global optimization algorithm named differential evolution (DE) is employed to search for a set of optimum combination of all error parameters in the error model to minimize the discrepancies of measured and predicted leg lengths. Numerical simulation and analysis are conducted by generating random manufacturing and assembly errors within the real error parameter tolerance range. Meanwhile, different measurement poses of the end-effector and the corresponding joint displacements of the serial mechanism are also randomly generated in the workspace to simulate the real physical behaviours. The simulation results show that the DE-based parameter identification method is robust and reliable, and all of the preset errors can be successfully recovered. The simulation also shows that the hybrid calibration method can avoid the external pose measurement of the connecting point between serial and parallel mechanism, and the pose measurement of the end-effector of serial—parallel robot can satisfy the calibration purpose effectively.

Complementing visual tracking of moving targets by fusion of tactile sensing

Robot control in uncertain and dynamic environments can be greatly improved using sensor-based control. Vision is a versatile low-cost sensory modality, but low sample rate, high sensor delay and uncertain measurements limit its usability, especially in strongly dynamic environments. Vision can be used to estimate a 6-DOF pose of an object by model-based pose-estimation methods, but the estimate is typically not accurate along all degrees of freedom. Force is a complementary sensory modality allowing accurate measurements of local object shape when a tooltip is in contact with the object. In multimodal sensor fusion, several sensors measuring different modalities are combined together to give a more accurate estimate of the environment. As force and vision are fundamentally different sensory modalities not sharing a common representation, combining the information from these sensors is not straightforward. We show that the fusion of tactile and visual measurements enables to estimate the pose of a moving target at high rate and accuracy. Making assumptions of the object shape and carefully modeling the uncertainties of the sensors, the measurements can be fused together in an extended Kalman filter. Experimental results show greatly improved pose estimates with the proposed sensor fusion.