Giacomo Ziliani

Iterative-Learning Hybrid Force/Velocity Control for Contour Tracking

In this paper, we propose a new method, which is based on an iterative-learning-control (ILC) algorithm, for the contour tracking of an object of unknown shape performed by an industrial robot manipulator. In particular, we consider (both implicit and explicit) hybrid force/velocity control whose performance is improved by repeating the task. Here, a time-based reference signal is not present, and therefore, a new approach has been developed, which is different from the typical applications of ILC. Experimental results show the effectiveness of the technique.

Gain scheduling for hybrid force/velocity control in contour tracking task

In this paper a gain scheduling approach is proposed for the hybrid force/velocity control of an industrial manipulator employed for the contour tracking of objects of unknown shape. The methodology allows to cope with the configuration dependent dynamics of the manipulator during a constrained motion and therefore a significant improvement of the performance results. Experimental results obtained with an industrial SCARA manipulator demonstrate the effectiveness of the technique.

A Mechatronic Design for Robotic Deburring

This paper deals with the implementation of a mechatronic methodology for the robotic deburring of planar workpieces with an unknown shape performed by an industrial manipulator. The approach is based on the use of a hybrid force/velocity control law and on a correlated suitable design of the adopted deburring tool. Experimental results, obtained with a two degrees-of-freedom SCARA industrial robot manipulator, show the effectiveness of the method. I. INTRODUCTION Robots that are able to autonomously adapt themselves to semi-unstructured tasks are nowadays a key issue in several industrial applications, such as grinding, debur- ring, chamfering and polishing, where the capability to cope with a workpiece of an unknown shape would significantly reduce the task programming phase, espe- cially when frequent changes in the production occur. Various methods to perform effectively a deburring task have been proposed in the literature. They are based for example on impedance control (1), on hybrid control with an internal position loop (2) (note that in this case the geometry of the workpiece is known), and on the so-called triangular control (3). Soft-computing techniques can be also adopted (see e.g. (4), (5)). The idea to design the deburring tool in conjunction with the control algorithm has been pursued in (6), (7). In particular, in (6), under the framework of impedance control, a roller bearing mounted on the force sensor is employed for the purpose of tracking. In this case however two force sensors are needed. Differently, in (7) an end-effector mounted jig and a proper design of the compliance of the manipulator are used to achieve accurate force guidance. An admittance control law is proposed but no results are given to prove the effectiveness of the method. In any case, when dealing with an unknown object, the robot has to perform a con- tour tracking task effectively, by imposing an appropriate normal force and an appropriate tangential velocity to the robot end-effector. In this context, the use of hybrid force/velocity control (8) is addressed in this paper. It will be shown that it can be effectively adopted for this task in conjunction with a suitable design of the milling tool. Indeed, by suitably adopting the measured torque along the vertical axis, it is possible to detect the contact angle in order to achieve an accurate tracking (9). Further the presence of burrs can be detected and consequently the reference tangential velocity can be suitably modified to improve the quality of the deburring.

Friction Compensation in Hybrid Force/Velocity Control for Contour Tracking Tasks

Nowadays robots in industrial settings are mainly used for repetitive tasks where they act as programmable devices reproducing previously recorded motions in a highly structured environment so that decision and initiative ca-pability is rarely exploited. Contour tracking is, on the contrary, an example of a complex task that requires the manipulator to continuously and autono-mously modify its path, coping with the uncertainties typical of unstructured environments (Siciliano & Villani, 1999). In many applications a robot is re-quired to follow a contour while applying a normal force; these tasks include grinding (Thomessen & Lien, 2000), deburring (Ferretti et al., 2000; Ziliani et al., 2005), shape recovery (Ahmad & Lee, 1990), polishing and kinematic cali-bration (Legnani et al., 2001). The problem of tracking (known and) unknown contours has been studied by many researchers in the last two decades. Hybrid force/velocity control (Raibert & Craig, 1981) appears to be suitable to be adopted in this context, because it explicitly controls the end-effector force in a selected direction and the end-effector velocity in the other complemen-tary directions. Actually, two kinds of hybrid force/velocity control can be implemented (Roy & Whitcomb, 2002): 1)

An iterative learning control algorithm for contour tracking of unknown objects

In this paper we propose a new method, based on an iterative learning control (ILC) algorithm, for the contour tracking of a planar object of unknown shape performed by an industrial SCARA robot manipulator. In particular, we adopt a hybrid force/velocity controller whose performance is improved by repeating the task. Differently from the typical applications of ILC, here a reference position signal is not present and therefore a new approach has been developed. Experimental results show the effectiveness of the technique

On the elasticity in the dynamic decoupling of hybrid force/velocity control in the contour tracking task

The paper investigates the decoupling of an explicit hybrid force/velocity control of robot manipulators employed in the contour tracking task of objects of unknown shape taking into account the elastic model of the contact and of the robot. The proposed controller allows the achievement of the decoupling of the normal force and tangential velocity control loops and can be expressed as a multi-input multi-output (MIMO) time-varying proportional-integral-derivative (PID) controller. Experimental results demonstrate the effectiveness of the method.

A comparison between implicit and explicit hybrid control for contour tracking

Abstract In this paper, the use of implicit (position based) and explicit hybrid force/velocity control for contour tracking task of unknown (planar) objects is discussed and compared from an industrial point of view. In particular, the joint friction compensation issue is addressed. A large number of experimental results obtained with a 2 degree-of-freedom SCARA industrial manipulator is shown to support the investigation.

Iterative learning explicit hybrid force/velocity control for contour tracking

In this paper we propose an Iterative Learning Control (ILC) algorithm for the contour tracking of unknown planar objects performed by an industrial SCARA robot manipulator with an explicit hybrid force/velocity controller. Conversely to the typical applications of ILC, the employed control law does not include a time-based reference position signal (and a position control loop) and therefore a different approach has been developed in order to exploit the repetitive nature of the task. Design choices are described in this context and experimental results show the effectiveness of the technique.

A GAIN SCHEDULING APPROACH FOR HYBRID FORCE/VELOCITY CONTROLLED ROBOT CONTOUR TRACKING

Abstract This paper deals with the implementation of a hybrid force/velocity controller for the contour tracking of an object of an unknown shape performed by an industrial robot manipulator. In particular we propose the use of a gain scheduling approach in order to cope with the configuration dependent dynamics of the manipulator in constrained motion and therefore in order to allow to obtain satisfactory performances in a very large portion of the robot workspace. Experimental results, obtained with a two degrees-of-freedom SCARA industrial robot manipulator show the effectiveness of the approach.