Search

New robot applications increasingly require external sensing to accomplish robustness and performance despite the variations and uncertainties that come with the increasing demands on flexibility for manufacturing of customized products. The demands on productivity then require high performance, which for well known feedback-control reasons means short response times from external process event to reaction on the robot motor control. The need for high- performance sensor interfaces connected to the motion control without excessive buffering and computations has therefore gained increasing attention over the last decade. Apart from high sampling frequencies and short latencies, there are also engineering aspects to consider. Control engineering tools should be possible to utilize, additional computing power is better connected via real-time Ethernet connections, and all interfaces should check compatibility without sacrificing performance. Furthermore, all that should preferably coexist with the original system, superimposing the effects on external sensing while still maintaining safety and consistency of data. The reason is that the original control software is model based and built on millions of lines of C-code, and tested beyond the possibilities of a research lab. Such a system has been accomplished for control and application development using ABB robots. A short description of the system is given, and its use for application/control development is exemplified. User specific feedback control can be done easily with short sampling periods and less than 1ms of latency from sensing to motor reaction.

In this paper, robotic sensor fusion of acceleration and force measurement is considered. We discuss the problem of using accelerometers close to the end-effectors of robotic manipulators and how it may improve the force control performance. We introduce a new model-based observer approach to sensor fusion of information from various different sensors. During contact transition, accelerometers and force sensors play a very important role and it can overcome many of the difficulties of uncertain models and unknown environments, which limit the domain of application of currents robots used without external sensory feedback. A model of the robot-grinding tool using the new sensors was obtained by system identification. An impedance control scheme was proposed to verify the improvement. The experiments were carried out on an ABB industrial robot with open control system architecture.

This paper presents theoretical and experimentalresults on observer-based impedance control for forcecontrol without velocity feedback. As the velocity may notavailable to measurement, which is often the case for industrialrobots, an observer was designed to reconstruct velocityin such a way that it be useful for stabilizing feedbackcontrol and to modification of the damping in the impedancerelationship. A good model of the robot joint used wasobtained by system identification. Experiments were carriedout on an ABB industrial robot 2000 to demonstrate resultson observer-based SPR feedback applied in the design.Stability was shown via a modified Popov criterion.

In attempt to have a flexible manufacturing environment with adaptive control systems, several researchers tend to integrate vision systems into the industrial systems to achieve that objective; the majority of the research works that have been done in this perspective tend to keep this kind of integration fully automated. Along the same direction, this paper presents an approach for integrating Robotic Sketching and Path Planning with Image Processing, and it introduces at the same time the ability for operators to remotely control and monitor the processing stages. It has been accomplished using network based architecture consisting of a vision system together with a server, a client and an ABB industrial robot. The aims of this paper are: (1) to provide an example that illustrates the benefit of interfacing the image processing techniques with the industrial shop floor system, and (2) to develop a web based system that allows an operator to remotely monitor and control the stages of the application using Java based applet. Further analyses have been done for the processing times of the application; this helped us to address the problem that consumed the largest proportion of the processing time. The work that has been done in this research provides additional supervision ability for the integrated system, and it demonstrates some of the challenges and the obstacles that may face this kind of integration., QCR 20170426

In attempt to have a flexible manufacturing environment with adaptive control systems, several researchers tend to integrate vision systems into the industrial systems to achieve that objective; the majority of the research works that have been done in this perspective tend to keep this kind of integration fully automated. Along the same direction, this paper presents an approach for integrating Robotic Sketching and Path Planning with Image Processing, and it introduces at the same time the ability for operators to remotely control and monitor the processing stages. It has been accomplished using network based architecture consisting of a vision system together with a server, a client and an ABB industrial robot. The aims of this paper are: (1) to provide an example that illustrates the benefit of interfacing the image processing techniques with the industrial shop floor system, and (2) to develop a web based system that allows an operator to remotely monitor and control the stages of the application using Java based applet. Further analyses have been done for the processing times of the application; this helped us to address the problem that consumed the largest proportion of the processing time. The work that has been done in this research provides additional supervision ability for the integrated system, and it demonstrates some of the challenges and the obstacles that may face this kind of integration.

When an autonomous vehicle has a mission to perform, it is of high importance that the robot has good knowledge about its position. Without good knowledge of the position, it will not be able to navigate properly and the data that it gathers, which could be of importance for the mission, might not be usable. A helicopter could for example be used to collect laser data from the terrain beneath it, which could be used to produce a 3D map of the terrain. If the knowledge about the position and orientation of the helicopter is poor then the collected laser data will be useless since it is unknown what the laser actually measures. A successful solution to position and orientation (pose) estimation of an autonomous helicopter, using an inertial measurement unit (IMU), a camera and a GPS, is proposed in this thesis. The problem is to estimate the unknown pose using sensors that measure different physical attributes and give readings containing noise. An extended Kalman filter solution to the simultaneous localisation and mapping (SLAM) is used to fuse data from the different sensors and estimate the pose of the robot. The scale invariant feature transform (SIFT) is used for feature extraction and the unified inverse depth parametrisation (UIDP) model is used to parametrise the landmarks. The orientation of the robot is described by quaternions. To be able to evaluate the performance of the filter an ABB industrial robot has been used as reference. The pose of the end tool of the robot is known with high accuracy and gives a good ground truth so that the estimations can be evaluated. The results shows that the algorithm performs well and that the pose is estimated with good accuracy., När en autonom farkost skall utföra ett uppdrag är det av högsta vikt att den har god kännedom av sin position. Utan detta kommer den inte att kunna navigera och den data som den samlar in, relevant för uppdraget, kan vara oanvändbar. Till exempel skulle en helikopter kunna användas för att samla in laser data av terrängen under den, för att skapa en 3D karta av terrängen. Om kännedomen av helikopterns position och orientering är dålig kommer de insamlade lasermätningarna att vara oanvändbara eftersom det inte är känt vad lasern faktiskt mäter. I detta examensarbete presenteras en väl fungerande lösning för position och orienterings estimering av autonom helikopter med hjälp av en inertial measurement unit (IMU), en kamera och GPS. Problemet är att skatta positionen och orienteringen med hjälp av sensorer som mäter olika fysiska storheter och vars mätningar innehåller brus. En extended Kalman filter (EKF) lösning för simultaneous localisation and mapping (SLAM) problemet används för att fusionera data från de olika sensorerna och estimera positionen och orienteringen. För feature extrahering används scale invariant feature transform (SIFT) och för att parametrisera landmärken används unified inverse depth parametrisation (UIDP). Orienteringen av roboten beskrivs med hjälp av qvartinjoner. För att evaluera skattningarna har en ABB robot används som referens vid datainsamling. Då roboten har god kännedom om position och orientering av sitt främre verktyg gör detta att prestandan i filtret kan undersökas. Resultaten visar att algorithmen fungerar bra och att skattningar har hög noggrannhet.

When an autonomous vehicle has a mission to perform, it is of high importance that the robot has good knowledge about its position. Without good knowledge of the position, it will not be able to navigate properly and the data that it gathers, which could be of importance for the mission, might not be usable. A helicopter could for example be used to collect laser data from the terrain beneath it, which could be used to produce a 3D map of the terrain. If the knowledge about the position and orientation of the helicopter is poor then the collected laser data will be useless since it is unknown what the laser actually measures. A successful solution to position and orientation (pose) estimation of an autonomous helicopter, using an inertial measurement unit (IMU), a camera and a GPS, is proposed in this thesis. The problem is to estimate the unknown pose using sensors that measure different physical attributes and give readings containing noise. An extended Kalman filter solution to the simultaneous localisation and mapping (SLAM) is used to fuse data from the different sensors and estimate the pose of the robot. The scale invariant feature transform (SIFT) is used for feature extraction and the unified inverse depth parametrisation (UIDP) model is used to parametrise the landmarks. The orientation of the robot is described by quaternions. To be able to evaluate the performance of the filter an ABB industrial robot has been used as reference. The pose of the end tool of the robot is known with high accuracy and gives a good ground truth so that the estimations can be evaluated. The results shows that the algorithm performs well and that the pose is estimated with good accuracy., När en autonom farkost skall utföra ett uppdrag är det av högsta vikt att den har god kännedom av sin position. Utan detta kommer den inte att kunna navigera och den data som den samlar in, relevant för uppdraget, kan vara oanvändbar. Till exempel skulle en helikopter kunna användas för att samla in laser data av terrängen under den, för att skapa en 3D karta av terrängen. Om kännedomen av helikopterns position och orientering är dålig kommer de insamlade lasermätningarna att vara oanvändbara eftersom det inte är känt vad lasern faktiskt mäter. I detta examensarbete presenteras en väl fungerande lösning för position och orienterings estimering av autonom helikopter med hjälp av en inertial measurement unit (IMU), en kamera och GPS. Problemet är att skatta positionen och orienteringen med hjälp av sensorer som mäter olika fysiska storheter och vars mätningar innehåller brus. En extended Kalman filter (EKF) lösning för simultaneous localisation and mapping (SLAM) problemet används för att fusionera data från de olika sensorerna och estimera positionen och orienteringen. För feature extrahering används scale invariant feature transform (SIFT) och för att parametrisera landmärken används unified inverse depth parametrisation (UIDP). Orienteringen av roboten beskrivs med hjälp av qvartinjoner. För att evaluera skattningarna har en ABB robot används som referens vid datainsamling. Då roboten har god kännedom om position och orientering av sitt främre verktyg gör detta att prestandan i filtret kan undersökas. Resultaten visar att algorithmen fungerar bra och att skattningar har hög noggrannhet.

[ES] En aquellas operaciones robóticas en las que un robot manipulador interactúa con su entorno resulta de extremada importancia poder controlar la fuerza que aquél ejerce sobre este. Con este objetivo, se suele colocar habitualmente un sensor de fuerza en la muñeca del manipulador, cerrando de esta forma el lazo de control. En lo que se refiere a estos sensores, uno de los principales problemas que plantean es que sus medidas están influenciadas no sólo por las fuerzas de contacto, sino también por las fuerzas relacionadas con la dinámica de la herramienta del manipulador, es decir, las fuerzas de inercia. En este artículo se presenta un nuevo estimador de la fuerza de contacto consistente en la integración de la información de sensores de fuerza, posición y aceleración. Además, se describe un procedimiento de calibración automático ‘plug -and- play’ para la identificación y ajuste de los parámetros de este observador. Tanto el observador de la fuerza de contacto como el procedimiento de calibración automático han sido verificados experimentalmente en un robot industrial ABB con arquitectura software abierta., [EN] It is well known that for a robotic manipulator without sensors on the end-effector, the end-effector has to follow a path in its workspace without regard to any feedback other than its joints shaft encoder or resolvers. This fact imposes severe limitations on certain tasks where an interaction between the robot and the environment is needed. However, with the help of sensors, a robot can exhibit an ”adaptive behavior”, the robot being able to deal flexibly with changes in its environment and to execute complicated skilled tasks. In this paper, we firstly propose a new control strategy based on multisensor fusion with three different sensors—that is, encoders mounted at each joint of the robot with six degrees of freedom, a wrist force sensor and an accelerometer—whose goal is to obtain a suitable contact force estimator for the three Cartesian axes. Secondly, the corresponding self-calibration method developed for the commented contact force observer is also proposed. This procedure aims at offering a ‘plug-and-play’ solution for the integration of different kind of accelerometers with the final goal of obtaining an observer capable of estimating the contact force exerted by an industrial robotic manipulator. To validate both the behaviour of the contact force observer and the performance of the self-calibrating procedure, different experiments were carried out on an ABB industrial robot with open control system architecture.

[ES] En aquellas operaciones robóticas en las que un robot manipulador interactúa con su entorno resulta de extremada importancia poder controlar la fuerza que aquél ejerce sobre este. Con este objetivo, se suele colocar habitualmente un sensor de fuerza en la muñeca del manipulador, cerrando de esta forma el lazo de control. En lo que se refiere a estos sensores, uno de los principales problemas que plantean es que sus medidas están influenciadas no sólo por las fuerzas de contacto, sino también por las fuerzas relacionadas con la dinámica de la herramienta del manipulador, es decir, las fuerzas de inercia. En este artículo se presenta un nuevo estimador de la fuerza de contacto consistente en la integración de la información de sensores de fuerza, posición y aceleración. Además, se describe un procedimiento de calibración automático ‘plug -and- play’ para la identificación y ajuste de los parámetros de este observador. Tanto el observador de la fuerza de contacto como el procedimiento de calibración automático han sido verificados experimentalmente en un robot industrial ABB con arquitectura software abierta., [EN] It is well known that for a robotic manipulator without sensors on the end-effector, the end-effector has to follow a path in its workspace without regard to any feedback other than its joints shaft encoder or resolvers. This fact imposes severe limitations on certain tasks where an interaction between the robot and the environment is needed. However, with the help of sensors, a robot can exhibit an ”adaptive behavior”, the robot being able to deal flexibly with changes in its environment and to execute complicated skilled tasks. In this paper, we firstly propose a new control strategy based on multisensor fusion with three different sensors—that is, encoders mounted at each joint of the robot with six degrees of freedom, a wrist force sensor and an accelerometer—whose goal is to obtain a suitable contact force estimator for the three Cartesian axes. Secondly, the corresponding self-calibration method developed for the commented contact force observer is also proposed. This procedure aims at offering a ‘plug-and-play’ solution for the integration of different kind of accelerometers with the final goal of obtaining an observer capable of estimating the contact force exerted by an industrial robotic manipulator. To validate both the behaviour of the contact force observer and the performance of the self-calibrating procedure, different experiments were carried out on an ABB industrial robot with open control system architecture.

[ES] En aquellas operaciones robóticas en las que un robot manipulador interactúa con su entorno resulta de extremada importancia poder controlar la fuerza que aquél ejerce sobre este. Con este objetivo, se suele colocar habitualmente un sensor de fuerza en la muñeca del manipulador, cerrando de esta forma el lazo de control. En lo que se refiere a estos sensores, uno de los principales problemas que plantean es que sus medidas están influenciadas no sólo por las fuerzas de contacto, sino también por las fuerzas relacionadas con la dinámica de la herramienta del manipulador, es decir, las fuerzas de inercia. En este artículo se presenta un nuevo estimador de la fuerza de contacto consistente en la integración de la información de sensores de fuerza, posición y aceleración. Además, se describe un procedimiento de calibración automático ‘plug -and- play’ para la identificación y ajuste de los parámetros de este observador. Tanto el observador de la fuerza de contacto como el procedimiento de calibración automático han sido verificados experimentalmente en un robot industrial ABB con arquitectura software abierta., [EN] It is well known that for a robotic manipulator without sensors on the end-effector, the end-effector has to follow a path in its workspace without regard to any feedback other than its joints shaft encoder or resolvers. This fact imposes severe limitations on certain tasks where an interaction between the robot and the environment is needed. However, with the help of sensors, a robot can exhibit an ”adaptive behavior”, the robot being able to deal flexibly with changes in its environment and to execute complicated skilled tasks. In this paper, we firstly propose a new control strategy based on multisensor fusion with three different sensors—that is, encoders mounted at each joint of the robot with six degrees of freedom, a wrist force sensor and an accelerometer—whose goal is to obtain a suitable contact force estimator for the three Cartesian axes. Secondly, the corresponding self-calibration method developed for the commented contact force observer is also proposed. This procedure aims at offering a ‘plug-and-play’ solution for the integration of different kind of accelerometers with the final goal of obtaining an observer capable of estimating the contact force exerted by an industrial robotic manipulator. To validate both the behaviour of the contact force observer and the performance of the self-calibrating procedure, different experiments were carried out on an ABB industrial robot with open control system architecture.

[ES] En aquellas operaciones robóticas en las que un robot manipulador interactúa con su entorno resulta de extremada importancia poder controlar la fuerza que aquél ejerce sobre este. Con este objetivo, se suele colocar habitualmente un sensor de fuerza en la muñeca del manipulador, cerrando de esta forma el lazo de control. En lo que se refiere a estos sensores, uno de los principales problemas que plantean es que sus medidas están influenciadas no sólo por las fuerzas de contacto, sino también por las fuerzas relacionadas con la dinámica de la herramienta del manipulador, es decir, las fuerzas de inercia. En este artículo se presenta un nuevo estimador de la fuerza de contacto consistente en la integración de la información de sensores de fuerza, posición y aceleración. Además, se describe un procedimiento de calibración automático ‘plug -and- play’ para la identificación y ajuste de los parámetros de este observador. Tanto el observador de la fuerza de contacto como el procedimiento de calibración automático han sido verificados experimentalmente en un robot industrial ABB con arquitectura software abierta., [EN] It is well known that for a robotic manipulator without sensors on the end-effector, the end-effector has to follow a path in its workspace without regard to any feedback other than its joints shaft encoder or resolvers. This fact imposes severe limitations on certain tasks where an interaction between the robot and the environment is needed. However, with the help of sensors, a robot can exhibit an ”adaptive behavior”, the robot being able to deal flexibly with changes in its environment and to execute complicated skilled tasks. In this paper, we firstly propose a new control strategy based on multisensor fusion with three different sensors—that is, encoders mounted at each joint of the robot with six degrees of freedom, a wrist force sensor and an accelerometer—whose goal is to obtain a suitable contact force estimator for the three Cartesian axes. Secondly, the corresponding self-calibration method developed for the commented contact force observer is also proposed. This procedure aims at offering a ‘plug-and-play’ solution for the integration of different kind of accelerometers with the final goal of obtaining an observer capable of estimating the contact force exerted by an industrial robotic manipulator. To validate both the behaviour of the contact force observer and the performance of the self-calibrating procedure, different experiments were carried out on an ABB industrial robot with open control system architecture.

[ES] En aquellas operaciones robóticas en las que un robot manipulador interactúa con su entorno resulta de extremada importancia poder controlar la fuerza que aquél ejerce sobre este. Con este objetivo, se suele colocar habitualmente un sensor de fuerza en la muñeca del manipulador, cerrando de esta forma el lazo de control. En lo que se refiere a estos sensores, uno de los principales problemas que plantean es que sus medidas están influenciadas no sólo por las fuerzas de contacto, sino también por las fuerzas relacionadas con la dinámica de la herramienta del manipulador, es decir, las fuerzas de inercia. En este artículo se presenta un nuevo estimador de la fuerza de contacto consistente en la integración de la información de sensores de fuerza, posición y aceleración. Además, se describe un procedimiento de calibración automático ‘plug -and- play’ para la identificación y ajuste de los parámetros de este observador. Tanto el observador de la fuerza de contacto como el procedimiento de calibración automático han sido verificados experimentalmente en un robot industrial ABB con arquitectura software abierta., [EN] It is well known that for a robotic manipulator without sensors on the end-effector, the end-effector has to follow a path in its workspace without regard to any feedback other than its joints shaft encoder or resolvers. This fact imposes severe limitations on certain tasks where an interaction between the robot and the environment is needed. However, with the help of sensors, a robot can exhibit an ”adaptive behavior”, the robot being able to deal flexibly with changes in its environment and to execute complicated skilled tasks. In this paper, we firstly propose a new control strategy based on multisensor fusion with three different sensors—that is, encoders mounted at each joint of the robot with six degrees of freedom, a wrist force sensor and an accelerometer—whose goal is to obtain a suitable contact force estimator for the three Cartesian axes. Secondly, the corresponding self-calibration method developed for the commented contact force observer is also proposed. This procedure aims at offering a ‘plug-and-play’ solution for the integration of different kind of accelerometers with the final goal of obtaining an observer capable of estimating the contact force exerted by an industrial robotic manipulator. To validate both the behaviour of the contact force observer and the performance of the self-calibrating procedure, different experiments were carried out on an ABB industrial robot with open control system architecture.

Introduction It is important to understand the mechanics of the lumbar spine, as it has been shown that much low back pain is attributable to mechanical factors. One important aspect of spinal mechanics is the neutral zone, defined as a region of little or no resistance to motion on either side of the neutral position for a motion segment. If the neutral zone is a significant feature of intervertebral joint mechanics then the spinal joints will have little intrinsic stability and rely on muscles to control their movement around the neutral position. This has significant implications for our understanding of how degenerative changes to the spinal joints might destabilise the spine. This study was performed to characterise the size of the neutral zone and the effect of axial preload for different spinal motions. Methods Using a 6 degree-of-freedom (DOF) ABB industrial robot incorporating a 6-DOF JR3 force sensor, six isolated ovine lumbar joint segments were subjected to 5 repetitive movements in 3 directions (6° extension / 15° flexion, +/- 7° lateral bend, +/- 3° axial twist) with 4 different preloads (0, 150, 300, 450N) under 2 conditions (facet joints intact and facets removed). For each direction, the fixed axis about which the joint would rotate with a minimal motion-opposing moment was determined in advance. In accordance with a previous study by this group, the neutral zone was defined as the region where absolute rotational stiffness is less than 0.05 Nm/°. Results When moving from 6° of extension to -15° (flexion) a neutral zone was usually observed starting around 0° and continuing as far as -8 or -9°. The neutral zone was in the same region when moving in the opposite direction, except when the specimen showed a considerable amount of hysteresis, in which case the neutral zone could start as early as -11° or -12° and usually continued to -2°or -3°. Increasing preload usually made the joint stiffer in the regions outside the neutral zone, but did not affect

In this masters thesis we consider a method for experimental identification of the inertial parameters of an industrial robot, using measured torques and joint angles. A dynamic model of the first three joints of the robot has been identified. To achieve good identification results, it is important to carefully choose the trajectory for the experimental identification. A method to generate trajectories using two suggested design criteria has been used and evaluated using an ABB industrial robot, and one of them yields good identification results., Denna rapport behandlar en metod för experimentell identifiering av en stelkroppsmodell för en industrirobot. Metoden använder sig av uppmätta moment och armvinklar för att identifiera parametrarna i en dynamikmodell för robotens tre huvudaxlar. För att erhålla bra identifieringsresultat är det viktigt att välja en lämplig identifieringsbana och i detta arbete har en metod för generering av banor använts och utvärderats för två olika designkriterier. Experiment har utförts på en industrirobot från ABB och vi har erhållit goda identifieringsresultat med ett av designkriterierna.

In this masters thesis we consider a method for experimental identification of the inertial parameters of an industrial robot, using measured torques and joint angles. A dynamic model of the first three joints of the robot has been identified. To achieve good identification results, it is important to carefully choose the trajectory for the experimental identification. A method to generate trajectories using two suggested design criteria has been used and evaluated using an ABB industrial robot, and one of them yields good identification results., Denna rapport behandlar en metod för experimentell identifiering av en stelkroppsmodell för en industrirobot. Metoden använder sig av uppmätta moment och armvinklar för att identifiera parametrarna i en dynamikmodell för robotens tre huvudaxlar. För att erhålla bra identifieringsresultat är det viktigt att välja en lämplig identifieringsbana och i detta arbete har en metod för generering av banor använts och utvärderats för två olika designkriterier. Experiment har utförts på en industrirobot från ABB och vi har erhållit goda identifieringsresultat med ett av designkriterierna.

In this work, we present implementation and experiment of the theory of dynamic force sensing for robotic manipulators, which uses a sensor fusion technique in order to extract the contact force exerted by the end-effector of the manipulator from those measured by a wrist force sensor, which are corrupted by the inertial forces on the end-effector. We propose a new control strategy based on multisensor fusion with three different sensors that is, encoders mounted at each joint of the robot with six degrees of freedom, a wrist force sensor and an accelerometer whose goal is to obtain a suitable contact force estimator for the three Cartesian axes. This new observer contributes to overcome many of the difficulties of uncertain world models and unknown environments, which limit the domain of application of currents robots used without external sensory fedback. An impedance control scheme was proposed to verify the improvement. The experiments were carried out on an ABB industrial robot with open control system architecture.

In robotic operations where a manipulator is involved, it is well-known that the quantities measured by a wrist force/torque sensor are corrupted by the dynamicsof the end effector and manipulator. To solve this problem,an observer, which fuses information from force sensors andaccelerometers, was designed recently in order to estimate the contact force exerted by a manipulator to its environment [1]. This paper introduced a high-speed, high-accuracy, versatile, simple, and fully autonomous technique for the calibration of this robotic manipulator 3D force observer by means of active motion. To verify the improvement, an impedance control scheme was used. A dynamic model of the robot-grinding tool using the new sensors was obtained by system identification. The experiments were carried out on an ABB industrial robot with open control system architecture.

In this paper, we propose a method for selfcalibration of a robotic manipulator force observer, which fuses information from force sensors and accelerometers in order to estimate the contact force exerted by a manipulator to its environment, by means of active motion. In robotic operation, during contact transition accelerometers and force sensors play a very important role and serve to overcome many of the difficulties of uncertain world models and unknown environments, limiting the domain of application of current robots used without external sensory provided. The calibration procedure helps to improve the performance as well as enhanced stability and robustness for the transition phase. A variety of accelerometers were used to validate the procedure. A dynamic model of the robot-grinding tool using the new sensors was obtained by system identification. An impedance control scheme was proposed to verify the improvement. The experiments were carried out on an ABB industrial robot with open control system architecture.

One way to increase the flexibility of industrial robots in manipulation tasks is to integrate additional sensors in the control systems. Cameras are an example of such sensors, and in recent years there has been an increased interest in vision based control. However, it is clear that most manipulation tasks can not be solved using position control alone, because of the risk of excessive contact forces. Therefore, it would be interesting to combine vision based position control with force feedback. In this thesis, we present a method for combining direct force control and visual servoing in the presence of unknown planar surfaces. The control algorithm involves a force feedback control loop and a vision based reference trajectory as a feed-forward signal. The vision system is based on a constrained image-based visual servoing algorithm, using an explicit 3D-reconstruction of the planar constraint surface. We show how calibration data calculated by a simple but efficient camera calibration method can be used in combination with force and position data to improve the reconstruction and reference trajectories. The task chosen involves force controlled drawing on an unknown surface. The robot will grasp a pen using visual servoing, and use the pen to draw lines between a number of points on a whiteboard. The force control will keep the contact force constant during the drawing. The method is validated through experiments carried out on a 6-degree-of-freedom ABB Industrial Robot 2000.

In situations where robots are performing manipulation tasks, e.g. welding, grinding, and drilling, it is necessary to avoid excessive forces in the contact with the environment. Because of geometric uncertainty and finite positioning accuracy this usually cannot be solved by position control alone. The subject of this thesis is robot force control, with particular interest devoted to impedance control. Two other force control strategies are also treated to some extent. These are compliance control and direct force control. Impedance control is a technique aimed at controlling the dynamic relation between position error and force error in interaction similar to Newton's second law of motion. The velocity is used to modify the damping in the impedance relation. A problem, however, is that the velocity is not available for measurement. For this reason an observer is designed to reconstruct the velocity signal. A good model of the robot joint used is obtained by system identification. Another important issue is stability, and two different approaches are used for the stability analysis. The first one uses theory for piecewise linear control systems, motivated by the fact that the contact force non-linearity can be modeled as piecewise linear. The other approach is based on results on observer-based SPR feedback and stability is proved by the Popov criterion. Experiments are carried out on an ABB industrial robot 2000 to verify the theoretical results.