R. Mahony | The Australian National University (original) (raw)
Papers by R. Mahony
2013 European Control Conference (ECC), 2013
In this paper we apply port-Hamiltonian theory with the bondgraph modelling approach to the probl... more In this paper we apply port-Hamiltonian theory with the bondgraph modelling approach to the problem of formation control using partial measurements of relative positions. We present a control design that drives a group of vehicles to a desired formation without requiring inter-vehicle communications or global position and velocity measurements to be available. Our generic approach is applicable to any form of relative measurement between vehicles, but we specifically consider the important cases of relative bearings and relative distances. In the case of bearings, our theory closely relates to the field of image-based visual servo (IBVS) control. We present simulation results to support the developed theory.
2007 European Control Conference (ECC)
This paper considers the question of fully image based visual servo control for a Vertical Takeof... more This paper considers the question of fully image based visual servo control for a Vertical Takeoff and landing (VTOL) Unmanned Aerial Vehicle. A new visual feature based on the combination of the first order un-normalized spherical moment and the image length of the target image constellation is proposed. To avoid the lack of the translational velocity measurement of the camera, we propose a nonlinear observer-based visual servo controller. Compared with earlier works, the observer-based visual servo controller allows to simplify measurement through observation techniques and provides a framework to implement algorithms to stabilize an UAV in front of an unknown target. The corresponding control algorithm is presented and simulation results are finally shown to illustrate performances of the proposed approach.
Proceedings of the 41st IEEE Conference on Decision and Control, 2002.
ABSTRACT
IEEE/RSJ International Conference on Intelligent Robots and System
This paper is concerned with choosing image features for image based visual servo control and how... more This paper is concerned with choosing image features for image based visual servo control and how this choice influences the closed-loop dynamics of the system. In prior work, image features tend to be chosen on the basis of image processing simplicity and noise sensitivity. In this paper we show that the choice of feature directly influences the closed-loop dynamics in task-space. We focus on the depth axis control of a visual servo system and compare analytically various approaches that have been reported recently in the literature. The theoretical predictions are verified by experiment.
A robust and reliable attitude estimator is a key technology enabler for the development of auton... more A robust and reliable attitude estimator is a key technology enabler for the development of autonomous aerial vehicles. This paper is an introduction to attitude estimation for aerial robotic systems. First, attitude definition and parameterizations are recalled and discussed. Then, several attitude estimation techniques – ranging from algebraic vector observation-based attitude determination algorithms to dynamic attitude filtering and estimation methodologies – are presented and commented upon in relation to practical implementation issues. Particular attention is devoted to the applications of a well-known nonlinear attitude observer (called explicit complementary filter in the literature) to aerial robotics, using a low-cost and light-weight inertial measurement unit, which can be complemented with a GPS or airspeed sensors.
2009 IEEE International Conference on Robotics and Automation, 2009
In the last decade, many vision-based robot controllers have been designed using Cartesian inform... more In the last decade, many vision-based robot controllers have been designed using Cartesian information encoded in the homography transformation that links two images of a planar object. For any approach, the performance of the closed-loop system depends on the quality of the homography estimates obtained. In this paper, we exploit the special linear Lie-group structure of the set of all homographies to develop a dynamic observer to estimate homographies online. The resulting estimates are effective and can be used to improve closed-loop response of several visual servoing algorithms.
![Research paper thumbnail of Attitude estimation on SO[3] based on direct inertial measurements](https://attachments.academia-assets.com/98073061/thumbnails/1.jpg)
](Proceedings 2006 IEEE International Conference on Robotics and Automation, 2006. ICRA 2006.
This paper considers the question of obtaining high quality attitude estimates from typical low c... more This paper considers the question of obtaining high quality attitude estimates from typical low cost inertial measurement units for applications in control of unmanned aerial vehicles. A nonlinear complementary filter exploiting the structure of Special Orthogonal Group S0(3) is proposed. The filter is expressed explicitly in terms of direct and untreated measurements. For a typical low cost inertial measurement where two inertial direction are measured (gravity and magnetic field), it is shown that the filter is well conditioned. If only a single direction is available (typically the gravity) associated with gyros measurements, it is shown that the full gyros bias vector is correctly estimated and that the estimated orientation converges to a set consistent with the measurements. Experimental results, conducted on a the HoverEye c UAV, demonstrate the efficiency of the proposed filter.
European Journal of Control, 2001
Control of a scale model autonomous helicopter during takeoff and landing manoeuvres has proved t... more Control of a scale model autonomous helicopter during takeoff and landing manoeuvres has proved to be an extremely difficult problem. This is a consequence of the slowly time-varying and environment dependent nature of the aerodynamic forces encountered along with the high sensitivity of the helicopter to collective pitch changes during these manoeuvres. In this paper we propose a novel approach to the control problem for such manoeuvres. The proposed control design uses the motor torque rather than collective pitch as the principal control input and takes advantage of its reduced sensitivity to aerodynamic effects and structural properties to develop a parametric adaptive control algorithm that estimates the principal aerodynamic effects on-line.
Transactions of the Institute of Measurement and Control, 2010
In this paper we propose a simple non-linear observer for attitude estimation based only on outpu... more In this paper we propose a simple non-linear observer for attitude estimation based only on output from a typical inertial measurement unit (IMU) and dynamic pressure sensor embarked on a low-cost unmanned aerial vehicle. In particular, we aim to provide a good quality attitude estimate in the absence of global positioning system (GPS) ground truth and with potential low-frequency bias and high-frequency noise in the IMU sensor measurements. In addition, the case where the IMU only provides gyrometer and accelerometer outputs is considered; that is, there is no magnetometer output or it cannot be used due to local magnetic disturbances such as are common on a vehicle with electric motors. The proposed observer uses a simple centripetal force model (based on gyrometer and dynamic pressure measurements), augmented by a first-order dynamic model for angle of attack, to estimate non-inertial components of the acceleration. This estimate is used to correct the accelerometer output to pro...
Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference, 2009
This paper considers the question of designing an attitude observer exploiting the structure of t... more This paper considers the question of designing an attitude observer exploiting the structure of the Special Orthogonal Group SO(3) for both inertial and body-fixedframe measurements. We consider measurements from a minimal sensor suite, typically a rate gyroscope along with several measurements of inertial and/or body-fixed vector directions. We propose fully nonlinear pose observers based directly on the vector measurements, allowing a mix of both inertial and body-fixed measurements. We provide a comprehensive stability analysis of the observer error dynamics. We show that even with a single vector measurement it is often possible to recover asymptotically stable observer error dynamics.
Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference, 2009
This paper describes a new visual-servo control algorithm to stabilize an aircraft along a circul... more This paper describes a new visual-servo control algorithm to stabilize an aircraft along a circular trajectory using a single visual landmark on textured flat ground. The control algorithm requires a sensor suite consisting of inertial measurement unit, along with a single embedded camera capable of detecting and tracking a visual feature and estimating optic flow. The proposed control algorithm does
Proceedings 2001 ICRA. IEEE International Conference on Robotics and Automation (Cat. No.01CH37164), 2001
A control design to stabilise a reduced scale autonomous helicopter equipped with a camera is pre... more A control design to stabilise a reduced scale autonomous helicopter equipped with a camera is presented. The proposed algorithm is motivated by recent work in image-based visual servoing control for under-actuated dynamic systems. This work is extended by considering models that contain weakly nonminimum phase zero dynamics. This is an important class of systems since an offset between the camera
2015 American Control Conference (ACC), 2015
This paper proposes an attitude estimation methodology for the case where attitude sensors provid... more This paper proposes an attitude estimation methodology for the case where attitude sensors provide discrete-time samples of vector measurements at different sample rates and with time delays. The proposed methodology is based on a cascade combination of an output predictor and an attitude observer or filter. The predictor compensates for the effect of sampling and delays in vector measurements and provides continuous-time predictions of outputs. These predictions are then used in an observer or filter to estimate the current attitude. The primary contribution of the paper is to exploit the underlying symmetry of the attitude kinematics to design a recursive predictor that is computationally simple and generic, in the sense that it can be combined with any asymptotically stable observer or filter. We prove that the predictor is able to reproduce the continuous time delay-free vector measurements. In a simulation example, we demonstrate good performance of the combined predictor-observer even in presence of measurement noise and delay uncertainties.
2007 46th IEEE Conference on Decision and Control, 2007
An image-based visual servo (IBVS) control strategy is proposed for a class of under actuated sys... more An image-based visual servo (IBVS) control strategy is proposed for a class of under actuated systems. The class of systems considered includes dynamic models of vehicles capable of quasi-stationary flight. The paper extends earlier works that stabilises an aerial vehicle using only direct visual and inertial measurements, and in particular, without a direct measurement of linear velocity. The approach taken
IFAC Proceedings Volumes, 2008
An Image-Based Visual Servo (IBVS) control strategy for stabilisation of Vertical TakeOff and Lan... more An Image-Based Visual Servo (IBVS) control strategy for stabilisation of Vertical TakeOff and Landing (VTOL) vehicles with respect to a fixed target is proposed. A visual feature based on the combination of the first order un-normalized spherical moment and the image length of the target image is considered. To avoid the lack of the measurement of the translational velocity of the camera and the thrust of the vehicle, a nonlinear observer-based visual servo controller is proposed. Local asymptotic stability of the system is proved and an estimate of the basin of attraction for the closed-loop system is provided. Simulation results are finally presented to illustrate the performance of the control algorithm.
IFAC Proceedings Volumes, 2013
This paper considers the design of nonlinear state observers for finite-dimensional equivariant k... more This paper considers the design of nonlinear state observers for finite-dimensional equivariant kinematics of mechanical systems. The observer design problem is approached by lifting the system kinematics onto the symmetry group and designing an observer for the lifted system. Two particular classes of lifted systems are identified, which we term type I and type II systems, that correspond to common configurations of sensor suites for mobile robotics applications. We consider type I systems in detail and define an error signal on the symmetry group using the group structure. We propose an observer structure with a pre-observer or internal model augmented by an equivariant innovation term that leads to autonomous error evolution. A control Lyapunov function construction is used to design the observer innovation that both ensures the required equivariance, and leads to strong convergence properties of the observer error dynamics.
Proceedings of the 45th IEEE Conference on Decision and Control, 2006
... This filter is a key aspect of the practical implementation. ... III. FILTER DESIGN For the e... more ... This filter is a key aspect of the practical implementation. ... III. FILTER DESIGN For the experimental system considered in this paper it is straightforward to derive a good quality ... indicate that the measured signal for W must be filtered to obtain a stable closed-loop control system. ...
International Journal of Optomechatronics, 2008
This article considers the question of dynamic image-based visual servo control for a vertical ta... more This article considers the question of dynamic image-based visual servo control for a vertical takeoff and landing unmanned aerial vehicle. The visual targets considered are colored blobs on a flat surface to which the normal direction is known. A fully nonlinear adaptive control design coupled with an asymptotic filter is provided that ensures the stability of the closed-loop system in a large domain. The image features used are a modified first order spherical moment for position regulation and optic flow for velocity damping.
2009 IEEE International Conference on Robotics and Automation, 2009
... A NON-LINEAR CONTROLLER FOR TERRAIN FOLLOWING In this section a control design ensuring terra... more ... A NON-LINEAR CONTROLLER FOR TERRAIN FOLLOWING In this section a control design ensuring terrain ... for all initial conditions such that d(0) = d0 > 0, the closed-loop trajectory exists ... 5). A Pyramidal implementation of the Lucas-Kanade [21] algorithm is used to compute ...
2013 European Control Conference (ECC), 2013
In this paper we apply port-Hamiltonian theory with the bondgraph modelling approach to the probl... more In this paper we apply port-Hamiltonian theory with the bondgraph modelling approach to the problem of formation control using partial measurements of relative positions. We present a control design that drives a group of vehicles to a desired formation without requiring inter-vehicle communications or global position and velocity measurements to be available. Our generic approach is applicable to any form of relative measurement between vehicles, but we specifically consider the important cases of relative bearings and relative distances. In the case of bearings, our theory closely relates to the field of image-based visual servo (IBVS) control. We present simulation results to support the developed theory.
2007 European Control Conference (ECC)
This paper considers the question of fully image based visual servo control for a Vertical Takeof... more This paper considers the question of fully image based visual servo control for a Vertical Takeoff and landing (VTOL) Unmanned Aerial Vehicle. A new visual feature based on the combination of the first order un-normalized spherical moment and the image length of the target image constellation is proposed. To avoid the lack of the translational velocity measurement of the camera, we propose a nonlinear observer-based visual servo controller. Compared with earlier works, the observer-based visual servo controller allows to simplify measurement through observation techniques and provides a framework to implement algorithms to stabilize an UAV in front of an unknown target. The corresponding control algorithm is presented and simulation results are finally shown to illustrate performances of the proposed approach.
Proceedings of the 41st IEEE Conference on Decision and Control, 2002.
ABSTRACT
IEEE/RSJ International Conference on Intelligent Robots and System
This paper is concerned with choosing image features for image based visual servo control and how... more This paper is concerned with choosing image features for image based visual servo control and how this choice influences the closed-loop dynamics of the system. In prior work, image features tend to be chosen on the basis of image processing simplicity and noise sensitivity. In this paper we show that the choice of feature directly influences the closed-loop dynamics in task-space. We focus on the depth axis control of a visual servo system and compare analytically various approaches that have been reported recently in the literature. The theoretical predictions are verified by experiment.
A robust and reliable attitude estimator is a key technology enabler for the development of auton... more A robust and reliable attitude estimator is a key technology enabler for the development of autonomous aerial vehicles. This paper is an introduction to attitude estimation for aerial robotic systems. First, attitude definition and parameterizations are recalled and discussed. Then, several attitude estimation techniques – ranging from algebraic vector observation-based attitude determination algorithms to dynamic attitude filtering and estimation methodologies – are presented and commented upon in relation to practical implementation issues. Particular attention is devoted to the applications of a well-known nonlinear attitude observer (called explicit complementary filter in the literature) to aerial robotics, using a low-cost and light-weight inertial measurement unit, which can be complemented with a GPS or airspeed sensors.
2009 IEEE International Conference on Robotics and Automation, 2009
In the last decade, many vision-based robot controllers have been designed using Cartesian inform... more In the last decade, many vision-based robot controllers have been designed using Cartesian information encoded in the homography transformation that links two images of a planar object. For any approach, the performance of the closed-loop system depends on the quality of the homography estimates obtained. In this paper, we exploit the special linear Lie-group structure of the set of all homographies to develop a dynamic observer to estimate homographies online. The resulting estimates are effective and can be used to improve closed-loop response of several visual servoing algorithms.
Proceedings 2006 IEEE International Conference on Robotics and Automation, 2006. ICRA 2006.
This paper considers the question of obtaining high quality attitude estimates from typical low c... more This paper considers the question of obtaining high quality attitude estimates from typical low cost inertial measurement units for applications in control of unmanned aerial vehicles. A nonlinear complementary filter exploiting the structure of Special Orthogonal Group S0(3) is proposed. The filter is expressed explicitly in terms of direct and untreated measurements. For a typical low cost inertial measurement where two inertial direction are measured (gravity and magnetic field), it is shown that the filter is well conditioned. If only a single direction is available (typically the gravity) associated with gyros measurements, it is shown that the full gyros bias vector is correctly estimated and that the estimated orientation converges to a set consistent with the measurements. Experimental results, conducted on a the HoverEye c UAV, demonstrate the efficiency of the proposed filter.
European Journal of Control, 2001
Control of a scale model autonomous helicopter during takeoff and landing manoeuvres has proved t... more Control of a scale model autonomous helicopter during takeoff and landing manoeuvres has proved to be an extremely difficult problem. This is a consequence of the slowly time-varying and environment dependent nature of the aerodynamic forces encountered along with the high sensitivity of the helicopter to collective pitch changes during these manoeuvres. In this paper we propose a novel approach to the control problem for such manoeuvres. The proposed control design uses the motor torque rather than collective pitch as the principal control input and takes advantage of its reduced sensitivity to aerodynamic effects and structural properties to develop a parametric adaptive control algorithm that estimates the principal aerodynamic effects on-line.
Transactions of the Institute of Measurement and Control, 2010
In this paper we propose a simple non-linear observer for attitude estimation based only on outpu... more In this paper we propose a simple non-linear observer for attitude estimation based only on output from a typical inertial measurement unit (IMU) and dynamic pressure sensor embarked on a low-cost unmanned aerial vehicle. In particular, we aim to provide a good quality attitude estimate in the absence of global positioning system (GPS) ground truth and with potential low-frequency bias and high-frequency noise in the IMU sensor measurements. In addition, the case where the IMU only provides gyrometer and accelerometer outputs is considered; that is, there is no magnetometer output or it cannot be used due to local magnetic disturbances such as are common on a vehicle with electric motors. The proposed observer uses a simple centripetal force model (based on gyrometer and dynamic pressure measurements), augmented by a first-order dynamic model for angle of attack, to estimate non-inertial components of the acceleration. This estimate is used to correct the accelerometer output to pro...
Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference, 2009
This paper considers the question of designing an attitude observer exploiting the structure of t... more This paper considers the question of designing an attitude observer exploiting the structure of the Special Orthogonal Group SO(3) for both inertial and body-fixedframe measurements. We consider measurements from a minimal sensor suite, typically a rate gyroscope along with several measurements of inertial and/or body-fixed vector directions. We propose fully nonlinear pose observers based directly on the vector measurements, allowing a mix of both inertial and body-fixed measurements. We provide a comprehensive stability analysis of the observer error dynamics. We show that even with a single vector measurement it is often possible to recover asymptotically stable observer error dynamics.
Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference, 2009
This paper describes a new visual-servo control algorithm to stabilize an aircraft along a circul... more This paper describes a new visual-servo control algorithm to stabilize an aircraft along a circular trajectory using a single visual landmark on textured flat ground. The control algorithm requires a sensor suite consisting of inertial measurement unit, along with a single embedded camera capable of detecting and tracking a visual feature and estimating optic flow. The proposed control algorithm does
Proceedings 2001 ICRA. IEEE International Conference on Robotics and Automation (Cat. No.01CH37164), 2001
A control design to stabilise a reduced scale autonomous helicopter equipped with a camera is pre... more A control design to stabilise a reduced scale autonomous helicopter equipped with a camera is presented. The proposed algorithm is motivated by recent work in image-based visual servoing control for under-actuated dynamic systems. This work is extended by considering models that contain weakly nonminimum phase zero dynamics. This is an important class of systems since an offset between the camera
2015 American Control Conference (ACC), 2015
This paper proposes an attitude estimation methodology for the case where attitude sensors provid... more This paper proposes an attitude estimation methodology for the case where attitude sensors provide discrete-time samples of vector measurements at different sample rates and with time delays. The proposed methodology is based on a cascade combination of an output predictor and an attitude observer or filter. The predictor compensates for the effect of sampling and delays in vector measurements and provides continuous-time predictions of outputs. These predictions are then used in an observer or filter to estimate the current attitude. The primary contribution of the paper is to exploit the underlying symmetry of the attitude kinematics to design a recursive predictor that is computationally simple and generic, in the sense that it can be combined with any asymptotically stable observer or filter. We prove that the predictor is able to reproduce the continuous time delay-free vector measurements. In a simulation example, we demonstrate good performance of the combined predictor-observer even in presence of measurement noise and delay uncertainties.
2007 46th IEEE Conference on Decision and Control, 2007
An image-based visual servo (IBVS) control strategy is proposed for a class of under actuated sys... more An image-based visual servo (IBVS) control strategy is proposed for a class of under actuated systems. The class of systems considered includes dynamic models of vehicles capable of quasi-stationary flight. The paper extends earlier works that stabilises an aerial vehicle using only direct visual and inertial measurements, and in particular, without a direct measurement of linear velocity. The approach taken
IFAC Proceedings Volumes, 2008
An Image-Based Visual Servo (IBVS) control strategy for stabilisation of Vertical TakeOff and Lan... more An Image-Based Visual Servo (IBVS) control strategy for stabilisation of Vertical TakeOff and Landing (VTOL) vehicles with respect to a fixed target is proposed. A visual feature based on the combination of the first order un-normalized spherical moment and the image length of the target image is considered. To avoid the lack of the measurement of the translational velocity of the camera and the thrust of the vehicle, a nonlinear observer-based visual servo controller is proposed. Local asymptotic stability of the system is proved and an estimate of the basin of attraction for the closed-loop system is provided. Simulation results are finally presented to illustrate the performance of the control algorithm.
IFAC Proceedings Volumes, 2013
This paper considers the design of nonlinear state observers for finite-dimensional equivariant k... more This paper considers the design of nonlinear state observers for finite-dimensional equivariant kinematics of mechanical systems. The observer design problem is approached by lifting the system kinematics onto the symmetry group and designing an observer for the lifted system. Two particular classes of lifted systems are identified, which we term type I and type II systems, that correspond to common configurations of sensor suites for mobile robotics applications. We consider type I systems in detail and define an error signal on the symmetry group using the group structure. We propose an observer structure with a pre-observer or internal model augmented by an equivariant innovation term that leads to autonomous error evolution. A control Lyapunov function construction is used to design the observer innovation that both ensures the required equivariance, and leads to strong convergence properties of the observer error dynamics.
Proceedings of the 45th IEEE Conference on Decision and Control, 2006
... This filter is a key aspect of the practical implementation. ... III. FILTER DESIGN For the e... more ... This filter is a key aspect of the practical implementation. ... III. FILTER DESIGN For the experimental system considered in this paper it is straightforward to derive a good quality ... indicate that the measured signal for W must be filtered to obtain a stable closed-loop control system. ...
International Journal of Optomechatronics, 2008
This article considers the question of dynamic image-based visual servo control for a vertical ta... more This article considers the question of dynamic image-based visual servo control for a vertical takeoff and landing unmanned aerial vehicle. The visual targets considered are colored blobs on a flat surface to which the normal direction is known. A fully nonlinear adaptive control design coupled with an asymptotic filter is provided that ensures the stability of the closed-loop system in a large domain. The image features used are a modified first order spherical moment for position regulation and optic flow for velocity damping.
2009 IEEE International Conference on Robotics and Automation, 2009
... A NON-LINEAR CONTROLLER FOR TERRAIN FOLLOWING In this section a control design ensuring terra... more ... A NON-LINEAR CONTROLLER FOR TERRAIN FOLLOWING In this section a control design ensuring terrain ... for all initial conditions such that d(0) = d0 > 0, the closed-loop trajectory exists ... 5). A Pyramidal implementation of the Lucas-Kanade [21] algorithm is used to compute ...