Pietro Maria Muraca | Università della Calabria (original) (raw)
Papers by Pietro Maria Muraca
IEEE Transactions on Automatic Control, 1994
Lai and H. Robbins, "Consistency and asymptotic efficiency of slope estimates in stochastic schem... more Lai and H. Robbins, "Consistency and asymptotic efficiency of slope estimates in stochastic schemes." Z Wahrscheinlichkeirs theorie y e w .
International Journal of Modelling and Simulation, 1992
This paper considers the problem of locating a sensor for the state estimation of a linear stocha... more This paper considers the problem of locating a sensor for the state estimation of a linear stochastic distributed parameter system of parabolic type, with observations taken at discrete times 1 1 ,12' .... Assuming that k measurements have been taken, the location of the measurement at time Ik+1 is selected, within a given set of admissible measurement points, in such a way as to minimize the spatial integral of the estimation error variance at time Ik+I' An algorithm is proposed, which incorporates the sclection of thc measurement point into a discrete-time Kalman filter that gives a finite-dimensional approximation of the state estimate. The characteristic features of the sequence of measurement points obtained by the algorithm are discussed on the basis of several numerical experiments. Finally, it is pointed out how the procedure can be applied to the solution of problems such as the allocation of a fixed sensor, the determination of the trajectory of a moving sensor, and the measurements scheduling among a predetermincd set of sensors.
21st Mediterranean Conference on Control and Automation, 2013
In this work we compare the performance of two algorithms, respectively based on the Extended Kal... more In this work we compare the performance of two algorithms, respectively based on the Extended Kalman Filter and the Unscented Kalman Filter, for the mobile robot localization and environment reconstruction problem. The proposed algorithms do not require any assumption on the robot working space: they are driven only by the measurements taken using ultrasonic sensors located onboard the robot. We also devise a switching sensors activation policy, which allows energy saving still achieving accurate tracking and reliable mapping of the workspace. The results show that the two filters work comparably well, in spite of the superior theoretical properties of the Unscented Filter.
2011 9th IEEE International Conference on Control and Automation (ICCA), 2011
ABSTRACT In this work we compare the performance of two well known filters for nonlinear models, ... more ABSTRACT In this work we compare the performance of two well known filters for nonlinear models, the Extended Kalman Filter and the Unscented Kalman Filter, in estimating the position and orientation of a mobile robot. The two filters are driven by measurements taken by ultrasonic sensors that are located onboard the robot, and a switching sensors activation policy is devised, which allows power saving and accurate tracking. The experimental results show a substantial equivalence of the two filters, although in principle the approximating properties of the UKF are much better.
IFAC Proceedings Volumes, 2014
A repetitive control scheme for asymptotic tracking of the fundamental frequency of periodic sign... more A repetitive control scheme for asymptotic tracking of the fundamental frequency of periodic signals is presented. The method uses an adaptive orthogonal signals generator based on a second order generalized integrator for identifying the fundamental frequency of the signal with accuracy depending on the adaptive gain. The robustness of the repetitive controller is established by investigating its sensitivities with respect to an inaccurate estimation of the period.
21st Mediterranean Conference on Control and Automation, 2013
The paper addresses the obstacle avoidance motion planning problem for ground vehicles operating ... more The paper addresses the obstacle avoidance motion planning problem for ground vehicles operating in uncertain environments. By resorting to set-theoretic ideas and sum of squares (SOS) decomposition techniques, a receding horizon control algorithm is proposed for robots modeled by polynomial systems subject to input, state and nonholonomic constraints. Sequences of inner ellipsoidal approximations of the exact one-step controllable sets are computed for all admissible obstacle scenarios and then on-line exploited to determine the more adequate control action to be applied in a receding horizon fashion. The results here proposed are a significant generalization of existing algorithms which are tailored only for linear time invariant plant descriptions. The resulting framework guarantees uniformly ultimate boundedness and constraints fulfilment regardless of any obstacle scenario occurrence.
18th Mediterranean Conference on Control and Automation, MED'10, 2010
This note presents the first experimental results about the application of a sensor switching pol... more This note presents the first experimental results about the application of a sensor switching policy to the estimation of the pose of a mobile robot. We assume that only a fraction of the available sensors can be used, to the sake of saving robot's batteries, and investigate sensors' better switching rule with respect to the reliability of the estimated pose under the usual hypotheses of Gaussian disturbances. The problem of possibly missing measurements, due to sensors malfunctioning or communication bottlenecks, is also addressed.
2009 IEEE International Conference on Control and Automation, 2009
... is the recent book [2]. The research has recently received a new impetus from the wide availa... more ... is the recent book [2]. The research has recently received a new impetus from the wide availability of cheap wireless sensors, for ... of the Finite Elements Method, Prentice-Hall, 1973 [11] D. Luenberger, Introduction to Dynamic Systems, Wiley, 1979 [12] BDO Anderson, JB Moore ...
2008 IEEE International Conference on Emerging Technologies and Factory Automation, 2008
ABSTRACT Wireless sensor networks are useful for many reasons, but they add at least two new issu... more ABSTRACT Wireless sensor networks are useful for many reasons, but they add at least two new issues to the Extended Kalman Filtering problem. First, they can be a further cause of divergence, as the information they send could not reach the filter. Second, batteries consumption must be taken into account: this leads to the need for a policy of querying, at each time instant, only a few sensors. In this paper we show how a wise sensor querying can improve the convergence rate of the filter, thus facing both the above problems. The querying criterion we suggest is simple to be implemented and adds a little computational overload to the filtering algorithm. The simulations we report, which refer to a mobile robot position estimation problem, show that it is effective in reducing the divergence rate of the filter.
[1992] Proceedings of the 31st IEEE Conference on Decision and Control
ABSTRACT A new regulator design for mechanical systems using the theory of the variable structure... more ABSTRACT A new regulator design for mechanical systems using the theory of the variable structure system (VSS) is presented. A relevant problem in VSS control is the eigenvalues assignment for the equivalent motion; taking into acount the acceleration of the desired trajectory, an analytical solution for this problem is proposed. Parameter variations and external disturbances can be easily considered in the design method. For robotic applications, the method can be applied to a robot with an high number of links. The problem of chattering is solved by introducing a saturation function
Proceedings of 1994 American Control Conference - ACC '94
ABSTRACT This paper presents a procedure for allocating measurement sensors on a rectangular plat... more ABSTRACT This paper presents a procedure for allocating measurement sensors on a rectangular plate, to achieve deformation measurement. The aim is to obtain a good reconstruction of the plate's state that enable us to reduce its vibrations through a state-feedback control law.
Proceedings of Tenth International Symposium on Intelligent Control, 1995
... Shape) and the more is the value of the regularization parameter for the reconstruction of th... more ... Shape) and the more is the value of the regularization parameter for the reconstruction of the Cylinder and the Double one; we summarize this result in the subsequent table in which it's measured the relative difference 2-norm between the ... [17] JL Schneiter, TB Sheridan. ...
Journal of Intelligent and Robotic Systems - JIRS, 1997
In this paper we propose a fuzzy rule-based algorithm for solvingclassification problems related ... more In this paper we propose a fuzzy rule-based algorithm for solvingclassification problems related to tactile sensing. A tactile sensing systemis a robotic device which gives an image of contacting objects. While thefine form recognition problem has been widely discussed and severaltechniques have been proposed for its solution (Bayesian approach or Neuralalgorithms), less attention has been paid to the problem of deciding whetherthe object belongs to a particular class or set of objects that share acommon feature, also known as tactile primitive. As input data we considerthe sum of the normal stresses at the sensing sites. Three levels ofclassification, hierarchically connected, are analyzed and, for each level,different basis variables with their membership functions are proposed andcalibrated using a training procedure. The output is an answer regarding thefeatures of the object at each level and is related to the truth values ofthe fuzzy classes. The numerical experiences show tha...
International Journal of Adaptive Control and Signal Processing, 1994
In this paper an observation strategy is presented, based on the concept of scanning observabilit... more In this paper an observation strategy is presented, based on the concept of scanning observability, for finite-dimensional linear systems with discrete observations. It aims to improve the convergence properties of the extended Kalman filter, regarded as an adaptive estimator of the state for the above class of systems. A significant application to a distributed parameter system is shown, related to the problem of determining a convenient trajectory for a moving sensor or a convenient scanning observation rule over a set of fixed sensors. The effectiveness of the resulting filter is analysed on the basis of several numerical experiments.
Control Engineering Practice, 1993
It is known that the application of the field-oriented control strategy to induction motors needs... more It is known that the application of the field-oriented control strategy to induction motors needs the knowledge of the rotor flux components and/or the rotor time constant. The aim of this paper is to give contributions to the solution of the above problem by means of real-time algorithms. The first part of the paper gives a short introduction about Kalman Filter theory. The second part is dedicated to the development of a suitable deterministic reduced order model with only the unrneasurable rotor fluxes as states, from which two reduced-order Kalman filters are obtained. The former one is called Reduced Order Linear Kalman Filter (RKF) and represents an effective algorithm for the estimation of the rotor flux components. The second filter, called Reduced Order Extended Kalman Filter (REKF) is obtained first developing a non linear model where the three state variables are the rotor flux components and the rotor time constant and then applying the extended Kalman Filter theory. The last part of the paper is related to the simulations results and the implementation tests both for RKF and REKF algorithms to verify their effectiveness in the problems of real-time applications.
2014 IEEE International Conference on Robotics and Automation (ICRA), 2014
In this paper the problem of estimating the relative orientation and position between a camera an... more In this paper the problem of estimating the relative orientation and position between a camera and an object is investigated. It is assumed that both the camera and the object are provided with an Inertial Measurement Unit (IMU) capable of measuring their attitude with respect to the gravity and the earth magnetic vectors. Furthermore, the object is assumed to contain a feature of n points, the position of which is known in the object coordinate frame. An algorithm is proposed, which uses the image provided by the camera and the information provided by the IMUs, to solve the PnP problem, i.e., to estimate the relative pose of the object in the camera reference frame. Two special cases will be studied. The first is the case where all the attitude information given by the IMU is used. In the second case only the measurements provided by inclinometers are used, neglecting those coming from the magnetometers, because they are usually quite noisy. The effectiveness of the proposed algorithms is tested either by numerical simulations and by experimental tests with cameras.
Proceedings of the 1992 International Conference on Industrial Electronics, Control, Instrumentation, and Automation
The problem is considered of estimating the dynamic characteristics of the elastic modes excited ... more The problem is considered of estimating the dynamic characteristics of the elastic modes excited by the control input in the flexible forearm of a two-joint robot. The complexity of the nonlinear model of the elastic link is substantially reduced, at least with reference to the estimation problem, by assuming the availability of direct measurements or estimates of the state variables
Mathematics and Computers in Simulation, 1994
ABSTRACT In this paper the problem of modeling two-joint planar robots with a flexible forearm is... more ABSTRACT In this paper the problem of modeling two-joint planar robots with a flexible forearm is considered. A simplified finite element approach is proposed in order to obtain simple and structured models which can easily take into account transient phenomena and external variable forces; gravity is also kept into account. The model can be easily generalised to robots with multiple elastic links. A comparison is made with the eigenfunction expansion model in stationary conditions by evaluating the eigenvalues of the elastic link. The remarkable structure of the model can be useful in control designing, owing to the fact that the usual properties of mechanical systems are still valid.
Proceedings of 32nd IEEE Conference on Decision and Control
ABSTRACT This paper proposes a new regulator design for mechanical systems, using the theory of v... more ABSTRACT This paper proposes a new regulator design for mechanical systems, using the theory of variable structure systems. The proposed control law needs no inversion of the inertia matrix and no evaluation of its derivative. The chattering phenomenon is avoided by using a pair of switching lines for each generalized coordinate
Machine Vision and Applications
In this work, we face the problem of estimating the relative position and orientation of a camera... more In this work, we face the problem of estimating the relative position and orientation of a camera and an object, when they are both equipped with inertial measurement units (IMUs), and the object exhibits a set of n landmark points with known coordinates (the so-called Pose estimation or PnP Problem). We present two algorithms that, fusing the information provided by the camera and the IMUs, solve the PnP problem with good accuracy. These algorithms only use the measurements given by IMUs’ inclinometers, as the magnetometers usually give inaccurate estimates of the Earth magnetic vector. The effectiveness of the proposed methods is assessed by numerical simulations and experimental tests. The results of the tests are compared with the most recent methods proposed in the literature.
IEEE Transactions on Automatic Control, 1994
Lai and H. Robbins, "Consistency and asymptotic efficiency of slope estimates in stochastic schem... more Lai and H. Robbins, "Consistency and asymptotic efficiency of slope estimates in stochastic schemes." Z Wahrscheinlichkeirs theorie y e w .
International Journal of Modelling and Simulation, 1992
This paper considers the problem of locating a sensor for the state estimation of a linear stocha... more This paper considers the problem of locating a sensor for the state estimation of a linear stochastic distributed parameter system of parabolic type, with observations taken at discrete times 1 1 ,12' .... Assuming that k measurements have been taken, the location of the measurement at time Ik+1 is selected, within a given set of admissible measurement points, in such a way as to minimize the spatial integral of the estimation error variance at time Ik+I' An algorithm is proposed, which incorporates the sclection of thc measurement point into a discrete-time Kalman filter that gives a finite-dimensional approximation of the state estimate. The characteristic features of the sequence of measurement points obtained by the algorithm are discussed on the basis of several numerical experiments. Finally, it is pointed out how the procedure can be applied to the solution of problems such as the allocation of a fixed sensor, the determination of the trajectory of a moving sensor, and the measurements scheduling among a predetermincd set of sensors.
21st Mediterranean Conference on Control and Automation, 2013
In this work we compare the performance of two algorithms, respectively based on the Extended Kal... more In this work we compare the performance of two algorithms, respectively based on the Extended Kalman Filter and the Unscented Kalman Filter, for the mobile robot localization and environment reconstruction problem. The proposed algorithms do not require any assumption on the robot working space: they are driven only by the measurements taken using ultrasonic sensors located onboard the robot. We also devise a switching sensors activation policy, which allows energy saving still achieving accurate tracking and reliable mapping of the workspace. The results show that the two filters work comparably well, in spite of the superior theoretical properties of the Unscented Filter.
2011 9th IEEE International Conference on Control and Automation (ICCA), 2011
ABSTRACT In this work we compare the performance of two well known filters for nonlinear models, ... more ABSTRACT In this work we compare the performance of two well known filters for nonlinear models, the Extended Kalman Filter and the Unscented Kalman Filter, in estimating the position and orientation of a mobile robot. The two filters are driven by measurements taken by ultrasonic sensors that are located onboard the robot, and a switching sensors activation policy is devised, which allows power saving and accurate tracking. The experimental results show a substantial equivalence of the two filters, although in principle the approximating properties of the UKF are much better.
IFAC Proceedings Volumes, 2014
A repetitive control scheme for asymptotic tracking of the fundamental frequency of periodic sign... more A repetitive control scheme for asymptotic tracking of the fundamental frequency of periodic signals is presented. The method uses an adaptive orthogonal signals generator based on a second order generalized integrator for identifying the fundamental frequency of the signal with accuracy depending on the adaptive gain. The robustness of the repetitive controller is established by investigating its sensitivities with respect to an inaccurate estimation of the period.
21st Mediterranean Conference on Control and Automation, 2013
The paper addresses the obstacle avoidance motion planning problem for ground vehicles operating ... more The paper addresses the obstacle avoidance motion planning problem for ground vehicles operating in uncertain environments. By resorting to set-theoretic ideas and sum of squares (SOS) decomposition techniques, a receding horizon control algorithm is proposed for robots modeled by polynomial systems subject to input, state and nonholonomic constraints. Sequences of inner ellipsoidal approximations of the exact one-step controllable sets are computed for all admissible obstacle scenarios and then on-line exploited to determine the more adequate control action to be applied in a receding horizon fashion. The results here proposed are a significant generalization of existing algorithms which are tailored only for linear time invariant plant descriptions. The resulting framework guarantees uniformly ultimate boundedness and constraints fulfilment regardless of any obstacle scenario occurrence.
18th Mediterranean Conference on Control and Automation, MED'10, 2010
This note presents the first experimental results about the application of a sensor switching pol... more This note presents the first experimental results about the application of a sensor switching policy to the estimation of the pose of a mobile robot. We assume that only a fraction of the available sensors can be used, to the sake of saving robot's batteries, and investigate sensors' better switching rule with respect to the reliability of the estimated pose under the usual hypotheses of Gaussian disturbances. The problem of possibly missing measurements, due to sensors malfunctioning or communication bottlenecks, is also addressed.
2009 IEEE International Conference on Control and Automation, 2009
... is the recent book [2]. The research has recently received a new impetus from the wide availa... more ... is the recent book [2]. The research has recently received a new impetus from the wide availability of cheap wireless sensors, for ... of the Finite Elements Method, Prentice-Hall, 1973 [11] D. Luenberger, Introduction to Dynamic Systems, Wiley, 1979 [12] BDO Anderson, JB Moore ...
2008 IEEE International Conference on Emerging Technologies and Factory Automation, 2008
ABSTRACT Wireless sensor networks are useful for many reasons, but they add at least two new issu... more ABSTRACT Wireless sensor networks are useful for many reasons, but they add at least two new issues to the Extended Kalman Filtering problem. First, they can be a further cause of divergence, as the information they send could not reach the filter. Second, batteries consumption must be taken into account: this leads to the need for a policy of querying, at each time instant, only a few sensors. In this paper we show how a wise sensor querying can improve the convergence rate of the filter, thus facing both the above problems. The querying criterion we suggest is simple to be implemented and adds a little computational overload to the filtering algorithm. The simulations we report, which refer to a mobile robot position estimation problem, show that it is effective in reducing the divergence rate of the filter.
[1992] Proceedings of the 31st IEEE Conference on Decision and Control
ABSTRACT A new regulator design for mechanical systems using the theory of the variable structure... more ABSTRACT A new regulator design for mechanical systems using the theory of the variable structure system (VSS) is presented. A relevant problem in VSS control is the eigenvalues assignment for the equivalent motion; taking into acount the acceleration of the desired trajectory, an analytical solution for this problem is proposed. Parameter variations and external disturbances can be easily considered in the design method. For robotic applications, the method can be applied to a robot with an high number of links. The problem of chattering is solved by introducing a saturation function
Proceedings of 1994 American Control Conference - ACC '94
ABSTRACT This paper presents a procedure for allocating measurement sensors on a rectangular plat... more ABSTRACT This paper presents a procedure for allocating measurement sensors on a rectangular plate, to achieve deformation measurement. The aim is to obtain a good reconstruction of the plate's state that enable us to reduce its vibrations through a state-feedback control law.
Proceedings of Tenth International Symposium on Intelligent Control, 1995
... Shape) and the more is the value of the regularization parameter for the reconstruction of th... more ... Shape) and the more is the value of the regularization parameter for the reconstruction of the Cylinder and the Double one; we summarize this result in the subsequent table in which it's measured the relative difference 2-norm between the ... [17] JL Schneiter, TB Sheridan. ...
Journal of Intelligent and Robotic Systems - JIRS, 1997
In this paper we propose a fuzzy rule-based algorithm for solvingclassification problems related ... more In this paper we propose a fuzzy rule-based algorithm for solvingclassification problems related to tactile sensing. A tactile sensing systemis a robotic device which gives an image of contacting objects. While thefine form recognition problem has been widely discussed and severaltechniques have been proposed for its solution (Bayesian approach or Neuralalgorithms), less attention has been paid to the problem of deciding whetherthe object belongs to a particular class or set of objects that share acommon feature, also known as tactile primitive. As input data we considerthe sum of the normal stresses at the sensing sites. Three levels ofclassification, hierarchically connected, are analyzed and, for each level,different basis variables with their membership functions are proposed andcalibrated using a training procedure. The output is an answer regarding thefeatures of the object at each level and is related to the truth values ofthe fuzzy classes. The numerical experiences show tha...
International Journal of Adaptive Control and Signal Processing, 1994
In this paper an observation strategy is presented, based on the concept of scanning observabilit... more In this paper an observation strategy is presented, based on the concept of scanning observability, for finite-dimensional linear systems with discrete observations. It aims to improve the convergence properties of the extended Kalman filter, regarded as an adaptive estimator of the state for the above class of systems. A significant application to a distributed parameter system is shown, related to the problem of determining a convenient trajectory for a moving sensor or a convenient scanning observation rule over a set of fixed sensors. The effectiveness of the resulting filter is analysed on the basis of several numerical experiments.
Control Engineering Practice, 1993
It is known that the application of the field-oriented control strategy to induction motors needs... more It is known that the application of the field-oriented control strategy to induction motors needs the knowledge of the rotor flux components and/or the rotor time constant. The aim of this paper is to give contributions to the solution of the above problem by means of real-time algorithms. The first part of the paper gives a short introduction about Kalman Filter theory. The second part is dedicated to the development of a suitable deterministic reduced order model with only the unrneasurable rotor fluxes as states, from which two reduced-order Kalman filters are obtained. The former one is called Reduced Order Linear Kalman Filter (RKF) and represents an effective algorithm for the estimation of the rotor flux components. The second filter, called Reduced Order Extended Kalman Filter (REKF) is obtained first developing a non linear model where the three state variables are the rotor flux components and the rotor time constant and then applying the extended Kalman Filter theory. The last part of the paper is related to the simulations results and the implementation tests both for RKF and REKF algorithms to verify their effectiveness in the problems of real-time applications.
2014 IEEE International Conference on Robotics and Automation (ICRA), 2014
In this paper the problem of estimating the relative orientation and position between a camera an... more In this paper the problem of estimating the relative orientation and position between a camera and an object is investigated. It is assumed that both the camera and the object are provided with an Inertial Measurement Unit (IMU) capable of measuring their attitude with respect to the gravity and the earth magnetic vectors. Furthermore, the object is assumed to contain a feature of n points, the position of which is known in the object coordinate frame. An algorithm is proposed, which uses the image provided by the camera and the information provided by the IMUs, to solve the PnP problem, i.e., to estimate the relative pose of the object in the camera reference frame. Two special cases will be studied. The first is the case where all the attitude information given by the IMU is used. In the second case only the measurements provided by inclinometers are used, neglecting those coming from the magnetometers, because they are usually quite noisy. The effectiveness of the proposed algorithms is tested either by numerical simulations and by experimental tests with cameras.
Proceedings of the 1992 International Conference on Industrial Electronics, Control, Instrumentation, and Automation
The problem is considered of estimating the dynamic characteristics of the elastic modes excited ... more The problem is considered of estimating the dynamic characteristics of the elastic modes excited by the control input in the flexible forearm of a two-joint robot. The complexity of the nonlinear model of the elastic link is substantially reduced, at least with reference to the estimation problem, by assuming the availability of direct measurements or estimates of the state variables
Mathematics and Computers in Simulation, 1994
ABSTRACT In this paper the problem of modeling two-joint planar robots with a flexible forearm is... more ABSTRACT In this paper the problem of modeling two-joint planar robots with a flexible forearm is considered. A simplified finite element approach is proposed in order to obtain simple and structured models which can easily take into account transient phenomena and external variable forces; gravity is also kept into account. The model can be easily generalised to robots with multiple elastic links. A comparison is made with the eigenfunction expansion model in stationary conditions by evaluating the eigenvalues of the elastic link. The remarkable structure of the model can be useful in control designing, owing to the fact that the usual properties of mechanical systems are still valid.
Proceedings of 32nd IEEE Conference on Decision and Control
ABSTRACT This paper proposes a new regulator design for mechanical systems, using the theory of v... more ABSTRACT This paper proposes a new regulator design for mechanical systems, using the theory of variable structure systems. The proposed control law needs no inversion of the inertia matrix and no evaluation of its derivative. The chattering phenomenon is avoided by using a pair of switching lines for each generalized coordinate
Machine Vision and Applications
In this work, we face the problem of estimating the relative position and orientation of a camera... more In this work, we face the problem of estimating the relative position and orientation of a camera and an object, when they are both equipped with inertial measurement units (IMUs), and the object exhibits a set of n landmark points with known coordinates (the so-called Pose estimation or PnP Problem). We present two algorithms that, fusing the information provided by the camera and the IMUs, solve the PnP problem with good accuracy. These algorithms only use the measurements given by IMUs’ inclinometers, as the magnetometers usually give inaccurate estimates of the Earth magnetic vector. The effectiveness of the proposed methods is assessed by numerical simulations and experimental tests. The results of the tests are compared with the most recent methods proposed in the literature.