Ponències/Comunicacions de congressos
http://hdl.handle.net/2117/2254
2015-11-27T06:32:11ZDimensionality reduction and motion coordination in learning trajectories with dynamic movement primitives
http://hdl.handle.net/2117/77467
Dimensionality reduction and motion coordination in learning trajectories with dynamic movement primitives
Colomé Figueras, Adrià; Torras, Carme
Dynamic Movement Primitives (DMP) are nowadays widely used as movement parametrization for learning trajectories, because of their linearity in the parameters, rescaling robustness and continuity. However, when learning a movement with a robot using DMP, many parameters may need to be tuned, requiring a prohibitive number of experiments/simulations to converge to a solution with a locally or globally optimal reward. We propose here strategies to palliate this dimensionality problem: the first is to explore only along the most significant directions in the parameter space, and the second is to add a reduced second set of Gaussians that would optimize the trajectory after fixing the Gaussians approximating the demonstrated movement. Both strategies result in less Gaussian computations and better performance on learning algorithms.
To further speed up the learning and allow for a better biased exploration, we also propose to coordinate the motion of different joints, by computing a coordination matrix initialized with the demonstrated movement and then automatically updating it by eliminating the degrees of freedom least affecting task performance.
Our three proposals have been experimentally tested and the obtained results show that similar (or even better) performance can be obtained at a significantly lower computational cost by reducing the dimensionality of the exploration space.
2015-10-07T17:08:15ZColomé Figueras, AdriàTorras, CarmeDynamic Movement Primitives (DMP) are nowadays widely used as movement parametrization for learning trajectories, because of their linearity in the parameters, rescaling robustness and continuity. However, when learning a movement with a robot using DMP, many parameters may need to be tuned, requiring a prohibitive number of experiments/simulations to converge to a solution with a locally or globally optimal reward. We propose here strategies to palliate this dimensionality problem: the first is to explore only along the most significant directions in the parameter space, and the second is to add a reduced second set of Gaussians that would optimize the trajectory after fixing the Gaussians approximating the demonstrated movement. Both strategies result in less Gaussian computations and better performance on learning algorithms.
To further speed up the learning and allow for a better biased exploration, we also propose to coordinate the motion of different joints, by computing a coordination matrix initialized with the demonstrated movement and then automatically updating it by eliminating the degrees of freedom least affecting task performance.
Our three proposals have been experimentally tested and the obtained results show that similar (or even better) performance can be obtained at a significantly lower computational cost by reducing the dimensionality of the exploration space.Multimodal feedback fusion of laser, image and temporal information
http://hdl.handle.net/2117/77457
Multimodal feedback fusion of laser, image and temporal information
Huerta Casado, Iván; Ferrer Mínguez, Gonzalo; Herrero Cotarelo, Fernando; Prati, Andrea; Sanfeliu Cortés, Alberto
In the present paper, we propose a highly accurate and robust people detector, which works well under highly variant and uncertain conditions, such as occlusions, false positives and false detections. These adverse conditions, which initially motivated this research, occur when a robotic platform navigates in an urban environment, and although the scope is originally within the robotics field, the authors believe that our contributions can be extended to other fields. To this end, we propose a multimodal information fusion consisting of laser and monocular camera information. Laser information is modelled using a set of weak classifiers (Adaboost) to detect people. Camera information is processed by using HOG descriptors to classify person/non person based on a linear SVM. A multi-hypothesis tracker trails the position and velocity of each of the targets, providing temporal information to the fusion, allowing recovery of detections even when the laser segmentation fails. Experimental results show that our feedback-based system outperforms previous state-of-the-art methods in performance and accuracy, and that near real-time detection performance can be achieved.
2015-10-07T15:43:05ZHuerta Casado, IvánFerrer Mínguez, GonzaloHerrero Cotarelo, FernandoPrati, AndreaSanfeliu Cortés, AlbertoIn the present paper, we propose a highly accurate and robust people detector, which works well under highly variant and uncertain conditions, such as occlusions, false positives and false detections. These adverse conditions, which initially motivated this research, occur when a robotic platform navigates in an urban environment, and although the scope is originally within the robotics field, the authors believe that our contributions can be extended to other fields. To this end, we propose a multimodal information fusion consisting of laser and monocular camera information. Laser information is modelled using a set of weak classifiers (Adaboost) to detect people. Camera information is processed by using HOG descriptors to classify person/non person based on a linear SVM. A multi-hypothesis tracker trails the position and velocity of each of the targets, providing temporal information to the fusion, allowing recovery of detections even when the laser segmentation fails. Experimental results show that our feedback-based system outperforms previous state-of-the-art methods in performance and accuracy, and that near real-time detection performance can be achieved.On generalized dual Euler angles
http://hdl.handle.net/2117/77065
On generalized dual Euler angles
Rull Sanahuja, Aleix; Thomas, Federico
This paper first explores the generalization of Euler angles to the case in which the rotation axes are not necessarily members of an orthonormal triad, and presents a concise solution to their computation that relies on the calculation of standard Euler angles. Then, this generalization is taken one step further by introducing translations, that is, by defining generalized Euler angles about screw axes using a variation of the principle of transference that avoids the use of dual numbers. As an example, the obtained formulation is applied to solve the inverse kinematics of a 3C manipulator.
2015-09-23T16:43:10ZRull Sanahuja, AleixThomas, FedericoThis paper first explores the generalization of Euler angles to the case in which the rotation axes are not necessarily members of an orthonormal triad, and presents a concise solution to their computation that relies on the calculation of standard Euler angles. Then, this generalization is taken one step further by introducing translations, that is, by defining generalized Euler angles about screw axes using a variation of the principle of transference that avoids the use of dual numbers. As an example, the obtained formulation is applied to solve the inverse kinematics of a 3C manipulator.Evaluation of random forests on large-scale classification problems using a bag-of-visual-words representation
http://hdl.handle.net/2117/77001
Evaluation of random forests on large-scale classification problems using a bag-of-visual-words representation
Soler, Xavier; Ramisa Ayats, Arnau; Torras, Carme
Random Forest is a very efficient classification method that has shown success in tasks like image segmentation or object detection, but has not been applied yet in large-scale image classification scenarios using a Bag-of-Visual-Words representation. In this work we evaluate the performance of Random Forest on the ImageNet dataset, and compare it to standard approaches in the state-of-the-art.
2015-09-21T17:18:38ZSoler, XavierRamisa Ayats, ArnauTorras, CarmeRandom Forest is a very efficient classification method that has shown success in tasks like image segmentation or object detection, but has not been applied yet in large-scale image classification scenarios using a Bag-of-Visual-Words representation. In this work we evaluate the performance of Random Forest on the ImageNet dataset, and compare it to standard approaches in the state-of-the-art.Estimación monocular y eficiente de la pose usando modelos 3D complejos
http://hdl.handle.net/2117/76830
Estimación monocular y eficiente de la pose usando modelos 3D complejos
Rubio Romano, Antonio; Villamizar Vergel, Michael Alejandro; Ferraz Colomina, Luis; Peñate Sánchez, Adrián; Sanfeliu Cortés, Alberto; Moreno-Noguer, Francesc
El siguiente documento presenta un método robusto y eficiente para estimar la pose de una cámara. El método propuesto asume el conocimiento previo de un modelo 3D del entorno, y compara una nueva imagen de entrada únicamente con un conjunto pequeño de imágenes similares seleccionadas previamente por un algoritmo de
2015-09-15T18:53:23ZRubio Romano, AntonioVillamizar Vergel, Michael AlejandroFerraz Colomina, LuisPeñate Sánchez, AdriánSanfeliu Cortés, AlbertoMoreno-Noguer, FrancescEl siguiente documento presenta un método robusto y eficiente para estimar la pose de una cámara. El método propuesto asume el conocimiento previo de un modelo 3D del entorno, y compara una nueva imagen de entrada únicamente con un conjunto pequeño de imágenes similares seleccionadas previamente por un algoritmo deLeveraging feature uncertainty in the PnP problem
http://hdl.handle.net/2117/76828
Leveraging feature uncertainty in the PnP problem
Ferraz Colomina, Luis; Binefa, Xavier; Moreno-Noguer, Francesc
We propose a real-time and accurate solution to the Perspective-n-Point (PnP) problem –estimating the pose of a calibrated camera from n 3D-to-2D point correspondences– that exploits the fact that in practice the 2D position of not all 2D features is estimated with the same accuracy. Assuming a model of such feature uncertainties is known in advance, we reformulate the PnP problem as a maximum likelihood minimization approximated by an unconstrained Sampson error function, which naturally penalizes the most noisy correspondences. The advantages of this approach are thoroughly demonstrated in synthetic experiments where feature uncertainties are exactly known.
Pre-estimating the features uncertainties in real experiments is, though, not easy. In this paper we model feature uncertainty as 2D Gaussian distributions representing the sensitivity of the 2D feature detectors to different camera viewpoints. When using these noise models with our PnP formulation we still obtain promising pose estimation results that outperform the most recent approaches.
2015-09-15T17:10:09ZFerraz Colomina, LuisBinefa, XavierMoreno-Noguer, FrancescWe propose a real-time and accurate solution to the Perspective-n-Point (PnP) problem –estimating the pose of a calibrated camera from n 3D-to-2D point correspondences– that exploits the fact that in practice the 2D position of not all 2D features is estimated with the same accuracy. Assuming a model of such feature uncertainties is known in advance, we reformulate the PnP problem as a maximum likelihood minimization approximated by an unconstrained Sampson error function, which naturally penalizes the most noisy correspondences. The advantages of this approach are thoroughly demonstrated in synthetic experiments where feature uncertainties are exactly known.
Pre-estimating the features uncertainties in real experiments is, though, not easy. In this paper we model feature uncertainty as 2D Gaussian distributions representing the sensitivity of the 2D feature detectors to different camera viewpoints. When using these noise models with our PnP formulation we still obtain promising pose estimation results that outperform the most recent approaches.Two-layer scheduling scheme for pump stations
http://hdl.handle.net/2117/76686
Two-layer scheduling scheme for pump stations
Sun, Congcong; Puig Cayuela, Vicenç; Cembrano Gennari, Gabriela
In this paper, a two-layer scheduling scheme for pump stations in a water distribution network has been proposed. The upper layer, which works in one-hour sampling time, uses Model Predictive Control (MPC) to produce continuous flow set-points for the lower layer. While in the lower layer, a scheduling algorithm has been used to translate the continuous flow set-points to a discrete (ON-OFF) control operation sequence of the pump stations with the constraints that pump stations should draw the same amount of water as the continuous flow set-points provided by the upper layer.
The tuning parameters of such algorithm are the lower layer control sampling period and the number of parallel pumps in the pump station. The proposed method has been tested in the Richmond case study.
2015-09-08T11:31:04ZSun, CongcongPuig Cayuela, VicençCembrano Gennari, GabrielaIn this paper, a two-layer scheduling scheme for pump stations in a water distribution network has been proposed. The upper layer, which works in one-hour sampling time, uses Model Predictive Control (MPC) to produce continuous flow set-points for the lower layer. While in the lower layer, a scheduling algorithm has been used to translate the continuous flow set-points to a discrete (ON-OFF) control operation sequence of the pump stations with the constraints that pump stations should draw the same amount of water as the continuous flow set-points provided by the upper layer.
The tuning parameters of such algorithm are the lower layer control sampling period and the number of parallel pumps in the pump station. The proposed method has been tested in the Richmond case study.Very fast solution to the PnP problem with algebraic outlier rejection
http://hdl.handle.net/2117/76612
Very fast solution to the PnP problem with algebraic outlier rejection
Ferraz Colomina, Luis; Binefa, Xavier; Moreno-Noguer, Francesc
We propose a real-time, robust to outliers and accurate solution to the Perspective-n-Point (PnP) problem. The main advantages of our solution are twofold: first, it integrates the outlier rejection within the pose estimation pipeline with a negligible computational overhead; and second, its scalability to arbitrarily large number of correspondences. Given a set of 3D-to-2D matches, we formulate pose estimation problem as a low-rank homogeneous system where the solution lies on its 1D null space. Outlier correspondences are those rows of the linear system which perturb the null space and are progressively detected by projecting them on an iteratively estimated solution of the null space. Since our outlier removal process is based on an algebraic criterion which does not require computing the full-pose and reprojecting back all 3D points on the image plane at each step, we achieve speed gains of more than 100× compared to RANSAC strategies. An extensive experimental evaluation will show that our solution yields accu- rate results in situations with up to 50% of outliers, and can process more than 1000 correspondences in less than 5ms.
2015-09-04T09:11:51ZFerraz Colomina, LuisBinefa, XavierMoreno-Noguer, FrancescWe propose a real-time, robust to outliers and accurate solution to the Perspective-n-Point (PnP) problem. The main advantages of our solution are twofold: first, it integrates the outlier rejection within the pose estimation pipeline with a negligible computational overhead; and second, its scalability to arbitrarily large number of correspondences. Given a set of 3D-to-2D matches, we formulate pose estimation problem as a low-rank homogeneous system where the solution lies on its 1D null space. Outlier correspondences are those rows of the linear system which perturb the null space and are progressively detected by projecting them on an iteratively estimated solution of the null space. Since our outlier removal process is based on an algebraic criterion which does not require computing the full-pose and reprojecting back all 3D points on the image plane at each step, we achieve speed gains of more than 100× compared to RANSAC strategies. An extensive experimental evaluation will show that our solution yields accu- rate results in situations with up to 50% of outliers, and can process more than 1000 correspondences in less than 5ms.Active learning of manipulation sequences
http://hdl.handle.net/2117/76605
Active learning of manipulation sequences
Martínez Martínez, David; Alenyà Ribas, Guillem; Jimenez Schlegl, Pablo; Torras, Carme; Rossmann, Jürgen; Wantia, Nils; Eren Erdal, Aksoy; Haller, Simon; Piater, Justus
We describe a system allowing a robot to learn goal-directed manipulation sequences such as steps of an assembly task. Learning is based on a free mix of exploration and instruction by an external teacher, and may be active in the sense that the system tests actions to maximize learning progress and asks the teacher if needed. The main component is a symbolic planning engine that operates on learned rules, defined by actions and their pre- and postconditions. Learned by model-based reinforcement learning, rules are immediately available for planning. Thus, there are no distinct learning and application phases. We show how dynamic plans, replanned after every action if necessary, can be used for automatic execution of manipulation sequences, for monitoring of observed manipulation sequences, or a mix of the two, all while extending and refining the rule base on the fly. Quantitative results indicate fast convergence using few training examples, and highly effective teacher intervention at early stages of learning.
2015-09-03T14:21:07ZMartínez Martínez, DavidAlenyà Ribas, GuillemJimenez Schlegl, PabloTorras, CarmeRossmann, JürgenWantia, NilsEren Erdal, AksoyHaller, SimonPiater, JustusWe describe a system allowing a robot to learn goal-directed manipulation sequences such as steps of an assembly task. Learning is based on a free mix of exploration and instruction by an external teacher, and may be active in the sense that the system tests actions to maximize learning progress and asks the teacher if needed. The main component is a symbolic planning engine that operates on learned rules, defined by actions and their pre- and postconditions. Learned by model-based reinforcement learning, rules are immediately available for planning. Thus, there are no distinct learning and application phases. We show how dynamic plans, replanned after every action if necessary, can be used for automatic execution of manipulation sequences, for monitoring of observed manipulation sequences, or a mix of the two, all while extending and refining the rule base on the fly. Quantitative results indicate fast convergence using few training examples, and highly effective teacher intervention at early stages of learning.Proactive kinodynamic planning using the extended social force model and human motion prediction in urban environments
http://hdl.handle.net/2117/76556
Proactive kinodynamic planning using the extended social force model and human motion prediction in urban environments
Ferrer Mínguez, Gonzalo; Sanfeliu Cortés, Alberto
This paper presents a novel approach for robot navigation in crowded urban environments where people and objects are moving simultaneously while a robot is navigating. Avoiding moving obstacles at their corresponding precise moment motivates the use of a robotic planner satisfying both dynamic and nonholonomic constraints, also referred as kynodynamic constraints. We present a proactive navigation approach with respect its environment, in the sense that the robot calculates the reaction produced by its actions and provides the minimum impact on nearby pedestrians. As a consequence, the proposed planner integrates seamlessly planning and prediction and calculates a complete motion prediction of the scene for each robot propagation. Making use of the Extended Social Force Model (ESFM) allows an enormous simplification for both the prediction model and the planning system under differential constraints. Simulations and real experiments have been carried out to demonstrate the success of the proactive kinodynamic planner.
2015-09-02T12:45:32ZFerrer Mínguez, GonzaloSanfeliu Cortés, AlbertoThis paper presents a novel approach for robot navigation in crowded urban environments where people and objects are moving simultaneously while a robot is navigating. Avoiding moving obstacles at their corresponding precise moment motivates the use of a robotic planner satisfying both dynamic and nonholonomic constraints, also referred as kynodynamic constraints. We present a proactive navigation approach with respect its environment, in the sense that the robot calculates the reaction produced by its actions and provides the minimum impact on nearby pedestrians. As a consequence, the proposed planner integrates seamlessly planning and prediction and calculates a complete motion prediction of the scene for each robot propagation. Making use of the Extended Social Force Model (ESFM) allows an enormous simplification for both the prediction model and the planning system under differential constraints. Simulations and real experiments have been carried out to demonstrate the success of the proactive kinodynamic planner.