Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 40
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Word ordering and document adjacency for large loop closure detection in 2D laser maps

  • Deray, Jeremie
  • Solà, Joan
  • Andrade-Cetto, Juan
We address in this paper the problem of loop closure detection for laser-based simultaneous localization and mapping (SLAM) of very large areas. Consistent with the state of the art, the map is encoded as a graph of poses, and to cope with very large mapping capabilities, loop closures are asserted by comparing the features extracted from a query laser scan against a previously acquired corpus of scan features using a bag-ofwords (BoW) scheme. Two contributions are here presented. First, to benefit from the graph topology, feature frequency scores in the BoW are computed not only for each individual scan but also from neighboring scans in the SLAM graph. This has the effect of enforcing neighbor relational information during document matching. Secondly, a weak geometric check that takes into account feature ordering and occlusions is introduced that substantially improves loop closure detection performance. The two contributions are evaluated both separately and jointly on four common SLAM datasets, and are shown to improve the state-of-the-art performance both in terms of precision and recall in most of the cases. Moreover, our current implementation is designed to work at nearly frame rate, allowing loop closure query resolution at nearly 22 Hz for the best case scenario and 2 Hz for the worst case scenario., This work was supported by the Spanish Ministry of Economy and Competitiveness project ROBINSTRUCT TIN2014-58178-R, the EU H2020 LOGMATIC project 687534, and the Industrial Doctoral program of the Catalan Agency for Management of University and Research Grants., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

Global model with local interpretation for dynamic shape reconstruction

  • Agudo Martínez, Antonio
  • Moreno-Noguer, Francesc
Trabajo presentado a la IEEE Winter Conference on Applications of Computer Vision (WACV), celebrada en Santa Rosa, California (US) del 24 al 31 de marzo de 2017., The most standard approach to resolve the inherent ambiguities of the non-rigid structure from motion problem is using low-rank models that approximate deforming shapes by a linear combination of rigid basis. These models are typically global, i.e., each shape basis contributes equally to all points of the surface. While this approach has been shown effective to represent smooth deformations, its performance degrades for surfaces composed of various regions, each following a different deformation rule. Piecewise methods attempt to capture this type of behavior by locally modeling surface patches, although they subsequently require enforcing global constraints to assemble back the patches. In this paper we propose an approach that combines the best of global and local models: it locally considers low-rank models but, by construction, does not need to impose global constraints to guarantee local patch continuity. We achieve this by a simple expectation maximization strategy that besides learning global shape bases, it locally adapts their contribution to each specific surface region. Furthermore, as a side contribution, in order to split the surface into different local patches, we propose a novel physically-based mesh segmentation approach that obeys an energy criterion. The complete framework is evaluated in both synthetic and real datasets, and shows an improved performance to competing methods., This work has been partially supported by the Spanish Ministry of Science and Innovation under project RobInstruct TIN2014-58178-R, by the ERA-net CHISTERA project I-DRESS PCIN-2015-147 and by the EU project LOGIMATIC H2020-Galileo-2015-1-687534., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

DUST: Dual union of spatio-temporal subspaces for monocular multiple object 3D reconstruction

  • Agudo Martínez, Antonio
  • Moreno-Noguer, Francesc
Trabajo presentado a la 30th IEEE Conference on Computer Vision and Pattern Recognition (CVPR), celebrada en Honolulu, Hawaii (US) del 21 al 26 de julio de 2016., We present an approach to reconstruct the 3D shape of multiple deforming objects from incomplete 2D trajectories acquired by a single camera. Additionally, we simultaneously provide spatial segmentation (i.e., we identify each of the objects in every frame) and temporal clustering (i.e., we split the sequence into primitive actions). This advances existing work, which only tackled the problem for one single object and non-occluded tracks. In order to handle several objects at a time from partial observations, we model point trajectories as a union of spatial and temporal subspaces, and optimize the parameters of both modalities, the non-observed point tracks and the 3D shape via augmented Lagrange multipliers. The algorithm is fully unsupervised and results in a formulation which does not need initialization. We thoroughly validate the method on challenging scenarios with several human subjects performing different activities which involve complex motions and close interaction. We show our approach achieves state-of-the-art 3D reconstruction results, while it also provides spatial and temporal segmentation., This work has been partially supported by the Spanish Ministry of Science and Innovation under project RobInstruct TIN2014-58178-R, by the ERA-net CHISTERA project I-DRESS PCIN-2015-147 and by the EU project LOGIMATIC H2020-Galileo-2015-1-687534., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Boosted Random Ferns for object detection

  • Villamizar, Michael
  • Andrade-Cetto, Juan
  • Sanfeliu, Alberto
  • Moreno-Noguer, Francesc
In this paper we introduce the Boosted Random Ferns (BRFs) to rapidly build discriminative classifiers for learning and detecting object categories. At the core of our approach we use standard random ferns, but we introduce four main innovations that let us bring ferns from an instance to a category level, and still retain efficiency. First, we define binary features on the histogram of oriented gradients-domain (as opposed to intensity-), allowing for a better representation of intra-class variability. Second, both the positions where ferns are evaluated within the sliding window, and the location of the binary features for each fern are not chosen completely at random, but instead we use a boosting strategy to pick the most discriminative combination of them. This is further enhanced by our third contribution, that is to adapt the boosting strategy to enable sharing of binary features among different ferns, yielding high recognition rates at a low computational cost. And finally, we show that training can be performed online, for sequentially arriving images. Overall, the resulting classifier can be very efficiently trained, densely evaluated for all image locations in about 0.1 seconds, and provides detection rates similar to competing approaches that require expensive and significantly slower processing times. We demonstrate the effectiveness of our approach by thorough experimentation in publicly available datasets in which we compare against state-of-the-art, and for tasks of both 2D detection and 3D multi-view estimation., This work was partially supported by the Catalan Agency for Management of University and Research Grants for the Consolidated Group VIS (2014 SGR 897). A.S. acknowledges support also from project ColRobTransp (DPI2016-78957-R), and J.A. and F.M. from project RobInstruct (TIN2014-58178-R), both funded by the Spanish Ministry of Economy, Industry and Competitiveness. J.A., A.S., and F.M. were also partially supported by the European Union projects AEROARMS (H2020-ICT-2014-1-644271) and LOGIMATIC (H2020-Galileo-2015-1-687534)., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

Low resolution lidar-based multi object tracking for driving applications

  • Pino, Iván del
  • Vaquero, Victor
  • Masini, Beatrice
  • Solà, Joan
  • Moreno-Noguer, Francesc
  • Sanfeliu, Alberto
  • Andrade-Cetto, Juan
Trabajo presentado a la Third Iberian Robotics Conference, celebrada en Sevilla del 22 al 24 de noviembre de 2017., Vehicle detection and tracking in real scenarios are key com- ponents to develop assisted and autonomous driving systems. Lidar sen- sors are specially suitable for this task, as they bring robustness to harsh weather conditions while providing accurate spatial information. How- ever, the resolution provided by point cloud data is very scarce in com- parison to camera images. In this work we explore the possibilities of Deep Learning (DL) methodologies applied to low resolution 3D lidar sensors such as the Velodyne VLP-16 (PUCK), in the context of vehicle detection and tracking. For this purpose we developed a lidar-based sys- tem that uses a Convolutional Neural Network (CNN), to perform point- wise vehicle detection using PUCK data, and Multi-Hypothesis Extended Kalman Filters (MH-EKF), to estimate the actual position and veloci- ties of the detected vehicles. Comparative studies between the proposed lower resolution (VLP-16) tracking system and a high-end system, using Velodyne HDL-64, were carried out on the Kitti Tracking Benchmark dataset. Moreover, to analyze the influence of the CNN-based vehicle detection approach, comparisons were also performed with respect to the geometric-only detector. The results demonstrate that the proposed low resolution Deep Learning architecture is able to successfully accom- plish the vehicle detection task, outperforming the geometric baseline approach. Moreover, it has been observed that our system achieves a similar tracking performance to the high-end HDL-64 sensor at close range. On the other hand, at long range, detection is limited to half the distance of the higher-end sensor., This work has been supported by the Spanish Ministry of Economy and Competitiveness projects ROBINSTRUCT (TIN2014-58178-R) and COLROBTRANSP (DPI2016-78957-R), by the Spanish Ministry of Education FPU grant (FPU15/04446), the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI (MDM-2016-0656) and by the EU H2020 project LOGIMATIC (H2020-Galileo-2015-1-687534)., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

Deconvolutional networks for point-cloud vehicle detection and tracking in driving scenarios

  • Vaquero, Victor
  • Pino, Iván del
  • Moreno-Noguer, Francesc
  • Solà, Joan
  • Sanfeliu, Alberto
  • Andrade-Cetto, Juan
Trabajo presentado a la European Conference on Mobile Robots (ECMR), celebrada en Paris (Francia) del 6 al 8 de septiembre de 2017., Vehicle detection and tracking is a core ingredient for developing autonomous driving applications in urban scenarios. Recent image-based Deep Learning (DL) techniques are obtaining breakthrough results in these perceptive tasks. However, DL research has not yet advanced much towards processing 3D point clouds from lidar range-finders. These sensors are very common in autonomous vehicles since, despite not providing as semantically rich information as images, their performance is more robust under harsh weather conditions than vision sensors. In this paper we present a full vehicle detection and tracking system that works with 3D lidar information only. Our detection step uses a Convolutional Neural Network (CNN) that receives as input a featured representation of the 3D information provided by a Velodyne HDL-64 sensor and returns a per-point classification of whether it belongs to a vehicle or not. The classified point cloud is then geometrically processed to generate observations for a multi-object tracking system implemented via a number of Multi-Hypothesis Extended Kalman Filters (MH-EKF) that estimate the position and velocity of the surrounding vehicles. The system is thoroughly evaluated on the KITTI tracking dataset, and we show the performance boost provided by our CNN-based vehicle detector over a standard geometric approach. Our lidar-based approach uses about a 4% of the data needed for an image-based detector with similarly competitive results., This work has been supported by the Spanish Ministry of Economy and Competitiveness projects ROBINSTRUCT (TIN2014-58178-R) and COLROBTRANSP (DPI2016-78957-R), by the Spanish Ministry of Education FPU grant (FPU15/04446), the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI MDM-2016-0656) and by the EU H2020 project LOGIMATIC (H2020-Galileo-015-1-687534)., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

Factor descent optimization for sparsification in graph SLAM

  • Vallvé, Joan
  • Solà, Joan
  • Andrade-Cetto, Juan
Trabajo presentado a la European Conference on Mobile Robots (ECMR), celebrada en Paris (Francia) del 6 al 8 de septiembre de 2017., In the context of graph-based simultaneous localization and mapping, node pruning consists in removing a subset of nodes from the graph, while keeping the graph¿s information content as close as possible to the original. One often tackles this problem locally by isolating the Markov blanket sub-graph of a node, marginalizing this node and sparsifying the dense result. It means computing an approximation with a new set of factors. For a given approximation topology, the factors¿ mean and covariance that best approximate the original distribution can be obtained through minimization of the Kullback-Liebler divergence. For simple topologies such as Chow-Liu trees, there is a closed form for the optimal solution. However, a tree is oftentimes too sparse to explain some graphs. More complex topologies require nonlinear iterative optimization. In the present paper we propose Factor Descent, a new iterative optimization method to sparsify the dense result of node marginalization, which works by iterating factor by factor. We also provide a thorough comparison of our approach with state-of-the-art methods in real world datasets with regards to the obtained solution and convergence rates., This work has been supported by the Spanish Ministry of Economy and Competitiveness under Project ROBINSTRUCT (TIN2014-58178-R), by the EU H2020 Project LOGIMATIC (H2020-Galileo-2015-1-687534) and by the María de Maeztu Seal of Excellence to IRI MDM-2016-0656., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

Hallucinating dense optical flow from sparse lidar for autonomous vehicles

  • Vaquero, Victor
  • Sanfeliu, Alberto
  • Moreno-Noguer, Francesc
Trabajo presentado en la 24th International Conference on Pattern Recognition (ICPR), celebrada en Beijing (China), del 20 al 24 de agosto de 2018, In this paper we propose a novel approach to estimate dense optical flow from sparse lidar data acquired on an autonomous vehicle. This is intended to be used as a drop-in replacement of any image-based optical flow system when images are not reliable due to e.g. adverse weather conditions or at night. In order to infer high resolution 2D flows from discrete range data we devise a three-block architecture of multiscale filters that combines multiple intermediate objectives, both in the lidar and image domain. To train this network we introduce a dataset with approximately 20K lidar samples of the Kitti dataset which we have augmented with a pseudo ground-truth image-based optical flow computed using FlowNet2. We demonstrate the effectiveness of our approach on Kitti, and show that despite using the low-resolution and sparse measurements of the lidar, we can regress dense optical flow maps which are at par with those estimated with image-based methods., This work has been supported by the Spanish Ministry of Economy, Industry and Competitiveness projects COLROBTRANSP (DPI2016-78957-R), HuMoUR (TIN2017-90086-R), the Spanish State Research Agency through the Maria de Maeztu Seal of Excellence (MDM-2016- 0656), and the EU project LOGIMATIC (H2020-Galileo2015-1-687534). We also thank Nvidia for hardware donation under the GPU Grant Program., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Otros (Other)

LOGIMATIC Autonomous Navigation

  • Rizzo, Carlos
  • Andrade-Cetto, Juan
  • Solà, Joan
  • Vallvé, Joan
  • Serrano López, Daniel
  • González Villodres, Jesús Pablo
  • Tsitsamis, Dimitris
  • Rodríguez, Hugo
Advanced autonomous navigation solution based on the integration of Global Navigation Satellite Systems (GNSS) and sensors onboard straddle carrier vehicles for port automation., We acknowledge financial support from the Spanish Science Ministry (MINECO) through projects EB-SLAM (DPI2017-89564-P) and María de Maeztu Seal of Excellence (MDM-2016-0656); and from the European GNSS Agency Grant LOGIMATIC (H2020-Galileo-2015-1-687534)., No
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Pose-graph SLAM sparsification using factor descent

  • Vallvé, Joan
  • Solà, Joan
  • Andrade-Cetto, Juan
Since state of the art simultaneous localization and mapping (SLAM) algorithms are not constant time, it is often necessary to reduce the problem size while keeping as much of the original graph's information content. In graph SLAM, the problem is reduced by removing nodes and rearranging factors. This is normally faced locally: after selecting a node to be removed, its Markov blanket sub-graph is isolated, the node is marginalized and its dense result is sparsified. The aim of sparsification is to compute an approximation of the dense and non-relinearizable result of node marginalization with a new set of factors. Sparsification consists on two processes: building the topology of new factors, and finding the optimal parameters that best approximate the original dense distribution. This best approximation can be obtained through minimization of the Kullback–Liebler divergence between the two distributions. Using simple topologies such as Chow–Liu trees, there is a closed form for the optimal solution. However, a tree is oftentimes too sparse and produces bad distribution approximations. On the contrary, more populated topologies require nonlinear iterative optimization. In the present paper, the particularities of pose-graph SLAM are exploited for designing new informative topologies and for applying the novel factor descent iterative optimization method for sparsification. Several experiments are provided comparing the proposed topology methods and factor descent optimization with state-of-the-art methods in synthetic and real datasets with regards to approximation accuracy and computational cost., This work has been supported by the Spanish Ministry of Economy and Competitiveness Project EB-SLAM [DPI2017-89564-P]; by the EU H2020 Project LOGIMATIC [H2020-Galileo-2015-1-687534]; and by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI [MDM-2016-0656].
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

Timed-elastic smooth curve optimization for mobile-base planning

  • Deray, Jeremie
  • Magyar, Bence
  • Solà, Joan
  • Andrade-Cetto, Juan
Trabajo presentado en la IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), celebrada en Macao (China), del 3 al 8 de noviembre de 2019, This paper proposes the use of piecewise Cn smooth curve for mobile-base motion planning and control, coined Timed-Elastic Smooth Curve (TESC) planner. Based on a Timed-Elastic Band, the problem is defined so that the trajectory lies on a spline in SE(2) with non-vanishing n-th derivatives at every point. Formulated as a multi-objective nonlinear optimization problem, it allows imposing soft constraints such as collision-avoidance, velocity, acceleration and jerk limits, and more. The planning process is realtime-capable allowing the robot to navigate in dynamic complex scenarios. The proposed method is compared against the state-of-the-art in various scenarios. Results show that trajectories generated by the TESC planner have smaller average acceleration and are more efcient in terms of total curvature and pseudo-kinetic energy while being produced with more consistency than state-of-the-art planners do., This work was partially supported by the EU H2020 project LOGIMATIC (H2020-Galileo-2015-1-687534), and by the Spanish State Research Agency through projects EB-SLAM (DPI2017-89564-P) and the Mar´ıa de Maeztu Seal of Excellence to IRI (MDM-2016-0656).
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Contribución de congreso (ConferenceOutput)

Joint on-manifold self-calibration of odometry model and sensor extrinsics using pre-integration

  • Deray, Jeremie
  • Solà, Joan
  • Andrade-Cetto, Juan
Trabajo presentado en la European Conference on Mobile Robots (ECMR), celebrada en Praga (República Checa), del 4 al 6 de septiembre de 2019, This paper describes a self-calibration procedure that jointly estimates the extrinsic parameters of an exteroceptive sensor able to observe ego-motion, and the intrinsic parameters of an odometry motion model, consisting of wheel radii and wheel separation. We use iterative nonlinear onmanifold optimization with a graphical representation of the state, and resort to an adaptation of the pre-integration theory, initially developed for the IMU motion sensor, to be applied to the differential drive motion model. For this, we describe the construction of a pre-integrated factor for the differential drive motion model, which includes the motion increment, its covariance, and a first-order approximation of its dependence with the calibration parameters. As the calibration parameters change at each solver iteration, this allows a posteriori factor correction without the need of re-integrating the motion data. We validate our proposal in simulations and on a real robot and show the convergence of the calibration towards the true values of the parameters. It is then tested online in simulation and is shown to accommodate to variations in the calibration parameters when the vehicle is subject to physical changes such as loading and unloading a freight., This work has been supported by the Spanish Ministry of Science, Innovation, and Universities project EB-SLAM (DPI2017-89564-P), by the EU H2020 project LOGIMATIC (H2020-Galileo-2015-1-687534) and by the Spanish State Research Agency through the Maria de Maeztu Seal of Excellence to IRI MDM-2016-0656. J. Deray acknowledges support from the Industrial Doctorate Program of the Catalan Agency for Management of University and Research Grants.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Dual-Branch CNNs for Vehicle Detection and Tracking on LiDAR Data

  • Vaquero, Victor
  • Pino, Iván del
  • Moreno-Noguer, Francesc
  • Solà, Joan
  • Sanfeliu, Alberto
  • Andrade-Cetto, Juan
We present a novel vehicle detection and tracking system that works solely on 3D LiDAR information. Our approach segments vehicles using a dual-view representation of the 3D LiDAR point cloud on two independently trained convolutional neural networks, one for each view. A bounding box growing algorithm is applied to the fused output of the networks to properly enclose the segmented vehicles. Bounding boxes are grown using a probabilistic method that takes into account also occluded areas. The final vehicle bounding boxes act as observations for a multi-hypothesis tracking system which allows to estimate the position and velocity of the observed vehicles. We thoroughly evaluate our system on the KITTI benchmarks both for detection and tracking separately and show that our dual-branch classifier consistently outperforms previous single-branch approaches, improving or directly competing to other state of the art LiDAR-based methods., This work is supported by the EU project LOGIMATIC (H2020-Galileo2015-1-687534), the Spanish State Research Agency through the Mar´ıa de Maeztu Seal of Excellence to IRI (MDM-2016-0656), the ColRobTransp project (DPI2016-78957-R AEI/FEDER EU), the EB-SLAM project (DPI2017-89564-P), and the FPU grant FPU15/04446.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Wolf: A Modular Estimation Framework for Robotics Based on Factor Graphs

  • Solà, Joan
  • Vallvé, Joan
  • Casalsm Joaquim
  • Deray, Jeremie
  • Fourmy, Mederic
  • Atchuthan, Dinesh
  • Corominas Murtra, Andreu
  • Andrade-Cetto, Juan
This letter introduces Wolf, a C++ estimation framework based on factor graphs and targeted at mobile robotics. Wolf can be used beyond SLAM to handle self-calibration, model identification, or the observation of dynamic quantities other than localization. The architecture of Wolf allows for a modular yet tightly-coupled estimator. Modularity is enhanced via reusable plugins that are loaded at runtime depending on application setup. This setup is achieved conveniently through YAML files, allowing users to configure a wide range of applications without the need of writing or compiling code. Most procedures are coded as abstract algorithms in base classes with varying levels of specialization. Overall, all these assets allow for coherent processing and favor code re-usability and scalability. Wolf can be used with ROS, and is made publicly available and open to collaboration., This work was partially supported by the EU H2020 projects LOGIMATIC (Galileo- 2015-1-687534), GAUSS (Galileo-2017-1-776293), TERRINET (INFRAIA- 2017-1-730994), and MEMMO (ICT-25-2017-1-780684); and by the Spanish State Research Agency through projects EB-SLAM (DPI2017-89564- P), EBCON (PID2020-119244GB-I00), and the Mar´ıa de Maeztu Seal of Excellence to IRI (MDM-2016-0656).
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Global model with local interpretation for dynamic shape reconstruction

  • Agudo Martínez, Antonio|||0000-0001-6845-4998
  • Moreno-Noguer, Francesc
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., The most standard approach to resolve the inherent ambiguities of the non-rigid structure from motion problem is using low-rank models that approximate deforming shapes by a linear combination of rigid basis. These models are typically global, i.e., each shape basis contributes equally to all points of the surface. While this approach has been shown effective to represent smooth deformations, its performance degrades for surfaces composed of various regions, each following a different deformation rule. Piecewise methods attempt to capture this type of behavior by locally modeling surface patches, although they subsequently require enforcing global constraints to assemble back the patches. In this paper we propose an approach that combines the best of global and local models: it locally considers low-rank models but, by construction, does not need to impose global constraints to guarantee local patch continuity. We achieve this by a simple expectation maximization strategy that besides learning global shape bases, it locally adapts their contribution to each specific surface region. Furthermore, as a side contribution, in order to split the surface into different local patches, we propose a novel physically-based mesh segmentation approach that obeys an energy criterion. The complete framework is evaluated in both synthetic and real datasets, and shows an improved performance to competing methods., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Pose-graph SLAM sparsification using factor descent

  • Vallvé Navarro, Joan|||0000-0002-1029-356X
  • Solà Ortega, Joan|||0000-0002-2933-3381
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
Since state of the art simultaneous localization and mapping (SLAM) algorithms are not constant time, it is often necessary to reduce the problem size while keeping as much of the original graph’s information content. In graph SLAM, the problem is reduced by removing nodes and rearranging factors. This is normally faced locally: after selecting a node to be removed, its Markov blanket sub-graph is isolated, the node is marginalized and its dense result is sparsified. The aim of sparsification is to compute an approximation of the dense and non-relinearizable result of node marginalization with a new set of factors. Sparsification consists on two processes: building the topology of new factors, and finding the optimal parameters that best approximate the original dense distribution. This best approximation can be obtained through minimization of the Kullback-Liebler divergence between the two distributions. Using simple topologies such as Chow-Liu trees, there is a closed form for the optimal solution. However, a tree is oftentimes too sparse and produces bad distribution approximations. On the contrary, more populated topologies require nonlinear iterative optimization. In the present paper, the particularities of pose-graph SLAM are exploited for designing new informative topologies and for applying the novel factor descent iterative optimization method for sparsification. Several experiments are provided comparing the proposed topology methods and factor descent optimization with state-of-the-art methods in synthetic and real datasets with regards to approximation accuracy and computational cost., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Graph SLAM sparsification with populated topologies using factor descent optimization

  • Vallvé Navarro, Joan|||0000-0002-1029-356X
  • Solà Ortega, Joan|||0000-0002-2933-3381
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Current solutions to the simultaneous localization and mapping (SLAM) problem approach it as the optimization of a graph of geometric constraints. Scalability is achieved by reducing the size of the graph, usually in two phases. First, some selected nodes in the graph are marginalized and then, the dense and non-relinearizable result is sparsified. The sparsified network has a new set of relinearizable factors and is an approximation to the original dense one. Sparsification is typically approached as a Kullback-Liebler divergence (KLD) minimization between the dense marginalization result and the new set of factors. For a simple topology of the new factors, such as a tree, there is a closed form optimal solution. However, more populated topologies can achieve a much better approximation because more information can be encoded, although in that case iterative optimization is needed to solve the KLD minimization. Iterative optimization methods proposed by the state-of-art sparsification require parameter tuning which strongly affect their convergence. In this paper, we propose factor descent and non-cyclic factor descent, two simple algorithms for SLAM sparsification that match the state-of-art methods without any parameters to be tuned. The proposed methods are compared against the state of the art with regards to accuracy and CPU time, in both synthetic and real world datasets., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Factor descent optimization for sparsification in graph SLAM

  • Vallvé Navarro, Joan|||0000-0002-1029-356X
  • Solà Ortega, Joan|||0000-0002-2933-3381
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., In the context of graph-based simultaneous localization and mapping, node pruning consists in removing a subset of nodes from the graph, while keeping the graph’s information content as close as possible to the original. One often tackles this problem locally by isolating the Markov blanket sub-graph of a node, marginalizing this node and sparsifying the dense result. It means computing an approximation with a new set of factors. For a given approximation topology, the factors’ mean and covariance that best approximate the original distribution can be obtained through minimization of the Kullback-Liebler divergence. For simple topologies such as Chow-Liu trees, there is a closed form for the optimal solution. However, a tree is oftentimes too sparse to explain some graphs. More complex topologies require nonlinear iterative optimization. In the present paper we propose Factor Descent, a new iterative optimization method to sparsify the dense result of node marginalization, which works by iterating factor by factor. We also provide a thorough comparison of our approach with state-of-the-art methods in real world datasets with regards to the obtained solution and convergence rates., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Timed-elastic smooth curve optimization for mobile-base motion planning

  • Deray, Jeremie
  • Magyar, Bence
  • Solà Ortega, Joan|||0000-0002-2933-3381
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., This paper proposes the use of piecewise Cn smooth curve for mobile-base motion planning and control, coined Timed-Elastic Smooth Curve (TESC) planner. Based on a Timed-Elastic Band, the problem is defined so that the trajectory lies on a spline in SE(2) with non-vanishing n-th derivatives at every point. Formulated as a multi-objective nonlinear optimization problem, it allows imposing soft constraints such as collision-avoidance, velocity, acceleration and jerk limits, and more. The planning process is realtime-capable allowing the robot to navigate in dynamic complex scenarios. The proposed method is compared against the state-of-the-art in various scenarios. Results show that trajectories generated by the TESC planner have smaller average acceleration and are more efcient in terms of total curvature and pseudo-kinetic energy while being produced with more consistency than state-of-the-art planners do., This work was partially supported by the EU H2020 project LOGIMATIC(H2020-Galileo-2015-1-687534), and by the Spanish State Research Agencythrough projects EB-SLAM (DPI2017-89564-P) and the María de Maeztu Seal of Excellence to IRI (MDM-2016-065), Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Word ordering and document adjacency for large loop closure detection in 2D laser maps

  • Deray, Jeremie
  • Solà Ortega, Joan|||0000-0002-2933-3381
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works, We address in this paper the problem of loop closure detection for laser-based simultaneous localization and mapping (SLAM) of very large areas. Consistent with the state of the art, the map is encoded as a graph of poses, and to cope with very large mapping capabilities, loop closures are asserted by comparing the features extracted from a query laser scan against a previously acquired corpus of scan features using a bag-ofwords (BoW) scheme. Two contributions are here presented. First, to benefit from the graph topology, feature frequency scores in the BoW are computed not only for each individual scan but also from neighboring scans in the SLAM graph. This has the effect of enforcing neighbor relational information during document matching. Secondly, a weak geometric check that takes into account feature ordering and occlusions is introduced that substantially improves loop closure detection performance. The two contributions are evaluated both separately and jointly on four common SLAM datasets, and are shown to improve the state-of-the-art performance both in terms of precision and recall in most of the cases. Moreover, our current implementation is designed to work at nearly frame rate, allowing loop closure query resolution at nearly 22 Hz for the best case scenario and 2 Hz for the worst case scenario., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Joint on-manifold self-calibration of odometry model and sensor extrinsics using pre-integration

  • Deray, Jeremie
  • Solà Ortega, Joan|||0000-0002-2933-3381
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., This paper describes a self-calibration procedure that jointly estimates the extrinsic parameters of an exteroceptive sensor able to observe ego-motion, and the intrinsic parameters of an odometry motion model, consisting of wheel radii and wheel separation. We use iterative nonlinear onmanifold optimization with a graphical representation of the state, and resort to an adaptation of the pre-integration theory, initially developed for the IMU motion sensor, to be applied to the differential drive motion model. For this, we describe the construction of a pre-integrated factor for the differential drive motion model, which includes the motion increment, its covariance, and a first-order approximation of its dependence with the calibration parameters. As the calibration parameters change at each solver iteration, this allows a posteriori factor correction without the need of re-integrating the motion data. We validate our proposal in simulations and on a real robot and show the convergence of the calibration towards the true values of the parameters. It is then tested online in simulation and is shown to accommodate to variations in the calibration parameters when the vehicle is subject to physical changes such as loading and unloading a freight., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Dual-branch CNNs for vehicle detection and tracking on LiDAR data

  • Vaquero Gómez, Víctor|||0000-0003-2516-2551
  • del Pino Bastida, Iván
  • Moreno-Noguer, Francesc
  • Solà Ortega, Joan|||0000-0002-2933-3381
  • Sanfeliu Cortés, Alberto|||0000-0003-3868-9678
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., We present a novel vehicle detection and tracking system that works solely on 3D LiDAR information. Our approach segments vehicles using a dual-view representation of the 3D LiDAR point cloud on two independently trained convolutional neural networks, one for each view. A bounding box growing algorithm is applied to the fused output of the networks to properly enclose the segmented vehicles. Bounding boxes are grown using a probabilistic method that takes into account also occluded areas. The final vehicle bounding boxes act as observations for a multi-hypothesis tracking system which allows to estimate the position and velocity of the observed vehicles. We thoroughly evaluate our system on the KITTI benchmarks both for detection and tracking separately and show that our dual-branch classifier consistently outperforms previous single-branch approaches, improving or directly competing to other state of the art LiDAR-based methods., This work was supported in part by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI under Grant MDM-2016-0656, in part by the ColRobTransp Project under Grant DPI2016-78957-RAEI/FEDER EU, in part by the EB-SLAM Project under Grant DPI2017-89564-P, and in part by the FPU Grant under Grant FPU15/04446. The Associate Editor for this article was Z. Duric. (Víctor Vaquero and Iván del Pino contributed equally to this work.) (Corresponding author: Víctor Vaquero.), Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Deconvolutional networks for point-cloud vehicle detection and tracking in driving scenarios

  • Vaquero Gómez, Víctor|||0000-0003-2516-2551
  • del Pino Bastida, Iván
  • Moreno-Noguer, Francesc
  • Solà Ortega, Joan|||0000-0002-2933-3381
  • Sanfeliu Cortés, Alberto|||0000-0003-3868-9678
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Vehicle detection and tracking is a core ingredient for developing autonomous driving applications in urban scenarios. Recent image-based Deep Learning (DL) techniques are obtaining breakthrough results in these perceptive tasks. However, DL research has not yet advanced much towards processing 3D point clouds from lidar range-finders. These sensors are very common in autonomous vehicles since, despite not providing as semantically rich information as images, their performance is more robust under harsh weather conditions than vision sensors. In this paper we present a full vehicle detection and tracking system that works with 3D lidar information only. Our detection step uses a Convolutional Neural Network (CNN) that receives as input a featured representation of the 3D information provided by a Velodyne HDL-64 sensor and returns a per-point classification of whether it belongs to a vehicle or not. The classified point cloud is then geometrically processed to generate observations for a multi-object tracking system implemented via a number of Multi-Hypothesis Extended Kalman Filters (MH-EKF) that estimate the position and velocity of the surrounding vehicles. The system is thoroughly evaluated on the KITTI tracking dataset, and we show the performance boost provided by our CNN-based vehicle detector over a standard geometric approach. Our lidar-based approach uses about a 4% of the data needed for an image-based detector with similarly competitive results., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Hallucinating dense optical flow from sparse lidar for autonomous vehicles

  • Vaquero Gómez, Víctor|||0000-0003-2516-2551
  • Sanfeliu Cortés, Alberto|||0000-0003-3868-9678
  • Moreno-Noguer, Francesc
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., In this paper we propose a novel approach to estimate dense optical flow from sparse lidar data acquired on an autonomous vehicle. This is intended to be used as a drop-in replacement of any image-based optical flow system when images are not reliable due to e.g. adverse weather conditions or at night. In order to infer high resolution 2D flows from discrete range data we devise a three-block architecture of multiscale filters that combines multiple intermediate objectives, both in the lidar and image domain. To train this network we introduce a dataset with approximately 20K lidar samples of the Kitti dataset which we have augmented with a pseudo ground-truth image-based optical flow computed using FlowNet2. We demonstrate the effectiveness of our approach on Kitti, and show that despite using the low-resolution and sparse measurements of the lidar, we can regress dense optical flow maps which are at par with those estimated with image-based methods., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Artículo científico (JournalArticle)

Boosted Random ferns for object detection

  • Villamizar Vergel, Michael Alejandro
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
  • Sanfeliu Cortés, Alberto|||0000-0003-3868-9678
  • Moreno-Noguer, Francesc
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., In this paper we introduce the Boosted Random Ferns (BRFs) to rapidly build discriminative classifiers for learning and detecting object categories. At the core of our approach we use standard random ferns, but we introduce four main innovations that let us bring ferns from an instance to a category level, and still retain efficiency. First, we define binary features on the histogram of oriented gradients-domain (as opposed to intensity-), allowing for a better representation of intra-class variability. Second, both the positions where ferns are evaluated within the sliding window, and the location of the binary features for each fern are not chosen completely at random, but instead we use a boosting strategy to pick the most discriminative combination of them. This is further enhanced by our third contribution, that is to adapt the boosting strategy to enable sharing of binary features among different ferns, yielding high recognition rates at a low computational cost. And finally, we show that training can be performed online, for sequentially arriving images. Overall, the resulting classifier can be very efficiently trained, densely evaluated for all image locations in about 0.1 seconds, and provides detection rates similar to competing approaches that require expensive and significantly slower processing times. We demonstrate the effectiveness of our approach by thorough experimentation in publicly available datasets in which we compare against state-of-the-art, and for tasks of both 2D detection and 3D multi-view estimation., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

DUST: dual union of spatio-temporal subspaces for monocular multiple object 3D reconstruction

  • Agudo Martínez, Antonio|||0000-0001-6845-4998
  • Moreno-Noguer, Francesc
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., We present an approach to reconstruct the 3D shape of multiple deforming objects from incomplete 2D trajectories acquired by a single camera. Additionally, we simultaneously provide spatial segmentation (i.e., we identify each of the objects in every frame) and temporal clustering (i.e., we split the sequence into primitive actions). This advances existing work, which only tackled the problem for one single object and non-occluded tracks. In order to handle several objects at a time from partial observations, we model point trajectories as a union of spatial and temporal subspaces, and optimize the parameters of both modalities, the non-observed point tracks and the 3D shape via augmented Lagrange multipliers. The algorithm is fully unsupervised and results in a formulation which does not need initialization. We thoroughly validate the method on challenging scenarios with several human subjects performing different activities which involve complex motions and close interaction. We show our approach achieves state-of-the-art 3D reconstruction results, while it also provides spatial and temporal segmentation., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publication Contribución de congreso (ConferenceOutput)

Low resolution lidar-based multi object tracking for driving applications

  • del Pino Bastida, Iván
  • Vaquero Gómez, Víctor|||0000-0003-2516-2551
  • Massini, Beatrice
  • Solà Ortega, Joan|||0000-0002-2933-3381
  • Moreno-Noguer, Francesc
  • Sanfeliu Cortés, Alberto|||0000-0003-3868-9678
  • Andrade-Cetto, Juan|||0000-0002-6354-8941
The final publication is available at link.springer.com, Vehicle detection and tracking in real scenarios are key com- ponents to develop assisted and autonomous driving systems. Lidar sen- sors are specially suitable for this task, as they bring robustness to harsh weather conditions while providing accurate spatial information. How- ever, the resolution provided by point cloud data is very scarce in com- parison to camera images. In this work we explore the possibilities of Deep Learning (DL) methodologies applied to low resolution 3D lidar sensors such as the Velodyne VLP-16 (PUCK), in the context of vehicle detection and tracking. For this purpose we developed a lidar-based sys- tem that uses a Convolutional Neural Network (CNN), to perform point- wise vehicle detection using PUCK data, and Multi-Hypothesis Extended Kalman Filters (MH-EKF), to estimate the actual position and veloci- ties of the detected vehicles. Comparative studies between the proposed lower resolution (VLP-16) tracking system and a high-end system, using Velodyne HDL-64, were carried out on the Kitti Tracking Benchmark dataset. Moreover, to analyze the influence of the CNN-based vehicle detection approach, comparisons were also performed with respect to the geometric-only detector. The results demonstrate that the proposed low resolution Deep Learning architecture is able to successfully accom- plish the vehicle detection task, outperforming the geometric baseline approach. Moreover, it has been observed that our system achieves a similar tracking performance to the high-end HDL-64 sensor at close range. On the other hand, at long range, detection is limited to half the distance of the higher-end sensor., Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

DUST: dual union of spatio-temporal subspaces for monocular multiple object 3D reconstruction

  • Agudo Martínez, Antonio
  • Moreno-Noguer, Francesc
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., We present an approach to reconstruct the 3D shape of multiple deforming objects from incomplete 2D trajectories acquired by a single camera. Additionally, we simultaneously provide spatial segmentation (i.e., we identify each of the objects in every frame) and temporal clustering (i.e., we split the sequence into primitive actions). This advances existing work, which only tackled the problem for one single object and non-occluded tracks. In order to handle several objects at a time from partial observations, we model point trajectories as a union of spatial and temporal subspaces, and optimize the parameters of both modalities, the non-observed point tracks and the 3D shape via augmented Lagrange multipliers. The algorithm is fully unsupervised and results in a formulation which does not need initialization. We thoroughly validate the method on challenging scenarios with several human subjects performing different activities which involve complex motions and close interaction. We show our approach achieves state-of-the-art 3D reconstruction results, while it also provides spatial and temporal segmentation., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Boosted Random ferns for object detection

  • Villamizar Vergel, Michael Alejandro
  • Andrade-Cetto, Juan
  • Sanfeliu Cortés, Alberto
  • Moreno-Noguer, Francesc
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., In this paper we introduce the Boosted Random Ferns (BRFs) to rapidly build discriminative classifiers for learning and detecting object categories. At the core of our approach we use standard random ferns, but we introduce four main innovations that let us bring ferns from an instance to a category level, and still retain efficiency. First, we define binary features on the histogram of oriented gradients-domain (as opposed to intensity-), allowing for a better representation of intra-class variability. Second, both the positions where ferns are evaluated within the sliding window, and the location of the binary features for each fern are not chosen completely at random, but instead we use a boosting strategy to pick the most discriminative combination of them. This is further enhanced by our third contribution, that is to adapt the boosting strategy to enable sharing of binary features among different ferns, yielding high recognition rates at a low computational cost. And finally, we show that training can be performed online, for sequentially arriving images. Overall, the resulting classifier can be very efficiently trained, densely evaluated for all image locations in about 0.1 seconds, and provides detection rates similar to competing approaches that require expensive and significantly slower processing times. We demonstrate the effectiveness of our approach by thorough experimentation in publicly available datasets in which we compare against state-of-the-art, and for tasks of both 2D detection and 3D multi-view estimation., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Global model with local interpretation for dynamic shape reconstruction

  • Agudo Martínez, Antonio
  • Moreno-Noguer, Francesc
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., The most standard approach to resolve the inherent ambiguities of the non-rigid structure from motion problem is using low-rank models that approximate deforming shapes by a linear combination of rigid basis. These models are typically global, i.e., each shape basis contributes equally to all points of the surface. While this approach has been shown effective to represent smooth deformations, its performance degrades for surfaces composed of various regions, each following a different deformation rule. Piecewise methods attempt to capture this type of behavior by locally modeling surface patches, although they subsequently require enforcing global constraints to assemble back the patches. In this paper we propose an approach that combines the best of global and local models: it locally considers low-rank models but, by construction, does not need to impose global constraints to guarantee local patch continuity. We achieve this by a simple expectation maximization strategy that besides learning global shape bases, it locally adapts their contribution to each specific surface region. Furthermore, as a side contribution, in order to split the surface into different local patches, we propose a novel physically-based mesh segmentation approach that obeys an energy criterion. The complete framework is evaluated in both synthetic and real datasets, and shows an improved performance to competing methods., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Dual-branch CNNs for vehicle detection and tracking on LiDAR data

  • Vaquero Gómez, Víctor
  • del Pino Bastida, Iván
  • Moreno-Noguer, Francesc
  • Solà Ortega, Joan
  • Sanfeliu Cortés, Alberto
  • Andrade-Cetto, Juan
© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., We present a novel vehicle detection and tracking system that works solely on 3D LiDAR information. Our approach segments vehicles using a dual-view representation of the 3D LiDAR point cloud on two independently trained convolutional neural networks, one for each view. A bounding box growing algorithm is applied to the fused output of the networks to properly enclose the segmented vehicles. Bounding boxes are grown using a probabilistic method that takes into account also occluded areas. The final vehicle bounding boxes act as observations for a multi-hypothesis tracking system which allows to estimate the position and velocity of the observed vehicles. We thoroughly evaluate our system on the KITTI benchmarks both for detection and tracking separately and show that our dual-branch classifier consistently outperforms previous single-branch approaches, improving or directly competing to other state of the art LiDAR-based methods., This work was supported in part by the Spanish State Research Agency through the María de Maeztu Seal of Excellence to IRI under Grant MDM-2016-0656, in part by the ColRobTransp Project under Grant DPI2016-78957-RAEI/FEDER EU, in part by the EB-SLAM Project under Grant DPI2017-89564-P, and in part by the FPU Grant under Grant FPU15/04446. The Associate Editor for this article was Z. Duric. (Víctor Vaquero and Iván del Pino contributed equally to this work.) (Corresponding author: Víctor Vaquero.), Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Pose-graph SLAM sparsification using factor descent

  • Vallvé Navarro, Joan
  • Solà Ortega, Joan
  • Andrade-Cetto, Juan
Since state of the art simultaneous localization and mapping (SLAM) algorithms are not constant time, it is often necessary to reduce the problem size while keeping as much of the original graph’s information content. In graph SLAM, the problem is reduced by removing nodes and rearranging factors. This is normally faced locally: after selecting a node to be removed, its Markov blanket sub-graph is isolated, the node is marginalized and its dense result is sparsified. The aim of sparsification is to compute an approximation of the dense and non-relinearizable result of node marginalization with a new set of factors. Sparsification consists on two processes: building the topology of new factors, and finding the optimal parameters that best approximate the original dense distribution. This best approximation can be obtained through minimization of the Kullback-Liebler divergence between the two distributions. Using simple topologies such as Chow-Liu trees, there is a closed form for the optimal solution. However, a tree is oftentimes too sparse and produces bad distribution approximations. On the contrary, more populated topologies require nonlinear iterative optimization. In the present paper, the particularities of pose-graph SLAM are exploited for designing new informative topologies and for applying the novel factor descent iterative optimization method for sparsification. Several experiments are provided comparing the proposed topology methods and factor descent optimization with state-of-the-art methods in synthetic and real datasets with regards to approximation accuracy and computational cost., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Low resolution lidar-based multi object tracking for driving applications

  • del Pino Bastida, Iván
  • Vaquero Gómez, Víctor
  • Massini, Beatrice
  • Solà Ortega, Joan
  • Moreno-Noguer, Francesc
  • Sanfeliu Cortés, Alberto
  • Andrade-Cetto, Juan
The final publication is available at link.springer.com, Vehicle detection and tracking in real scenarios are key com- ponents to develop assisted and autonomous driving systems. Lidar sen- sors are specially suitable for this task, as they bring robustness to harsh weather conditions while providing accurate spatial information. How- ever, the resolution provided by point cloud data is very scarce in com- parison to camera images. In this work we explore the possibilities of Deep Learning (DL) methodologies applied to low resolution 3D lidar sensors such as the Velodyne VLP-16 (PUCK), in the context of vehicle detection and tracking. For this purpose we developed a lidar-based sys- tem that uses a Convolutional Neural Network (CNN), to perform point- wise vehicle detection using PUCK data, and Multi-Hypothesis Extended Kalman Filters (MH-EKF), to estimate the actual position and veloci- ties of the detected vehicles. Comparative studies between the proposed lower resolution (VLP-16) tracking system and a high-end system, using Velodyne HDL-64, were carried out on the Kitti Tracking Benchmark dataset. Moreover, to analyze the influence of the CNN-based vehicle detection approach, comparisons were also performed with respect to the geometric-only detector. The results demonstrate that the proposed low resolution Deep Learning architecture is able to successfully accom- plish the vehicle detection task, outperforming the geometric baseline approach. Moreover, it has been observed that our system achieves a similar tracking performance to the high-end HDL-64 sensor at close range. On the other hand, at long range, detection is limited to half the distance of the higher-end sensor., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Graph SLAM sparsification with populated topologies using factor descent optimization

  • Vallvé Navarro, Joan
  • Solà Ortega, Joan
  • Andrade-Cetto, Juan
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Current solutions to the simultaneous localization and mapping (SLAM) problem approach it as the optimization of a graph of geometric constraints. Scalability is achieved by reducing the size of the graph, usually in two phases. First, some selected nodes in the graph are marginalized and then, the dense and non-relinearizable result is sparsified. The sparsified network has a new set of relinearizable factors and is an approximation to the original dense one. Sparsification is typically approached as a Kullback-Liebler divergence (KLD) minimization between the dense marginalization result and the new set of factors. For a simple topology of the new factors, such as a tree, there is a closed form optimal solution. However, more populated topologies can achieve a much better approximation because more information can be encoded, although in that case iterative optimization is needed to solve the KLD minimization. Iterative optimization methods proposed by the state-of-art sparsification require parameter tuning which strongly affect their convergence. In this paper, we propose factor descent and non-cyclic factor descent, two simple algorithms for SLAM sparsification that match the state-of-art methods without any parameters to be tuned. The proposed methods are compared against the state of the art with regards to accuracy and CPU time, in both synthetic and real world datasets., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Deconvolutional networks for point-cloud vehicle detection and tracking in driving scenarios

  • Vaquero Gómez, Víctor
  • del Pino Bastida, Iván
  • Moreno-Noguer, Francesc
  • Solà Ortega, Joan
  • Sanfeliu Cortés, Alberto
  • Andrade-Cetto, Juan
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Vehicle detection and tracking is a core ingredient for developing autonomous driving applications in urban scenarios. Recent image-based Deep Learning (DL) techniques are obtaining breakthrough results in these perceptive tasks. However, DL research has not yet advanced much towards processing 3D point clouds from lidar range-finders. These sensors are very common in autonomous vehicles since, despite not providing as semantically rich information as images, their performance is more robust under harsh weather conditions than vision sensors. In this paper we present a full vehicle detection and tracking system that works with 3D lidar information only. Our detection step uses a Convolutional Neural Network (CNN) that receives as input a featured representation of the 3D information provided by a Velodyne HDL-64 sensor and returns a per-point classification of whether it belongs to a vehicle or not. The classified point cloud is then geometrically processed to generate observations for a multi-object tracking system implemented via a number of Multi-Hypothesis Extended Kalman Filters (MH-EKF) that estimate the position and velocity of the surrounding vehicles. The system is thoroughly evaluated on the KITTI tracking dataset, and we show the performance boost provided by our CNN-based vehicle detector over a standard geometric approach. Our lidar-based approach uses about a 4% of the data needed for an image-based detector with similarly competitive results., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Joint on-manifold self-calibration of odometry model and sensor extrinsics using pre-integration

  • Deray, Jeremie
  • Solà Ortega, Joan
  • Andrade-Cetto, Juan
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., This paper describes a self-calibration procedure that jointly estimates the extrinsic parameters of an exteroceptive sensor able to observe ego-motion, and the intrinsic parameters of an odometry motion model, consisting of wheel radii and wheel separation. We use iterative nonlinear onmanifold optimization with a graphical representation of the state, and resort to an adaptation of the pre-integration theory, initially developed for the IMU motion sensor, to be applied to the differential drive motion model. For this, we describe the construction of a pre-integrated factor for the differential drive motion model, which includes the motion increment, its covariance, and a first-order approximation of its dependence with the calibration parameters. As the calibration parameters change at each solver iteration, this allows a posteriori factor correction without the need of re-integrating the motion data. We validate our proposal in simulations and on a real robot and show the convergence of the calibration towards the true values of the parameters. It is then tested online in simulation and is shown to accommodate to variations in the calibration parameters when the vehicle is subject to physical changes such as loading and unloading a freight., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Artículo científico (JournalArticle)

Word ordering and document adjacency for large loop closure detection in 2D laser maps

  • Deray, Jeremie
  • Solà Ortega, Joan
  • Andrade-Cetto, Juan
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works, We address in this paper the problem of loop closure detection for laser-based simultaneous localization and mapping (SLAM) of very large areas. Consistent with the state of the art, the map is encoded as a graph of poses, and to cope with very large mapping capabilities, loop closures are asserted by comparing the features extracted from a query laser scan against a previously acquired corpus of scan features using a bag-ofwords (BoW) scheme. Two contributions are here presented. First, to benefit from the graph topology, feature frequency scores in the BoW are computed not only for each individual scan but also from neighboring scans in the SLAM graph. This has the effect of enforcing neighbor relational information during document matching. Secondly, a weak geometric check that takes into account feature ordering and occlusions is introduced that substantially improves loop closure detection performance. The two contributions are evaluated both separately and jointly on four common SLAM datasets, and are shown to improve the state-of-the-art performance both in terms of precision and recall in most of the cases. Moreover, our current implementation is designed to work at nearly frame rate, allowing loop closure query resolution at nearly 22 Hz for the best case scenario and 2 Hz for the worst case scenario., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Factor descent optimization for sparsification in graph SLAM

  • Vallvé Navarro, Joan
  • Solà Ortega, Joan
  • Andrade-Cetto, Juan
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., In the context of graph-based simultaneous localization and mapping, node pruning consists in removing a subset of nodes from the graph, while keeping the graph’s information content as close as possible to the original. One often tackles this problem locally by isolating the Markov blanket sub-graph of a node, marginalizing this node and sparsifying the dense result. It means computing an approximation with a new set of factors. For a given approximation topology, the factors’ mean and covariance that best approximate the original distribution can be obtained through minimization of the Kullback-Liebler divergence. For simple topologies such as Chow-Liu trees, there is a closed form for the optimal solution. However, a tree is oftentimes too sparse to explain some graphs. More complex topologies require nonlinear iterative optimization. In the present paper we propose Factor Descent, a new iterative optimization method to sparsify the dense result of node marginalization, which works by iterating factor by factor. We also provide a thorough comparison of our approach with state-of-the-art methods in real world datasets with regards to the obtained solution and convergence rates., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Hallucinating dense optical flow from sparse lidar for autonomous vehicles

  • Vaquero Gómez, Víctor
  • Sanfeliu Cortés, Alberto
  • Moreno-Noguer, Francesc
© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., In this paper we propose a novel approach to estimate dense optical flow from sparse lidar data acquired on an autonomous vehicle. This is intended to be used as a drop-in replacement of any image-based optical flow system when images are not reliable due to e.g. adverse weather conditions or at night. In order to infer high resolution 2D flows from discrete range data we devise a three-block architecture of multiscale filters that combines multiple intermediate objectives, both in the lidar and image domain. To train this network we introduce a dataset with approximately 20K lidar samples of the Kitti dataset which we have augmented with a pseudo ground-truth image-based optical flow computed using FlowNet2. We demonstrate the effectiveness of our approach on Kitti, and show that despite using the low-resolution and sparse measurements of the lidar, we can regress dense optical flow maps which are at par with those estimated with image-based methods., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publication Contribución de congreso (ConferenceOutput)

Timed-elastic smooth curve optimization for mobile-base motion planning

  • Deray, Jeremie
  • Magyar, Bence
  • Solà Ortega, Joan
  • Andrade-Cetto, Juan
© 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., This paper proposes the use of piecewise Cn smooth curve for mobile-base motion planning and control, coined Timed-Elastic Smooth Curve (TESC) planner. Based on a Timed-Elastic Band, the problem is defined so that the trajectory lies on a spline in SE(2) with non-vanishing n-th derivatives at every point. Formulated as a multi-objective nonlinear optimization problem, it allows imposing soft constraints such as collision-avoidance, velocity, acceleration and jerk limits, and more. The planning process is realtime-capable allowing the robot to navigate in dynamic complex scenarios. The proposed method is compared against the state-of-the-art in various scenarios. Results show that trajectories generated by the TESC planner have smaller average acceleration and are more efcient in terms of total curvature and pseudo-kinetic energy while being produced with more consistency than state-of-the-art planners do., This work was partially supported by the EU H2020 project LOGIMATIC(H2020-Galileo-2015-1-687534), and by the Spanish State Research Agencythrough projects EB-SLAM (DPI2017-89564-P) and the María de Maeztu Seal of Excellence to IRI (MDM-2016-065), Peer Reviewed, Postprint (published version)