2011 IEEE/RSJ International Conference on Intelligent Robots and Systems

IROS Workshop on Perception and Navigation for Autonomous Vehicles in Human Environment

Full Day Workshop

September 30th, 2011 San Francisco, California, USA

Workshop Proceedings, Program

Contact : Professor Philippe Martinet
IRCCYN-CNRS Laboratory, Ecole Centrale de Nantes,
1 rue de la Noë
44321 Nantes Cedex 03, France
Phone: +33 237407975, Fax: +33 237406934,
Email: Philippe.Martinet@irccyn.ec-nantes.fr,
Home page: http://www.irccyn.ec-nantes.fr/~martinet



Organizers

Research Director Christian Laugier, INRIA, Emotion project, INRIA Rhône-Alpes, 655 Avenue de l'Europe, 38334 Saint Ismier Cedex, France, Phone: +33 4 7661 5222, Fax : +33 4 7661 5477, Email: Christian.Laugier@inrialpes.fr,
Home page: http://emotion.inrialpes.fr/laugier

Professor Philippe Martinet, LASMEA-CNRS Laboratory, Blaise Pascal University, Campus des Cezeaux, 63177 Aubiere, Cedex, France, Phone: +33 473 407 653, Sec : +33 473 407 261, Fax : +33 473 407 262, Email: martinet@lasmea.univ-bpclermont.fr,
Home page: http://wwwlasmea.univ-bpclermont.fr/Personnel/Philippe.Martinet

Professor Urbano Nunes, Department of Electrical and Computer Engineering of the Faculty of Sciences and Technology of University of Coimbra, 3030-290 Coimbra, Portugal, GABINETE 3A.10, Phone: +351 239 796 287, Fax: +351 239 406 672, Email: urbano@deec.uc.pt,
Home page: http://www.isr.uc.pt/~urbano

General Scope

Autonomous driving and navigation is a major research issue which would affect our lives in near future. The purpose of this workshop is to discuss topics related to the challenging problems of autonomous navigation and of driving assistance in open and dynamic environments. Technologies related to application fields such as unmanned outdoor vehicles or intelligent road vehicles will be considered from both the theoretical and technological point of views. Several research questions located on the cutting edge of the state of the art will be addressed. Among the many application areas that robotics is addressing, transportation of people and goods seem to be a domain that will dramatically benefit from intelligent automation. Fully automatic driving is emerging as the approach to dramatically improve efficiency while at the same time leading to the goal of zero fatalities. Theses new technologies can be applied efficiently for other application field such as unmanned vehicles, mobile service robots, or mobile devices for motion assistance to elderly or disable peoples. Technologies related to this area, such as autonomous outdoor vehicles, achievements, challenges and open questions would be presented.

Main Topics

  • Road scene understanding
  • Lane detection and lane keeping
  • Pedestrian and vehicle detection
  • Detection, tracking and classification
  • Feature extraction and feature selection
  • Cooperative techniques
  • Collision prediction and avoidance
  • Driver assistance systems
  • Collision prediction and avoidance
  • Environment perception, vehicle localization and autonomous navigation
  • Real-time perception and sensor fusion
  • SLAM in dynamic environments
  • Real-time motion planning in dynamic environments
  • 3D Modelling and reconstruction
  • Human-Robot Interaction
  • Behavior modeling and learning
  • Robust sensor-based 3D reconstruction
  • Modeling and Control of mobile robot
  • Multi-agent based architectures
  • Cooperative unmanned vehicles (not restricted to ground transportation)
  • Multi autonomous vehicles studies, models,techniques and simulations
  • International Program Committee

  • Alberto Broggi (VisLab, Parma University, Italy)
  • Stefano Cattani (VisLab, Parma University, Italy)
  • Javier Ibanez-Guzman (Renault, France)
  • Christian Laugier (Emotion, INRIA, France)
  • Philippe Martinet (Blaise Pascal University, France)
  • Urbano Nunes (Coimbra University, Portugal),
  • Cedric Pradalier, (ETH Zurich, Switzerland)
  • Anya Petrovskaya (Stanford University, USA)
  • Christoph Stiller (Karlruhe University, Germany)
  • Roland Siegwart, (ETH Zurich, Switzerland)
  • Luciano Oliveira, (UFBA, Brasil)
  • Preliminary program

    Introduction to the workshop 8:20

    Session I: Path Planning & Navigation systems 8:30
    Chairman: Urbano Nunes

    • Title: : Why can’t road positioning and integrity be friends? 8:30
      Keynote speaker: Rafael Toledo-Moreo (Technical University of Cartagena, Spain) 30min + 5min questions Related paper, Presentation

      Abstract: Today’s positioning systems work quite well in many situations. However, they lack some robustness, what hinders its exploitation for safety-critical and liability-critical applications. The warranty of the quality of the positioning service would open the way for further services, but for the community of the field it is unclear whether or not integrity is achievable for road positioning. There are several good reasons for that hesitation: The representation of the integrity of such a complex system by means of projecting all its possible errors onto an integrity parameter is difficult; the use of assumptions may cause that a non-modeled event disrupts the consistency of the error estimates; and also, the different ways in which the concept of integrity is approached in the literature lead to confusion and contradictions. This talk focuses on road positioning and its integrity, discussing aspects that play a role in this problem such as Global Navigation Satellite Systems, aiding sensors, the vehicle environment and its model, data fusion methods and map-matching algorithms.

    • Title: Autonomous Navigation in Crowded Campus Environments 9:05
      Authors: Z. J. Chong, B. Qin, T. Bandyopadhyay, T. Wongpiromsarn, E. S. Rankin, M. H. Ang Jr., E. Frazzoli, D. Rus, D. Hsu, K. H. Low 17min + 3min questions Paper, Presentation

      Abstract: This paper considers autonomous navigation in crowded city environments. An autonomous vehicle testbed is presented. We address two challenges of pedestrian detection and GPS-based localization in the presence of high-level build- ings. First, we augment the localization using local laser maps and show improved results. A pedestrian detection algorithm using a complementary vision and laser system is proposed. We implement this algorithm in our testbed and evaluate its performance using purely off the shelf components and open source software packages provided by ROS. We also show how utilizing existing infrastructural sensors can improve the performance of the system. Potential applications of this work include fully automated vehicle systems in urban environments typical in megacities in Asia.

    • Title: Integration of visual and depth information for vehicle detection 9:25
      Authors: A. Makris, M. Perrollaz, I. Paromtchik, C. Laugier 17min + 3min questions Paper, Presentation

      Abstract: In this work an object class recognition method is presented. The method uses local image features and follows the part based detection approach. It fuses intensity and depth information in a probabilistic framework. The depth of each local feature is used to weight the probability of finding the object at a given scale. To train the system for an object class only a database of annotated with bounding boxes images is required, thus automatizing the extension of the system to different object classes. We apply our method in the problem of detecting vehicles captured from a moving platform. The experiments in a data-set of stereo images captured in an urban environment show a significant improvement in performance when using both information modalities.

    Coffee Break 10:00

    Session II: Perception & Situation awareness 10:30
    Chairman: Rafael Toledo-Moreo
    • Title: Situation awareness & Risk based navigation in dynamic environments 10:30
      Keynote speaker: C. Laugier (Inria Grenoble, France) 30min + 5min questions Paper, Presentation
      Co-Authors: I. Paromtchik, M. Perrollaz, J.D. Yoder, C. Tay, K. Makhnacha, C. Fulgenzi, A. Spalanzani

      Abstract: This talk address the problem of safe navigation in dynamic environments, with a focus on intelligent vehicle application. After a global overview of the problem and of the state of the art, several key aspects of this problem will be addressed: Bayesian perception and sensor fusion, Motion prediction for sensed mobile obstacles (including maneuvers prediction at road intersections), Probabilistic collision risk assessment, and Risk based navigation. Results obtained with our equipped Lexus hybrid vehicle will be presented and discussed.

    • Title: From Structure to Actions: Semantic Navigation Planning in Office Environments 11:05
      Authors: K. Uhl, A. Roennau, R. Dillmann 17min + 3min questions Paper, Presentation

      Abstract: The use of meaning in mapping and navigation is inevitable if a robot has to interact with its environment in a goal-directed way. Moreover, a semantic environment model makes navigation planning more efficient and simplifies the review and communication of the robot’s knowledge. Existing work in this area decomposes the environment into places, which can be distinguished using the robot’s sensors. However, if important features of the environment cannot be detected by the robot’s sensors a different approach is needed. This paper introduces the Semantic Region Map, an envi- ronment model with complex metric, topological and semantic features. It shows how navigation points, so-called semantic positions, can be deduced from the map using a semantic description of the environment. Furthermore, the semantic positions are connected to a reachability graph, whose edges are labelled with robot actions, using a semantic description of the robot’s capabilities. An ontology consisting of a taxonomy and a set of rules are used to implement the semantic models. The concept of the Semantic Region Map is applied to a robot operating in an office environment.

    • Title: Situation Assessment and Trajectory Planning for AnnieWAY 11:25
      Authors: C. Stiller, J. Ziegler 17min + 3min questions Paper, Presentation, Video1

      Abstract: This contribution addresses machine perception of a priori unknown environment, situation recognition, and automated trajectory planning in urban traffic. We discuss how to represent and acquire metric, symbolic and conceptual knowledge from video and lidar data of a vehicle. A hardware and software architecture tailored to this knowledge structure for an autonomous vehicle is proposed. Emphasis is laid on methods for situation recognition employing geometrical and topological reasoning and Markov Logic Networks. Trajec- tory planning is conducted in spatiotemporal state lattices. The computational effort of the planning method is almost independent of the number of moving objects as these simply disable spatiotemporal nodes. The planning optimizes a quality measure that considers safety, efficiency, and comfort. Results are shown from the autonomous vehicle AnnieWAY that is able to autonomously travel in urban and platooning scenarios.

    Lunch break 12:00

    Session III: Interactive session 13:30
    Chairman: P. Martinet
    • Title: Proposition for propagated occupation grids for non-rigid moving objects tracking
      Authors: B. Lefaudeux, G. Gate, F. Nashashibi Paper

      Abstract: Autonomous navigation among humans is, however simple it might seems, a difficult subject which draws a lot a attention in our days of increasingly autonomous systems. From a typical scene from a human environment, diverse shapes, behaviours, speeds or colours can be gathered by a lot of sensors ; and a generic mean to perceive space and dynamics is all the more needed, if not easy. We propose an incremental evolution over the well-known occupancy grid paradigm, introducing grid cell propagation over time and a limited neighbourhood, handled by probabilistic calculus. Our algorithm runs in real-time from a GPU implementation, and considers completely generically space-cells propagation, without any a priori requirements. It produces a set of belief maps of our environment, handling occupancy, but also items dynamics, relative rigidity links, and an initial object classification. Observations from free-space sensors are thus turned into information needed for autonomous navigation.

    • Title: Probabilistic Road Geometry Estimation using a Millimetre-Wave Radar
      Authors: A. Hernandez-Gutierrez, J. I. Nieto, T. Bailey, E.M. Nebot Paper, Poster, Video1, Video2,

      Abstract: This paper presents a probabilistic framework for road geometry estimation using a millimetre wave radar. It aims at estimating the geometry of unpaved and unmarked roads, and also provides the vehicle location with respect to the edges of the road. This road tracking system employs a radar sensor due to its robustness to weather conditions such as fog, dust, rain and snow. The proposed approach is robust to noisy measurements because the radar target locations are modelled as Gaussian distributions. These observations are integrated into a Kalman Particle filter to estimate the posterior distribution of the parameters that best describe the geometry of the road. Experimental results using data acquired on a highway road are presented. The effectiveness of the proposed approach is demonstrated by a qualitative analysis of the results.

    • Title: Safety robotic lawnmower with precise and low-cost L1-only RTK-GPS positioning
      Authors: J.M. Codol, M. Poncelet, A. Monin, M. Devy Paper

      Abstract: In this paper, we will introduce an autonomous robotic lawnmower, equipped by a safety and low-cost RTK- DGPS centimetric positioning system available also in semi-urban environment. The GPS-RTK sensors are a pair of L1-only GPS receivers (L1-only GPS receivers are cheaper than dual-frequency ones because of the existence of patents on the usage of the second frequency). This work is an extension of a collaboration between NAV ON TIME and BELROBOTICS, consisting on evaluate GPS replacement for the current mower area limit (a buried wire). The objective of the latest work is to ensure the GPS mission realization, keeping the same safety as the buried wire one. In this context, this paper will present a complete statistical approach to L1-only RTK-positioning system in urban environment. The result of this approach have been embedded into the mower machine, by using a Linux operating system equipped with an ARM-9 processor running at 400MHz, and an UHF radio-communication to the reference station, this one having the role of realize path planning, geographical database managing, remote and IHM communication.

    • Title: Odometry from Planar landmarks
      Authors: K. Narayana, B. Steux Paper

      Abstract: This paper presents a new perception odometry approach using extracted stationary planar features to resolve 5 degrees of freedom of the robot motion. The approach exploits the geometrical properties of the extracted features to determine the transformation of the moving robot, which has perceived these landmarks. This way of localizing can help several applications in indoors and outdoors such as urban canyons, with plenty of planar features. The paper presents the concept and the algorithm, and validates them using a simulated scenario.

    • Title: Probabilistic autonomous navigation using Risk-RRT approach and models of human interaction
      Authors: J. Rios-Martinez, A. Spalanzani, C. Laugier Paper

      Abstract: Autonomous transportation in human environ- ments must follow social conventions. An autonomous wheelchair, for example, must respect proximity constraints but also respect people interacting, it should not break interaction between people talking, unless the user want to interact with them. In this case, the robot (i.e. the wheelchair) should find the best way to join the group. In this paper, we propose a risk-based navigation method which include risk of collision but also risk of disturbance. Results exhibit new emerging behavior showing how the robot takes into account social conventions in its navigation strategy.

    Session IV: 2D and 3D Mapping & Localization 14:30
    Chairman: C. Laugier
    • Title: 2D/3D mapping and localization 14:30
      Keynote speaker: C. Stiller (Karlsruhe Institute of Technology) 30min + 5min questions Presentation, Video1, Video2
      Co-Authors: A. Geiger, F. Moosmann

      Abstract: On-line environment modelling and mapping are of growing importance for autonomous robots and cognitive automobiles. Emerging from 2D or 2.5D flat world representations improved sensors and computing capabilities allow for 3D world representations and estimation of full 6 DOF robot motion. The choice of fast and expressive features enhances density while reducing noise. We present 3D dense maps acquired in real time from a lidar or stereo camera rig mounted on a traveling experimental vehicle without requiring any additional odometry or inertial sensors.

    • Title: A New Strategy for Feature Initialization in Visual SLAM 15:05
      Authors: G.Bresson, T. Feraud, R. Aufrere, P. Checchin and R. Chapuis 17min + 3min questions Paper, Presentation, Video1, Video2, Video3, Video4

      Abstract: This paper presents a Visual EKF-SLAM process using an original and very efficient strategy for initializing landmarks. Usually, with Cartesian coordinates, new points are created along the line-of-sight with a large variance. However, this type of initialization is subject to significant linearization issues making landmarks diverge from their real position. The immediate consequence is a failure of the Visual SLAM process. We propose here a new strategy that avoids or drastically limits the linearization errors. The first part of this strategy takes place during the tracking process where a coherent window is needed in order to successfully follow a point and make it converge. The second part concerns the update step. Due to linearization errors, a landmark in front of the observer can be updated behind it. We compute a corrective of the Kalman gain in order to preserve the integrity. We applied this strategy to real data illustrating its efficiency.

    • Title: Building Facade Detection, Segmentation, and Parameter Estimation for Mobile Robot Localization and Guidance 15:25
      Authors: J.A. Delmerico, P. David, J.J. Corso 17min + 3min questions Paper, Presentation

      Abstract: Building facade detection is an important problem in computer vision, with applications in mobile robotics and semantic scene understanding. In particular, mobile platform localization and guidance in urban environments can be enabled with an accurate segmentation of the various building facades in a scene. Toward that end, we present a system for segmenting and labeling an input image that for each pixel, seeks to answer the question “Is this pixel part of a building facade, and if so, which one?” The proposed method determines a set of candidate planes by sampling and clustering points from the image with RANSAC, using local normal estimates derived from PCA to inform the planar model. The corresponding disparity map and a discriminative classification provide prior information for a two-layer Markov Random Field model. This MRF problem is solved via Graph Cuts to obtain a labeling of building facade pixels at the mid-level, and a segmentation of those pixels into particular planes at the high-level. The results indicate a strong improvement in the accuracy of the binary building detection problem over the discriminative classifier alone, and the planar surface estimates provide a good approximation to the ground truth planes.

    Coffee break 16:00

    Session V: Mobile robot modeling and control 16:30
    Chairman: U. Nunes
    • Title: Generic algorithm for high accurate trajectory control in different conditions 16:30
      Keynote speaker: R. Lenain (Cemagref, France) 30min + 5min questions Presentation, Video1, Video2, Video3, Video4, Video5, Video6, Video7, Video8, Video9, Video10, Video11, Video12, Video13, Video14, Video15, Video16, Video17
      Co-Authors: B. Thuilot, C. Cariou, P. Martinet

      Abstract: From public transportation to agriculture, many fields of application may benefit from automation in the area of mobile robotics. As a result, research in that topic is subject of more and more investigation in order to propose new systems, from driver assistance (e.g automatic parking...) up to fully autonomous vehicles (such as autonomous robots acting in hazardous environment). In order to be fully effective, these innovations have to be accurate, safe, and able to act in various conditions. Many open problems then need to be addressed in order to propose such innovations in several part. If perception and navigation issues constitute important key points, the problem of motion control remains an important point since the control law to be embedded have to face a variability of conditions impacting directly their behaviour. These conditions rely on constant parameter, pending on the considered robot or vehicle (mechanical properties, actuators, specifications, …), but also depends on the variable interaction with the environment (grip conditions, terrain geometry, reachable velocity, ...). As a result in order to propose an efficient and accurate motion whatever the conditions variability, control laws have to account of the different dynamics encountered. This talk investigates the motion control of mobile robot in different conditions through the example of path tracking. It proposes several strategies to preserve the motion accuracy and safety whatever the encountered conditions. A correlation between the reachable velocity and the terrain complexity is proposed to extract the different effects which have to be accounted and related control objective. Based on this classification several modelling and control strategies are illustrated to face the considered phenomena. Starting from classical kinematic controller for simple path tracking task at low speed on flat terrain with good grip conditions, the talk investigates a rising complexity of situation. Adaptive control based on advanced kinematic model is proposed to face low grip conditions. This adaptive control is then associated with predictive control in order to preserve accuracy when increasing the velocity. Limitation of this controller with respect to the increasing speed and safety is pointed out and a new observer mixing kinematic and dynamic representation is proposed. This model permits also to account for 3D motion and permit to investigate the risk of instability rising at high speed. A control law acting on velocity in order to limit the rollover risk is then derived. This notion is then extended in a predictive way to adress the topic of obstacle avoidance and traversability for mobile robots. Finally, the notion of predictive control on velocity is extended to preserve the integrity of mobile robot, i.e, to preserve the stability, the controlability, and the accuracy of motion control. The capabilities of the different algorithm are investigated on actual experiments, using different kind of robots and vehicle, moving on different kind of ground.

    • Title: A control strategy taking advantage of inter-vehicle communication for platooning navigation in urban environment 17:05
      Authors: P. Avanzini, B. Thuilot, P. Martinet 17min + 3min questions Paper, Presentation

      Abstract: This paper deals with platooning navigation in the context of innovative solutions for urban transportation systems. More precisely, a sustainable approach centered on au- tomated electric vehicles in free-access is considered. To tackle the major problem of congestions in dense areas, cooperative navigation according to a platoon formation is investigated. With the aim to ensure the formation stability, i.e. longitudinal disturbances within the platoon do not grow when progressing down the chain, a global decentralized platoon control strategy is here proposed. It is supported by inter-vehicle communica- tions and relies on nonlinear control techniques. A wide range of experiments, carried out with up to four urban vehicles, demonstrates the capabilities of the proposed approach: two localization devices have been tested (RTK-GPS and monocular vision) along with two guidance modes (the path to be followed is either predefined or inferred on-line from the motion of the manually driven first vehicle).

    • Title: Semiautonomous Longitudinal Collision Avoidance Using a Probabilistic Decision Threshold 17:25
      Authors: J. Johnson, Y. Zhang, K. Hauser 17min + 3min questions Paper, Presentation

      Abstract: Automated emergency maneuvering systems can avoid or reduce the severity of collisions by taking control of a vehicle away from the driver during high-risk situations. The choice of when to switch to emergency control is challenging in the presence of uncertain information (imperfect sensors, road conditions, uncertain object behavior, etc.) and many dynamic obstacles. This paper considers longitudinal collision avoidance problems for a vehicle traveling along a known path. A probabilistic decision threshold framework is presented in which the user’s control is overridden if the probability that it would lead the system into an unsafe state exceeds some threshold.We apply the technique to collision imminent braking for obstacles traveling along the vehicle’s path, and present preliminary results extending the technique to the scenario of obstacles crossing the vehicle’s path.

    Clothing 17:50
    Author Information

      Format of the paper: Papers should be prepared according to the IROS11 final camera ready format and should be 4 to 6 pages long. The detailed information on the paper format is available from the IROS11 page. http://www.iros2011.org/author-instructions. Papers must be sent to the organizers by email.

      Important dates (preliminary)

      • Deadline for Paper submission: June 1st, 2011
      • Acceptance with review comments: June 15th, 2011
      • Deadline for final paper submission: June 30th, 12am at last, 2011

      Talk information

      • Invited talk: 35 min (30 min talk, 5 min question)
      • Other talk: 20 min (17 min talk, 3 min question)

      Poster session

      • Interactive and open session: 1h00

    Previous workshops

      Previously, six workshops were organized in the near same field. The 1st edition PPNIV'07 of this workshop was held in Roma during ICRA'07 (around 60 attendees), the second SNODE'07 in San Diego during IROS'07 (around 80 attendees), the third PPNIV'08 in Nice during IROS'08 (more than 90 registered people), the fourth edition SNODE'09 in Kobe during ICRA'09 (around 70 attendees), the fifth edition PPNIV'09 during IROS'09 in Saint-Louis (around 70 attendees) , and the last one RITS'10 was organized in the last ICRA'10 in Anchrorage (around 35 attendees).
      A special issue in IEEE Transaction on ITS, mainly focused on Car and ITS applications, has been published in September 2009. We are preparing proposal for special issue in IEEE RAS magazine and International Journal of Robotic Research Research. Best papers will be pushed to be extended and submitted to these special issues.

    Keynotes

      Proceedings: The workshop proceedings will be published within the IROS Workshop/Tutorial CDROM and electronically as a pdf file.

      Special issue: Selected papers will be considered for a special issue in an International Journal in connection with this workshop. We will issue an open call after the workshop, submissions will go through a separate peer review process.