We describe a LIDAR-based autonomous navigation approach applicable to both urban and non-urban e... more We describe a LIDAR-based autonomous navigation approach applicable to both urban and non-urban environments. At the core of the method is a set of “tentacles” that represent precalculated trajectories defined in the ego-centered coordinate system of the vehicle. Similar to an insect’s antennae, they fan out with different curvatures discretizing the basic driving options of the vehicle. We detail how the approach can be used for exploration of unknown environments and how it can be extended to combined GPS path following and obstacle avoidance allowing save road following in case of GPS offsets.
This paper describes an approach for autonomous offroad navigation for the Civilian European Land... more This paper describes an approach for autonomous offroad navigation for the Civilian European Landrobot Trial 2009 (C-ELROB). The main focus in this paper is how to cope with poor GPS conditions such as occurring in forest environments. Therefore the expected and detected problems are stated, and methods how to achieve good autonomous driving performance due to less weigthing of inaccurate GPS information are presented. But first, an introduction how to combine obstacle-avoidance and path following behavior within the reactive so-called tentacles approach is given. Extensions for utilizing visual information for tentacle evaluation as well as the integration of geodetic information to gather information on the road network are presented afterwards. Finally the results gathered from the successful C-ELROB trials are analyzed.
Results for the application of an adaptive background model to the problem of detecting changes o... more Results for the application of an adaptive background model to the problem of detecting changes on a combs surface are reported. It is demonstrated that the combined search for uncapped brood cells in the current image and the background image increases the overall detection rate of the system.
This paper describes a LIDAR-based perception system for ground robot mobility, consisting of 3D ... more This paper describes a LIDAR-based perception system for ground robot mobility, consisting of 3D object detection, classification and tracking. The presented system was demonstrated on-board our autonomous ground vehicle MuCAR-3, enabling it to safely navigate in urban traffic-like scenarios as well as in off-road convoy scenarios. The efficiency of our approach stems from the unique combination of 2D and 3D data processing techniques. Whereas fast segmentation of point clouds into objects is done in a 2 1 2 D occupancy grid, classifying the objects is done on raw 3D point clouds. For fast switching of domains, the occupancy grid is enhanced to act like a hash table for retrieval of 3D points. In contrast to most existing work on 3D point cloud classification, where realtime operation is often impossible, this combination allows our system to perform in real-time at 0.1s frame-rate.
This paper briefly describes the hardand software system of TAS/UniBw’s entry for the 2009 C-ELRO... more This paper briefly describes the hardand software system of TAS/UniBw’s entry for the 2009 C-ELROB robotic trials, participating at the autonomous navigation scenario. At the core of the system is the so-called tentacles approach to reactive robot navigation, that we already demonstrated successfully at the 2007 C-ELROB trials. Since then, the basic tentacles approach has been extended in many ways. We outline an efficient method for accumulating 3D LIDAR data into multi-layered occupancy grids to be used for tentacle evaluation. We then show how to combine obstacleavoidance and path following behavior within the reactive tentacles approach. Finally, we show how visual information can be used for tentacle evaluation and ERI-Card detection.
Zusammenfassung: Die Wahrnehmung des Fahrzeugumfelds ist Grundvoraussetzung für autonomes Fahren.... more Zusammenfassung: Die Wahrnehmung des Fahrzeugumfelds ist Grundvoraussetzung für autonomes Fahren. Hierfür wird sowohl eine hohe Sensorauflösung als auch ein großes Gesichtsfeld benötigt. Aktive Wahrnehmung, d.h. die aktive Änderung der Sensororientierung, ist eine Möglichkeit beides zu erreichen. Allerdings erfordert aktive Wahrnehmung die Auswahl einer geeigneten Sensororientierung. Dieser Beitrag präsentiert eine neue Methode für die Auswahl der Sensororientierung in städtischen Verkehrsszenarien. Die Auswahl der Sensororientierung basiert auf drei Kriterien: Der Bedeutung von Verkehrsteilnehmern für die aktuelle Situation, dem verfügbaren Wissen über einen Verkehrsteilnehmer unter Berücksichtigung alternativer Sensororientierungen sowie der räumlichen Erfassung des relevanten Fahrzeugumfelds durch den Sensor.
In this paper, we describe the hardware and software components of a fully autonomous prototype d... more In this paper, we describe the hardware and software components of a fully autonomous prototype delivery vehicle. Equipped with a robotic arm, the demonstrator is capable of delivering packages and picking up new ones by interacting with custom-made delivery boxes. As highly accurate positioning w. r. t. a box is required for successful handover of packages, we track the pose (position and orientation) of the box using a high-resolution on-board camera. The resulting estimate is relayed to our planning and control modules, which ensure that the vehicle reaches its required pose with centimeter-level accuracy.In order to deliver packages, the car needs to autonomously navigate our test facility, avoiding static and dynamic obstacles while obeying simple traffic rules. As one focus is on the practical challenges encountered when building a prototype, we cover issues ranging from sensor calibration and system identification to perception, planning, control, and the implementation of hi...
This paper describes a fast method for segmentation of large-size long-range 3D point clouds that... more This paper describes a fast method for segmentation of large-size long-range 3D point clouds that especially lends itself for later classification of objects. Our approach is targeted at high-speed autonomous ground robot mobility, so real-time performance of the segmentation method plays a critical role. This is especially true as segmentation is considered only a necessary preliminary for the more important
2010 Ieee Intelligent Vehicles Symposium, Jun 21, 2010
This paper describes a fast method for segmentation of large-size long-range 3D point clouds that... more This paper describes a fast method for segmentation of large-size long-range 3D point clouds that especially lends itself for later classification of objects. Our approach is targeted at high-speed autonomous ground robot mobility, so real-time performance of the segmentation method plays a critical role. This is especially true as segmentation is considered only a necessary preliminary for the more important
In this paper we describe how visual features can be incor- porated into the well known tentacles... more In this paper we describe how visual features can be incor- porated into the well known tentacles approach (1) which up to now has only used LIDAR and GPS data and was therefore limited to scenarios with signicant obstacles or non-at surfaces along roads. In addition we present a visual feature considering only color intensity which can be used to
This paper describes an approach for autonomous offroad navigation for the Civilian European Land... more This paper describes an approach for autonomous offroad navigation for the Civilian European Landrobot Trial 2009 (C-ELROB). The main focus in this paper is how to cope with poor GPS conditions such as occurring in forest environments. Therefore the expected and detected problems are stated, and methods how to achieve good autonomous driving performance due to less weigthing of inaccurate
2009 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2009
This paper describes a LIDAR-based perception system for ground robot mobility, consisting of 3D ... more This paper describes a LIDAR-based perception system for ground robot mobility, consisting of 3D object detection, classification and tracking. The presented system was demonstrated on-board our autonomous ground vehicle MuCAR-3, enabling it to safely navigate in urban traffic-like scenarios as well as in off-road convoy scenarios. The efficiency of our approach stems from the unique combination of 2D and 3D
2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2011
... Overview In this section, we present experimental results of road and intersection tracking o... more ... Overview In this section, we present experimental results of road and intersection tracking obtained with our ... Our first main sensor is the Velodyne LIDAR HDL-64ES2, a high-resolution laser ... Forvisual perception, we use our camera platform MarVEye 8, which is equipped with ...
Results for the application of an adaptive background model to the problem of detecting changes o... more Results for the application of an adaptive background model to the problem of detecting changes on a combs surface are reported. It is demonstrated that the combined search for uncapped brood cells in the current image and the background image increases the overall detection rate of the system.
This paper describes a fast method for segmentation of large-size long-range 3D point clouds that... more This paper describes a fast method for segmentation of large-size long-range 3D point clouds that especially lends itself for later classification of objects. Our approach is targeted at high-speed autonomous ground robot mobility, so real-time performance of the segmentation method plays a critical role. This is especially true as segmentation is considered only a necessary preliminary for the more important
We describe a LIDAR-based autonomous navigation approach applicable to both urban and non-urban e... more We describe a LIDAR-based autonomous navigation approach applicable to both urban and non-urban environments. At the core of the method is a set of “tentacles” that represent precalculated trajectories defined in the ego-centered coordinate system of the vehicle. Similar to an insect’s antennae, they fan out with different curvatures discretizing the basic driving options of the vehicle. We detail how the approach can be used for exploration of unknown environments and how it can be extended to combined GPS path following and obstacle avoidance allowing save road following in case of GPS offsets.
This paper describes an approach for autonomous offroad navigation for the Civilian European Land... more This paper describes an approach for autonomous offroad navigation for the Civilian European Landrobot Trial 2009 (C-ELROB). The main focus in this paper is how to cope with poor GPS conditions such as occurring in forest environments. Therefore the expected and detected problems are stated, and methods how to achieve good autonomous driving performance due to less weigthing of inaccurate GPS information are presented. But first, an introduction how to combine obstacle-avoidance and path following behavior within the reactive so-called tentacles approach is given. Extensions for utilizing visual information for tentacle evaluation as well as the integration of geodetic information to gather information on the road network are presented afterwards. Finally the results gathered from the successful C-ELROB trials are analyzed.
Results for the application of an adaptive background model to the problem of detecting changes o... more Results for the application of an adaptive background model to the problem of detecting changes on a combs surface are reported. It is demonstrated that the combined search for uncapped brood cells in the current image and the background image increases the overall detection rate of the system.
This paper describes a LIDAR-based perception system for ground robot mobility, consisting of 3D ... more This paper describes a LIDAR-based perception system for ground robot mobility, consisting of 3D object detection, classification and tracking. The presented system was demonstrated on-board our autonomous ground vehicle MuCAR-3, enabling it to safely navigate in urban traffic-like scenarios as well as in off-road convoy scenarios. The efficiency of our approach stems from the unique combination of 2D and 3D data processing techniques. Whereas fast segmentation of point clouds into objects is done in a 2 1 2 D occupancy grid, classifying the objects is done on raw 3D point clouds. For fast switching of domains, the occupancy grid is enhanced to act like a hash table for retrieval of 3D points. In contrast to most existing work on 3D point cloud classification, where realtime operation is often impossible, this combination allows our system to perform in real-time at 0.1s frame-rate.
This paper briefly describes the hardand software system of TAS/UniBw’s entry for the 2009 C-ELRO... more This paper briefly describes the hardand software system of TAS/UniBw’s entry for the 2009 C-ELROB robotic trials, participating at the autonomous navigation scenario. At the core of the system is the so-called tentacles approach to reactive robot navigation, that we already demonstrated successfully at the 2007 C-ELROB trials. Since then, the basic tentacles approach has been extended in many ways. We outline an efficient method for accumulating 3D LIDAR data into multi-layered occupancy grids to be used for tentacle evaluation. We then show how to combine obstacleavoidance and path following behavior within the reactive tentacles approach. Finally, we show how visual information can be used for tentacle evaluation and ERI-Card detection.
Zusammenfassung: Die Wahrnehmung des Fahrzeugumfelds ist Grundvoraussetzung für autonomes Fahren.... more Zusammenfassung: Die Wahrnehmung des Fahrzeugumfelds ist Grundvoraussetzung für autonomes Fahren. Hierfür wird sowohl eine hohe Sensorauflösung als auch ein großes Gesichtsfeld benötigt. Aktive Wahrnehmung, d.h. die aktive Änderung der Sensororientierung, ist eine Möglichkeit beides zu erreichen. Allerdings erfordert aktive Wahrnehmung die Auswahl einer geeigneten Sensororientierung. Dieser Beitrag präsentiert eine neue Methode für die Auswahl der Sensororientierung in städtischen Verkehrsszenarien. Die Auswahl der Sensororientierung basiert auf drei Kriterien: Der Bedeutung von Verkehrsteilnehmern für die aktuelle Situation, dem verfügbaren Wissen über einen Verkehrsteilnehmer unter Berücksichtigung alternativer Sensororientierungen sowie der räumlichen Erfassung des relevanten Fahrzeugumfelds durch den Sensor.
In this paper, we describe the hardware and software components of a fully autonomous prototype d... more In this paper, we describe the hardware and software components of a fully autonomous prototype delivery vehicle. Equipped with a robotic arm, the demonstrator is capable of delivering packages and picking up new ones by interacting with custom-made delivery boxes. As highly accurate positioning w. r. t. a box is required for successful handover of packages, we track the pose (position and orientation) of the box using a high-resolution on-board camera. The resulting estimate is relayed to our planning and control modules, which ensure that the vehicle reaches its required pose with centimeter-level accuracy.In order to deliver packages, the car needs to autonomously navigate our test facility, avoiding static and dynamic obstacles while obeying simple traffic rules. As one focus is on the practical challenges encountered when building a prototype, we cover issues ranging from sensor calibration and system identification to perception, planning, control, and the implementation of hi...
This paper describes a fast method for segmentation of large-size long-range 3D point clouds that... more This paper describes a fast method for segmentation of large-size long-range 3D point clouds that especially lends itself for later classification of objects. Our approach is targeted at high-speed autonomous ground robot mobility, so real-time performance of the segmentation method plays a critical role. This is especially true as segmentation is considered only a necessary preliminary for the more important
2010 Ieee Intelligent Vehicles Symposium, Jun 21, 2010
This paper describes a fast method for segmentation of large-size long-range 3D point clouds that... more This paper describes a fast method for segmentation of large-size long-range 3D point clouds that especially lends itself for later classification of objects. Our approach is targeted at high-speed autonomous ground robot mobility, so real-time performance of the segmentation method plays a critical role. This is especially true as segmentation is considered only a necessary preliminary for the more important
In this paper we describe how visual features can be incor- porated into the well known tentacles... more In this paper we describe how visual features can be incor- porated into the well known tentacles approach (1) which up to now has only used LIDAR and GPS data and was therefore limited to scenarios with signicant obstacles or non-at surfaces along roads. In addition we present a visual feature considering only color intensity which can be used to
This paper describes an approach for autonomous offroad navigation for the Civilian European Land... more This paper describes an approach for autonomous offroad navigation for the Civilian European Landrobot Trial 2009 (C-ELROB). The main focus in this paper is how to cope with poor GPS conditions such as occurring in forest environments. Therefore the expected and detected problems are stated, and methods how to achieve good autonomous driving performance due to less weigthing of inaccurate
2009 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2009
This paper describes a LIDAR-based perception system for ground robot mobility, consisting of 3D ... more This paper describes a LIDAR-based perception system for ground robot mobility, consisting of 3D object detection, classification and tracking. The presented system was demonstrated on-board our autonomous ground vehicle MuCAR-3, enabling it to safely navigate in urban traffic-like scenarios as well as in off-road convoy scenarios. The efficiency of our approach stems from the unique combination of 2D and 3D
2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2011
... Overview In this section, we present experimental results of road and intersection tracking o... more ... Overview In this section, we present experimental results of road and intersection tracking obtained with our ... Our first main sensor is the Velodyne LIDAR HDL-64ES2, a high-resolution laser ... Forvisual perception, we use our camera platform MarVEye 8, which is equipped with ...
Results for the application of an adaptive background model to the problem of detecting changes o... more Results for the application of an adaptive background model to the problem of detecting changes on a combs surface are reported. It is demonstrated that the combined search for uncapped brood cells in the current image and the background image increases the overall detection rate of the system.
This paper describes a fast method for segmentation of large-size long-range 3D point clouds that... more This paper describes a fast method for segmentation of large-size long-range 3D point clouds that especially lends itself for later classification of objects. Our approach is targeted at high-speed autonomous ground robot mobility, so real-time performance of the segmentation method plays a critical role. This is especially true as segmentation is considered only a necessary preliminary for the more important
Uploads
Papers by Michael Himmelsbach