In this paper, we present a stereovision algorithm for real-time 6DoF ego-motion estimation, which integrates image intensity information and 3D stereo data in the well-known Iterative Closest Point (ICP) scheme. The proposed method addresses a basic problem of standard ICP, i.e. its inability to perform the segmentation of data points and to deal with large displacements. Neither a-priori knowledge of the motion nor inputs from other sensors are required, while the only assumption is that the scene always contains visually distinctive features which can be tracked over subsequent stereo pairs. This generates what is usually called Visual Odometry. The paper details the various steps of the algorithm and presents the results of experimental tests performed with an allterrain mobile robot, proving the method to be as accurate as effective for autonomous navigation purposes. 1.

Real-time 3D localization and mapping is eventually needed in many service robotic applications. Toward a light and practical SLAM algorithm, we focus on feature extraction via segmentation of range images. Using horizontal and vertical traces of the range matrix, 2D observed polygons are considered for calculation of a one-dimensional measure of direction, called Bearing Angle (BA). BA is the incident angle between the laser beam and edges of the observed polygon by the scanner in the selected direction. Based on this measure, two different approaches to range image segmentation, region- and edge-based, are proposed and evaluated through a set of standard analysis. It is experimentally shown that for navigation applications, edge based approaches are more efficient. Extensive tests on real robots prove BA-based segmentation is successful for SLAM. 1.

Abstract—For mobile robots to be of value in practical situations a 3D perception and mapping capability will almost always prove essential. In this paper a 2D laser scanner is modified to produce 3D scans with a resolution of one degree updated every 3 seconds. This result is achieved by adding a rotating mirror to the original scanner in an inexpensive and relatively simple modification that is easily made to a standard mobile robot. The modified robot is then able to produces 3D scans at a various frequencies up to 1Hz and accurate to 0.02m over an 8m range. Index Terms—3D laser scanner, mobile robots, range sensors I.

A basic issue of mobile robotics is generating environment maps automatically. Outdoor terrain is challenging since the ground is uneven and the surrounding is structured irregularly. In earlier work, we have introduced 6D SLAM (Simultaneous Localization and Mapping) as a method to taking all six DOF of robot poses (x, y, z translation; roll, pitch, yaw angles) into account. This paper adds to 6D SLAM a method for extracting drivable surfaces in the 3D maps while they are being generated. Experiments have

Soon, in many service robotic applications, a realtime localization and 3D-mapping capability will be necessary for autonomous navigation. Toward a light and practical SLAM algorithm for indoor scenarios, we propose a fast SLAM algorithm which benefits from sensor geometry for feature extraction and enhance the mapping process using dominant orthogonality in the engineered structures of man-made environments. Range images obtained using a nodding SICK are segmented into planar patches with polygonal boundaries in linear time. Right corner features are constructed based on the recognized orthogonal planes and used for robot localization. In addition to these corners, the map also contains planar patches with inner and outer boundaries for 3D modeling and recognition of the major building structures. Experiments using a mobile robot in our laboratory hallway prove the effectiveness of our approach. Results of the algorithm are compared with hand-measured ground truth.

Abstract — A 3D laser scanner is built by adding a rotating mirror to a conventional 2D scanner. The scanners are deployed on four robots to build full 3D representations of an indoor environment. An original representation mechanism referred to as occupancy lists, rather than standard 2D free space grids, is used to maintain the 3D map. Localization is done by extracting horizontal sub-ceiling cross-sections. Taking cross-sections from near the ceiling in this way results in more reliable and time invariant maps. Experimental results show inter-robot sightings and the sharing of map data aid mapping by improving the reliability of localization. Mapping with four robots reduced the average position error from 0.35m for single robot operation to 0.1m when cooperating.

This thesis is concerned with three-dimensional scan registration, in particular of underground mine tunnels. Registration of partial range scans is an essential part of building 3D maps, and autonomous creation of reliable maps is a first step towards autonomous mining vehicles. The thesis presents a survey of relevant sensor technology and discusses the advantages and disadvantages of different sensors for use in mine environments. A survey of the state of the art in scan registration algorithms is also presented, as well as a number of relevant applications. A new algorithm for registration of 3D data is presented, which is a generalisation of the normal distributions transform (NDT) for 2D data developed by Biber and Straßer. A detailed quantitative and qualitative comparison of the new algorithm with existing registration algorithms is shown. Results with actual mine data, some of which were collected with a new prototype 3D laser scanner, show that the presented algorithm is faster and in many cases more

This paper describes a novel registration approach that is based on a combination of visual and 3D range information. To identify correspondences, local visual features obtained from images of a standard color camera are compared and the depth of matching features (and their position covariance) is determined from the range measurements of a 3D laser scanner. The matched depth-interpolated image features allows to apply registration with known correspondences. We compare several ICP variants in this paper and suggest an extension that considers the spatial distance between matching features to eliminate false correspondences. Experimental results are presented in both outdoor and indoor environments. In addition to pair-wise registration, we also propose a global registration method that registers all scan poses simultaneously.

This paper describes the development of a 3D laser scanner and an approach to 3D mapping and localization. The 3D scanner consists of a standard 2D laser scanner and a rotating mirror assembly. Employing multiple robots and mutual localization local 3D maps are built. Global localization within the maps is performed by extracting a cross-section of the map just below the ceiling and then using an exhaustive search algorithm to enable the merger of multiple local 3D maps. The quality of these maps is such that the poses estimated by this method are accurate to within 0.1m and 1 degree.

Abstract — Simple, fast and lightweight SLAM algorithms are necessary in many embedded robotic systems which soon will be used in houses and offices in order to do various service tasks. In this paper the Orthogonal SLAM algorithm is presented as an answer to this need. In continuation of our previous work, the algorithm is extended to generate 3D maps and empirically validated by mapping the long corridor of our lab with the accuracy comparable with hand measured ground truth. The main contribution resides in the idea of reducing the complexity by using orthogonality constraint in indoor environments. This is done by mapping only planes that are parallel or perpendicular to each other which represent the main structure of most indoor environments. Having this assumption, we use an inclined sensor setup (fixed 2D SICK laser range finders) to generate 3D orthogonal maps. The algorithm is extremely fast since in each step it just processes one line of laser measurements. I.