In this paper, we address the problem of robust data association for simultaneous vehicle localization and map building. We show that the classical gated nearest neighbor approach, which considers each matching between sensor observations and features independently, ignores the fact that measurement prediction errors are correlated. This leads to easily accepting incorrect matchings when clutter or vehicle errors increase. We propose a new measurement of the joint compatibility of a set of pairings that successfully rejects spurious matchings. We show experimentally that this restrictive criterion can be used to efficiently search for the best solution to data association. Unlike the nearest neighbor, this method provides a robust solution in complex situations, such as cluttered environments or when revisiting previously mapped regions.

This paper presents the design of a high accuracy outdoor navigation system based on standard dead reckoning sensors and laser range and bearing information. The data validation problem is addressed using laser intensity information. The beacon design aspect and location of landmarks are also discussed in relation to desired accuracy and required area of operation. The results are important for Simultaneous Localization and Map building applications, (SLAM), since the feature extraction and validation are resolved at the sensor level using laser intensity. This facilitates the use of additional natural landmarks to improve the accuracy of the localization algorithm. The modelling aspects to implement SLAM with beacons and natural features are also presented. These results are of fundamental importance because the implementation of the algorithm does not require the surveying of beacons. Furthermore we demonstrate that by using natural landmarks high accurate localization can be achieved by only requiring the initial estimate of the position of the vehicle. The algorithms are validated in outdoor environments using a standard utility car retrofitted with the navigation sensors and a 1 cm precision Kinematic GPS used as ground truth. 1

In this paper a new localization approach combining the metric and topological paradigm is presented. The main idea is to connect local metric maps by means of a global topological map. This allows a compact environment model which does not require global metric consistency and permits both precision and robustness. The method uses a 360 degree laser scanner in order to extract lines for the metric localization and doors, discontinuities and hallways for the topological approach. The approach has been widely tested in a 50 x 25 m portion of the institute building with the new fully autonomous robot Donald Duck. 25 randomly generated test missions have been performed with a success ratio of 96 % and a mean error at the goal point of 9 mm for an overall trajectory length of 1.15 km. Future work will focus on a similar hybrid approach for simultaneous localization and automatic mapping. 1.

A layered algorithm for extracting features in an indoor environment from planar range data is presented. At the lower (signal processing) level, the SEGMENTS algorithm exploits the speed and accuracy of two Extended Kalman filters working in tandem and processes sequentially the distance measurements within each scan. The first Kalman filter is responsible for initiating straight line segments and detecting clutter while the second estimates the parameters of each line segment (such as distance and orientation) and determines when this is interrupted. The dual filter combination is capable of detecting edges and straight line segments within the scene infront of the robot. At the higher (post-processing) level, the identified segments are combined to form more complex features such as extended walls, corners (concave and convex), doors and corridors. During the composition cycle, SEGMENTS makes use of (i) the parametric representation of the straight line segments and (ii) the prespecified topological models of the features this algorithm seeks. A list of the identified features along with their location with respect to the laser sensor is finally available to the user. The cluttered regions in the scene are also marked on a polar representation of the environment. The presented algorithm has been tested on a Pioneer2DX mobile robot equipped with a SICK, LMS 200 proximity laser scanner performing map-based localization. Low computational requirements, accuracy and robustness to uncertainty and noise characterize the performance ofthisnewmethod for feature extraction.

A layered algorithm for extracting features in an indoor environment from planar range data is presented. At the lower (signal processing) level, the SEGMENTS algorithm exploits the speed and accuracy of two Extended Kalman lters working in tandem and processes sequentially the distance measurements within each scan. The rst Kalman lter is responsible for initiating straight line segments and detecting clutter while the second estimates the parameters of each line segment (such as distance and orientation) and determines when this is interrupted. The dual lter combination is capable of detecting edges and straight line segments within the scene infront of the robot. At the higher (post-processing) level, the identi ed segments are combined to form more complex features such as extended walls, corners (concave and convex), doors and corridors. During the composition cycle, SEGMENTS makes use of (i) the parametric representation of the straight line segments and (ii) the prespeci ed topological models of the features this algorithm seeks. A list of the identi ed features along with their location with respect to the laser sensor is nally available to the user. The cluttered regions in the scene are also marked on a polar representation of the environment. The presented algorithm has been tested on a Pioneer2DX mobile robot equipped with a SICK, LMS 200 proximity laser scanner performing map-based localization. Low computational requirements, accuracy and robustness to uncertainty and noise characterize the performance ofthisnewmethod for feature extraction. 1

This paper presents a new appearance-based place recognition system for topological localization. The method uses a panoramic vision system to sense the environment. Color images are classified in real-time based on nearest-neighbor learning, image histogram matching, and a simple voting scheme. The system has been evaluated with eight cross-sequence tests in four unmodified environments, three indoors and one outdoors. In all eight cases, the system successfully tracked the mobile robot's position. The system correctly classified between 87% and 98% of the input color images. For the remaining images, the system was either momentarily confused or uncertain, but never classified an image incorrectly.