This article is protected by German copyright law. You may copy and distribute this article for your personal use only. Other use is only allowed with written permission by the copyright holder. it 5/2005 Schwerpunktthema Simultanous Mapping and Localization of Rescue Environments Gleichzeitiges Kartieren und Lokalisieren von Rescue-Umgebungen Hartmut Surmann, Kai Pervölz, Fraunhofer AIS, Sankt Augustin, Andreas Nüchter, Kai Lingemann, Joachim Hertzberg, University of Osnabrück, Matthias Hennig, Dresden University of Technology Summary Deploying rescue workers in an urban setting is often a perilous, time-, power-, and force-consuming job, and robot systems to assist in this effort are needed. A fundamen- tal task for rescuer is to localize and salvage injured persons. To this end, robotic systems are used for mapping a site and for re- mote inspection of suspicious objects. The mobile robot Kurt3D is the first rescue robot that is capable of mapping its envi- ronment in 3D and self localize in all six degrees of freedom, i. e., considering its x, y and z positions and the roll, yaw and pitch angles. This capability allowed the small robot to win the Vice World Championship at the RoboCup Rescue competition 2004 at Lisbon. ◮◮◮ Zusammenfassung Der Einsatz von Rettungskräften in Ballungsgebieten ist eine gefährliche, Zeit-, Energie- und kraftaufwändige Arbeit, bei der Robotersys- teme zur Unterstützung benötigt werden. Eine fundamentale Aufgabe der Retter ist die Lokalisierung und Bergung verletzter Personen. Robotersystem werden für die Kartierung der Umge- bung und der Inspektion von verdächtigen Objekten verwendet. Der mobile Roboter Kurt3D ist der erste Rescue-Roboter, der in der Lage ist, seine Umgebung dreidimensional zu kartie- ren und sich in der Karte selbst zu lokalisieren, und zwar in allen 6 Freiheitsgeraden, d. h. unter Berücksichtigung der drei Raumpositionen x, y und z und den Dreh-, Schwenk- und Nei- gewinkeln. Diese Fähigkeit ermöglichte es dem kleinen Roboter, die Vize-Weltmeisterschaft auf dem RoboCup Rescue Wettbe- werben 2004 in Lissabon zu gewinnen. KEYWORDS I.2.9 [Robotics] Mobile Robots, 6D-SLAM, RoboCup Rescue, Mapping, Localization 1 Introduction For protecting humans, it is nowa- days important to build robots that are able to operate in earthquake, fire, explosive and chemical disas- ter areas. The community of Urban Search and Rescue Robotics (USAR) grows very fast. Many robots are manufactured, both from research institutes and from industry. How- ever, until now, there have been no systems that can reliably map their environment in 3D under real- time conditions. The mobile robot Kurt3D was presented at RoboCup The paper is an extended version of the SSRR 2005 Best paper awarded pa- per “Mapping of Rescue Environments with Kurt3D” [18] Rescue 2004 in Lisbon (Fig. 1). The robot is capable of mapping its en- vironment in 3D and self localize in six degrees of freedom, i. e., consid- ering its x, y and z positions and the roll, yaw and pitch angles (6D SLAM) in real-time. Kurt3D’s driving and mapping system consists of four major parts. First is a non linear feed forward PI controller. Second is the precise pla- nar pose tracking algorithm HAYAI. Using 2D laser scans, features that correspond to natural landmarks are extracted and paired with features of previous scans. Third, a fast and reliable 3D scan matching proced- ure, employing a point reduction and approximate nearest neighbor search, is used to generate 3D maps and relocalize the robot. The ba- sis of the 3D scan matching is the well know iterative closest points (ICP) algorithm. The fourth mod- ule is an interactive, semi-automatic control program that enables the operator to interfere with the map- ping process for corrections. This paper focuses on Kurt3D’s driving, mapping and localization modules and their interaction in the rescue context, devising the similarities be- tween HAYAI and 6D SLAM. 1.1 Rescue Robotics Systems Current rescue robots are mainly designed for searching for vic- tims and paths through rubble 282 it – Information Technology 47 (2005) 5  Oldenbourg Verlag