ISBN 9783843903813

978-3-8439-0381-3, Reihe Robotik und Automation

Jürgen Gräfenstein
Simultaneous Localization and Mapping in Wireless Networks Localization in Ubiquitous Infrastructures

143 Seiten, Dissertation Eberhard-Karls-Universität Tübingen (2011), Hardcover, A5

Zusammenfassung / Abstract

The spanning topic of this work is the utilization of existing infrastructures for cost efficient and accurate localization of mobile robots. Introductory single aspects of this topic are treated: a generic architecture for the integration of mobile robots into intelligent infrastructures; requirements on the architecture for localization and their application on a selection of possible infrastructures. This shows the good suitability of wireless networks inter alia because of their pervasiveness, standardized interfaces and low cost. A thorough evaluation of the state of the art research shows that no existing method can achieve low cost and accurate localization at the same time.

Therefore, the focus is on developing a method for adhoc localization of mobile robots in wireless networks in unknown environments. The development is done with a holistic view of the topic starting with a thorough examination of the complete transmission path of the radio signals to derive the measurement model. This also shows that an angle measurement in conjunction with an integrity indicator allows for robust localization using the signal strength. The developed prodedure uses the directional characteristic of the receiving antenna for angle determination by matching the measured and the reference characteristic. The similarity of the two serves as an indicator for the integrity of the measurement. Experimental results show high accuracy in outdoor and indoor areas.

Based on this measuring principle a probabilistic framework was developed that by continuously considering the uncertainties allows to simultaneously localize the wireless nodes and the mobile robot in unknown environments. Experimental results show that accuracy requirements for applications in mobile robotics are met.