This paper provides an overview of a three year project conducted on the subject of self-localisation and autonomous navigation by a mobile robot. The research was carried out in two stages: Self-Localisation. Due to the fundamental unreliability of dead r

Self-Localisation and Autonomous Navigation by a Mobile Robot Tom Duckett and Ulrich Nehmzow Department of Computer Science University of Manchester Manchester M13 9PL United Kingdom 26.3.99 Abstract

duckettt@cs.man.ac.uk, ulrich@cs.man.ac.uk

This paper provides an overview of a three year project conducted on the subject of self-localisation and autonomous navigation by a mobile robot. The research was carried out in two stages: Self-Localisation. Due to the fundamental unreliability of dead reckoning, landmark-based methods were investigated. Several related issues were addressed, including performance evaluation, replication and comparison of existing work, and the development of a novel localisation system based on the results obtained. Autonomous Navigation. This work included the development of novel mechanisms for exploration and map building, as well as validating the previous work on self-localisation, in a complete, navigating mobile robot. Quantitative assessment of the di erent competences required for navigation was also carried out.

1 Introduction In recent years, there has been a great deal of interest in the topic of navigation by mobile robots. Reliable navigation systems are essential if mobile robots are to become widely used in industrial and domestic applications. However, while assembly robots are now ubiquitous in modern production plants, mobile robots are by comparison rarely found outside of scienti c research institutions at the present time. Perhaps the most fundamental competence required for autonomous navigation is that of self-localisation (location recognition). Without the ability to identify locations reliably, a mobile robot will inevitably become lost, and therefore unable to move purposefully between locations. We contend that the large scale use of mobile robots has been prevented so far, at least in part, by a lack of robust mechanisms for self-localisation. In this paper, we provide an overview of a three year PhD project carried out on the subject of mobile robot self-localisation and its application to the problem of autonomous mobile robot navigation. The work was carried out in two main stages, as follows: 1. Mobile Robot Self-Localisation. Due to the fundamental unreliability of self-localisation using dead reckoning, landmark-based methods were investigated. Several important issues were addressed, including quantitative assessment of performance, conducting empirical comparisons of existing approaches{ including replication of the work of other researchers, and the development of a novel localisation system based on the results obtained. 2. Autonomous Mobile Robot Navigation. A complete navigation system for a Nomad 200 mobile robot was developed, including novel mechanisms for exploration and map building, together with the previous work on self-localisation. Experiments were conducted in which complex, untreated environments of several hundred m2 were successfully mapped and navigated by the r

obot, requiring human intervention only to specify a goal location. Proc. TIMR 99\Towards Intelligent Mobile Robots", Bristol 1999. Technical Report Series, Department of Computer Science, Manchester University, ISSN 1361 - 6161. Report number UMCS-99-3-1. www.1mpi.com