Self-localization capable mobile sensor nodes

Abstract

Position sensing systems for indoor environments have one common problem: The setup effort of the whole system. GPS-supported localization is simple and fast to implement and the reference grid is ubiquitous. However, in locations where the signals of GPS are not available, e.g. indoor, or where the technique is not accurate enough, other position sensing systems have to be used. Usually, a reference grid will be provided by an administrator: Reference points must be deployed and programmed with the correct location information. Especially for temporary installations and dynamic surroundings the setup costs are very high. In our work, we developed a robotic platform, which is capable of deploying itself autonomously and which can operate as a mobile reference point equipped with ultra sound transmitters. Relying on a combination of two controllers and an odometry system, the platform is able to drive on a complex spline-like trace using only a few supporting points. We also investigated the problem for the localization system itself. The detection field of our ultra sound based system is a full hemisphere. Within this domain, we can measure the distances between nodes exploiting the time of flight technique as well as the angle of arrival. Finally, the system can also be used as a sonar system for passive obstacles detection.