Skip to Main Content
The experimental implementation and validation of a localization system based on a heterogeneous sensor network is described. The sensor network consists of ultrasound ranging sensors and Web cameras. They are used to localize a mobile robot under sensor communication constraints. Applying a recently proposed sensor fusion algorithm that explicitly takes communication delay and cost into account, it is shown that one can accurately trade off the estimation performance by using low quality ultrasound sensors with low processing time and low communication cost versus the use of the high-quality cameras with longer processing time and higher communication cost. It is shown that a periodic schedule of the sensors is suitable in many cases. The experimental setup is discussed in detail and experimental results are presented.