Robot localization using wireless networks

This paper describes how IEEE 802.11b wireless networks can be used to determine the position of a robot inside a building. For this purpose we carried on some experiments which have shown that localization using only a wireless network as the only sensorial information (without a motion model) is not possible without a preconstructed sensorial map. But with the use of a motion model (easy to implement if we are using robots), localization can be achieved using theoretical models and MArkov localization techniques and we don’t need to bound the localization to the environment. We also believe that even it has been said that orientation affects signal strength [10] the changes are to slight and it’s not possible to use this information to obtain a good estimation of the orientation.