Self-localisation in indoor environments combining learning and evolution with wireless networks

Abstract

This work combines wireless networks (WLAN) (Wireless LAN IEEE 802.11 b/g) with learning and evolution of artificial neural networks. Our main objective is to propose an architecture for a self-adaptive system, addressing alternative methods to the usage of GPS for self-localisation in autonomous mobile robots either in indoor or outdoor environments. We seek to describe alternatives and evaluation methods for localisation of mobile agents using the strength signal from Access Points (APs). The results show that the proposed method used with autonomous mobile robots does not require the use of special hardware, and hence is low cost, easy to operate, and fully autonomous.