Neural Fuzzy Based Indoor Localization by Extending Kalman Filtering with Propagation Channel Modeling

In this chapter, an indoor localization based on the Received Signal Strength Indication (RSSI) from the Wireless Local Area Network (WLAN) and the Adaptive Neural Fuzzy Inference system (ANFIS) is proposed. Over the last few years, the wireless local area networks based on IEEE802.11b (also named WiFi) and location-based service have been flourishing and increased demand. The wireless networks have become a critical part of the networking infrastructure, and capable for mobile devices equipped with the WLAN to receive the radio signal for networking. Location-aware computing is a recent interesting research issue that exploits the possibilities of modern communication technology due to the location-based service which has a great potential in areas such as library or museum tourguide, free mobility and nursing at home, patient transporting in the hospital and easy going capability in the shopping mall. The location of a mobile terminal can be estimated by using the strength of the radio signal with the WLAN. However, the unpredictability of signal propagation through indoor environments is the key challenge in the indoor positioning from the strength of the WLAN signals. It is difficult to provide an adequate likelihood model of signal strength measurements. Thus, the main focus of research in this area have been on the development of the technique that can generate the accurate empirical model from the training data collected in the test area and the real-time recursive estimation for the mobile user. There are many important propagation models based on the localization techniques. (Bahl & Padmanabhan, 2000; Wassi et al., 2005; Li, B.; Salter & Dempster, 2006) addressed the range based approach, such as pattern matching that relies on the range measurement to compute the position of the unknown node. (Kotanen et al., 2003) offered the design and implementation of a Bluetooth Local Positioning Application (BLPA). First, they converted the received signal power to the distance, according to a theoretical radio propagation model, and then, the Extended Kalman Filter (EKF) is used to compute 3-D position with the basis of distance estimates. (Ito & Kawaguchi, 2005) introduced a Bayesian frame work for indoor positioning over the IEEE 802.11 infrastructure, which investigated the direction and distribution of the signal strength for the pre-observation model, and then a location of the mobile device can be computed by using the Bayesian filtering (Fox et al., 2003). (Seshadri, 2005) provided a