QRPp1-5: Hybrid TOA-RSS Based Localization Using Neural Networks

Recently considerable attention has been given to indoor geolocation using WLAN and WPAN technologies. With respect to positioning, the difference between the two technologies is the system bandwidth, which is around 25 MHz for IEEE 802.11 WLAN devices and at least 500 MHz for IEEE 802.15.3a WPAN devices employing UWB technology. Various algorithms using Received Signal Strength (RSS) or Time of Arrival (TOA) have been implemented for indoor positioning using these technologies. The performance of the TOA based system is sensitive to bandwidth and occurrence of undetected direct path (UDP) conditions, which degrade performance significantly. The performance of the RSS systems is less sensitive to the bandwidth and occurrence of UDP conditions. This paper presents a new hybrid RSS-TOA based localization algorithm using neural networks. The performance of this algorithm is compared with traditional TOA and RSS based algorithms in a common repeatable framework, representing a typical office environment.