Recent development in wireless ad-hoc metropolitan area networks (WMANs) has increased the demand for a new service portfolio based on client location. Potential services include location sensitive billing, push-activated social-networking, location-based experiences, and applications exploiting least-cost routing. This paper proposes a discrete time solution for a scenario in which a client application measures and collects data from the indoor and outdoor radio environments and uploads the compressed bit-stream to a de-centralised server. The location of the mobile station is then estimated, within acceptable accuracies, at this server using a neural network approach.