AECID Fingerprinting Positioning Performance

The paper presents results from a field trial of the adaptive enhanced cell ID (AECID) fingerprinting positioning method that has been implemented by Ericsson AB for the GSM cellular system. The algorithm exploits own and neighbor cell global identities (CGIs), measured timing advance (TA) and quantized signal strength measurements to derive the radio fingerprint. The algorithm operates by collection of A-GPS measurements of opportunity, followed by a tagging by the fingerprint. All A-GPS measurements with the same fingerprint are then stored in separate clusters. In this way a large number of clusters covering the cellular network are created. Finally, a polygon is computed to describe the boundary of each cluster. The field trial of the AECID algorithm was performed in Shanghai, PRC, covering 15 dense urban and 11 suburban cells. About 4600 A-GPS reference measurements were collected for clustering and for performance evaluation. The results show that the inaccuracy of the AECID method is less than 25% of the inaccuracy of the CGI-TA positioning method, measured in terms of the area of the reported geographical regions.