Return on Investment (roi) for Multi-Technology son juan Ramiro, Mark Austin and Khalid Hamied

A.4. Scenario and Working Assumptions

Both topology and traffic conditions may have an impact on the frequency and the characteristics (potential multipath propagation, number of reported sites, etc.) of the collected MRs, and hence, on the achievable accuracy. This Appendix studies two different RNCs corresponding to a sub-urban area in Europe and an urban area in US. Forty-five minutes of call traces have been collected in each scenario containing enough data, that is, more than 200 000 events, for a detailed analysis. Accuracy is measured by comparing the estimated geolocation with a drive-test performed during the period in which call traces were collected in the case of Europe, and with two static locations in the US scenario. Figure A.1 shows where MRs were collected during the drive test route followed in the European scenario. As commented in Section A.1, there are long periods of time with almost total absence of MRs, which makes proper traffic estimation in those areas more difficult.

A.5. Results

The presented analysis shows the performance of the proposed geo-location algorithm based on OTDs under two different scenarios from real networks. All results consider only events that can be directly located, i.e. those reporting at least three different sites, except in Section A.5.4, where the impact of using PD measurements is studied.

A.5.1. Reported Sites per Event

The amount of different reported sites is a key factor for an event to be properly geo-located. Figure A.2 shows the Probability Density Function (PDF) of the number of reported sites per event in both scenarios. As expected, events reporting fewer sites are more common, but it is important to focus on the weight that two-site events, which cannot be directly located, have in the PDF distribution. For the scenario in US, this percentage reaches around 60%, which may i ndicate that this network has better dominance, i.e. lower number of interferers as c ompared to the network from Europe.

19.18 19.185 19.19 19.195 19.2 19.205 19.21 19.215 19.22
Longitude [°]
Figure A.1 Location of MRs during the drive test route followed in the European scenario.

2 3 4 5 6 7 8 9
Number of reported sites per event [–]
Figure A.2 PDF of the number of different reported sites per event for scenarios in US and Europe.

Wrong SC
Multipath
Handover effect
No synchro. ref
Diverge
Geo-located
0 10 20 30 40 50 60 70 80
Frequency [%]
Figure A.3 PDF of the event status for scenarios in US and Europe, including only events reporting at least three different sites.