 |
Efficient Quality of Service (QoS) Measurements in Wireless Mobile Networks
QoS measurements are essential for quality assurance and benchmarking; but usually they lead to high effort and costs. The direct application of mathematical methods is not straight forward; knowledge of the system and of the propagation conditions is needed to get verifiable results. Because of the effort the number of measurement samples is small. In this project we develop techniques to optimize QoS measurements and benchmarking based on knowledge of the considered system, propagation conditions and air interface.
Some main topics:
Route selection and measuring methods for reduced variance
The existing measurement data contains additional information that could be used to reduce the variance of the measurement results (e.g. space correlation could be considered as well as soft indicators like "almost dropped call").
Key Performance Indicators (KPIs)
For voice - and meanwhile for GPRS - the KPIs are well defined. But for benchmarking and for the characterization of the QoS performance of a network some "meta indicators" based on the KPIs are desirable.
Future services and networks
Just now UMTS services are starting or will start soon; this means there are new technologies and completely different system architectures. The applicability of the KPIs is to investigate or new QoS parameters have to be defined. WLAN services provided by hot-spots and overlay networks are also of interest.
Statistical extrapolation
Especially indoor measurements show a high measurement effort; usually it is only possible to measure inside public buildings. A statistical model, that is able to extrapolate the results of the outdoor measurements to get information about the indoor KPIs, would reduce the effort dramatically.
|
|
