Draft report ran wg4

R4-021479 Unit of layer 3 filtering (Motorola)

Sari Korpela (Nokia) questioned what layer 1 filter is used, Motorola clarified that this simulation is using an ideal layer 1 filter, measurements are provided each 200 ms exactly.

R4-021484 L3 filtering (Nokia)

Nokia remarks in this contribution that GSM RSSI uses logarithmic filtering, and therefore when comparing GSM power levels to UTRA power level the same type of filtering should be used. Motorola objected that a faster filter response is required in UTRA.

Nokia notes that the main influence on the measurement coming out of the layer 3 filter is not the layer 3 filter itself, but the layer 1 filter; therefore the difference in the results between log or linear layer 3 is small.

Qualcomm warned that it may be concluded from the graphics in this paper that the linear filtering would provide a response around 4 dBs away from the actual average value, which is incorrect.

R4-021534 L3 Filtering (Qualcomm)

Josef Blanz (Qualcomm) remarked that the linear filter follows faster the real signal in the case there is an increasing ramp, which means that the cell is getting closer and eventually would get included in the active set. Josef noted that it is of highest importance that a cell gets into the active set as fast as possible.

Don Zelmer asked the UE manufacturers if any had real products and data taken from real life on the behaviour of these two filters.
The chairman summarized that the option left are 3: WG4 could recommend to make mandatory in the specification either logarithmic, linear or leave it free for the manufacturers to chose. He also reminded that the reason GSM selected logarithmic was because at that time floating point processing was not available to introduce in the UE.
Finally, no agreement was reached. The chairman will report it to TSG RAN and WG4 will continue to study the issue for at least another meeting.