1.

1 Optimization of Mobility Indicators

Mobility indicators mainly fall into three categories soft handovers, hard
handovers and inter-RAT handovers. The handover problems usually happen in
the following cases.
The best server changes too fast or there is no best server due to pilot pollution.
The handover is not prompt or there are pingpong handovers due to improper
parameter configuration.
Adjust the engineering parameters for antennas in areas with severe pilot pollution.
And adjust the handover parameters, such as the values of A, !, "#$, TTT %time
to trigger&, 'ysteresis and so on, to solve the problem that the handover is not
prompt or there are pingpong handovers. This section tries to solve this (ind of
problems through $M" data analysis and parameter optimi)ation.
1.1.1 Optimization of Soft Handovers
Missed neighboring cell configuration, pilot pollution, improper soft handover
parameter configuration, and e*uipment malfunctions are the major causes of soft
handover failures. To confirm the problems, the field +Ts are re*uired.
1.1.1.1 Soft Handover Optimization Flowchart
,igure -- .oft handover optimi)ation flowchart
1.1.1.2 Checin! Missed "ei!hborin! Cell Confi!#ration
/enerally spea(ing, most of the call drops at the beginning of the optimi)ation are
caused by missed neighboring cell configuration. The following methods are often
used to judge whether the call drops are caused by missed configuration of co-
fre*uency neighboring cells.
$bserve the active set Ec/Io information recorded by the 01 and the Best Server
Ec/Io information recorded by the .canner before the call drop. #f the former record is
very bad but the latter record is very good, then chec( whether the Best Server
scrambling code recorded by the .canner appears in the latest list of the neighboring
cells under intra-fre*uency measurement control. #f it does not, then the call drop is
caused by missed neighboring cell configuration.
#f the 01 re-accesses immediately after the call drop and the cell scrambling codes
during the 01 reaccess and those during the call drop are different, then the call drop
may also be caused by missed neighboring cell configuration. 2ou can confirm it
through measurement control %loo( bac(wards from the message of the call drop
event for the latest intra-fre*uency measurement control message and chec( the
neighboring cell list of this message&.
.ome 01 may report the Detected Set information. #f the corresponding scrambling
code appears in the Detected Set information before the call drop, then the call drop is
caused by missed neighboring cell configuration.
1.1.1.$ Checin! %ilot %oll#tion
&efinition of pilot poll#tion' 13cessive strong pilot signals e3ist at a certain point,
but none of them is strong enough to be the best server. To form pilot pollution, the
following points in this definition should be satisfied.
Stron! pilot si!nal' The absolute pilot signal strength is used to judge whether the
pilot signal is a strong one. The pilot signal strength can be evaluated through the pilot
R."4. #f the pilot R."4 e3ceeds a threshold, it is considered a strong pilot signal. The
formula is5
Absolute RSCP
Th RSCP CPICH
_
_ >
()cessive' The number of pilot signals is used to judge whether there are e3cessive
pilot signals at a certain point. #f the number e3ceeds a threshold, it is regarded that
e3cessive pilot signals e3ist at this point. The formula is5
N
Th Number CPICH > _
"one of them is stron! eno#!h to be the best server' The relative strength of a
pilot signal is a (ey factor in judging whether the pilot signal is strong enough. !ased
on the above definition and formulae, if the difference between the strength of the
strongest pilot signal and that of the ) 1 ( +
N
Th strongest pilot signal at this point is
less than a threshold, it is regarded that there is no pilot signal strong enough to be
the best server at this point. The formula is5
lative RSCP th Th best
Th RSCP CPICH RSCP CPICH
N
Re _ ) 1 (
) _ _ ( <
+
According to the above description, it is regarded that pilot pollution e3ists if the
following conditions are both satisfied.
The number of pilot signals that meet the condition
Absolute RSCP
Th RSCP CPICH
_
_ >
is more
than
N
Th
.

lative RSCP th Th best

Th RSCP CPICH RSCP CPICH
N
Re _ ) 1 (
) _ _ ( <
+
,or e3ample, during the networ( optimi)ation, suppose5
dBm Th
Absolute RSCP
95
_
=
,
3 =
N
Th
, and
dB Th
lative RSCP
5
Re _
=
, if the following
conditions are both satisfied, then it is regarded that pilot pollution e3ists.
The number of the pilot signals that meet the condition
dBm RSCP CPICH 95 _ >
is
more than 6.

dB RSCP CPICH RSCP CPICH

th best
5 ) _ _ (
4
<
1.1.1.* +d,#stin! Soft Handover +l!orithm %arameters
2ou can solve the following (inds of problems by adjusting handover algorithm
parameters5
. The handover is not prompt.
,rom the perspective of the ". service signaling flow, the symptom of this problem is
that the 01 cannot receive Active Set Update %physical channel reallocation in the
case of the intra-fre*uency hard handover& because after the 01 reports the
measurement report, the source cell has a fast reduction in 1c7#o. 8hen the R9"
sends Active Set Update, the 01 has closed the transmitter due to the loss of
downlin( synchroni)ation. :iewed from the 01 side, it cannot receive Active Set
Update. #n the 4. services, if the 01 cannot receive Active Set Update or TR! resets
before the handover, the handover will also fail.
,rom the perspective of signals, the following phenomena may accompany this
problem.
"orner effect5 1c7#o of the source cell decreases drastically, and 1c7#o of
the target cell increases sharply %very high when it appears&.
,ast fading5 1c7#o of the source cell decreases *uic(ly for a while and
then increases, and 1c7#o of the target cell increases for a short while.
;. 4ingpong handovers. The following phenomena may accompany this problem.
The best server changes *uic(ly5 Two or more cells ta(e turns to be the best server.
!ut as the best server, none of the cells can last long though they has good R."4s
and 1c7#os.
There is no best server5 There are multiple cells. Their R."4s are normal and similar
to each other. !ut 1c7#o of every cell is very bad.
,rom the perspective of the signaling flow, 1vent A is reported immediately after
one cell is deleted. !ecause the 01 cannot receive Active Set Update from the
R9", the handover fails.
1.1.1.- (.#ipment Malf#nctions
,irst chec( the alarm console to see whether there are abnormal alarms, and
analy)e the message traces at the same time. ,ind out in which step the soft
handover fails. "hec( the failure message, and contact the local product
maintaining engineer to confirm whether the e*uipment has malfunctions.
1.1.1./ Sol#tions
"orresponding adjustments should be ta(en for the confirmed problems.
'andover failures caused by pilot pollution5 Adjust the engineering parameters of a
certain antenna to set this antenna as the best server in this interfered location. #f the
power of one of its sectors is reduced, then #o of the pilot pollution area will decrease<
even if the powers of other pilots are not adjusted, 1c7#o will also increase. Thereby
the 1c7#o differences with other scrambling codes in the active set will become larger
and pilot pollution will be eliminated. Through a lot of research, =T1 has proved that
the reduction in the pilot transmit power will not change cell capacity greatly. #f
condition allows, new base stations can be added to cover this area.
1*uipment malfunctions5 "onsult the customer service engineers, and as( them to
help chec( whether there are alarms and whether the transport layer is abnormal. #f
there are alarms, coordinate with the customer service engineers and the engineering
personnel to solve the problems.
"all drops because the handovers are not prompt5
Adjust the antenna to e3pand the handover )one.
"onfigure the 1vent A handover parameters to ma(e the handover
easier to happen.
#ncrease CIO to ma(e the handover happen earlier in the target cell. The
sum of CIO and the actually measured value is used for judging the 01
events, including the 01 intra-fre*uency handover. CIO helps shift the cell
border in the handover algorithm. #f CIO is configured with a larger value,
the handover will be easier to happen and there will be more 01 in the
soft-handover status, but more resources will be occupied. #f CIO is
configured with a smaller value, the soft handover will be more difficult to
happen and the receiving *uality may be impaired. A CIO of about >d! is
*uite good for eliminating the fast fading and the corner effect, but this
configuration has some side effects, such as the increase of handover
proportion.
"all drops caused by pingpong handovers5 Adjust the antenna to form a best server in
its coverage )one or set the 1vent 1B handover parameters %increase the threshold of
1vent 1B, the 1vent 1B hysteresis or the time to trigger 1vent 1B& to increase the
difficulty in deleting the active set.