WO 95/24810 ~ PCT/US95/02517
1
METHOD AND SYSTEM FOR CHANNEL ALLOCATION
USING POWER CONTROL AND MOBILE-ASSISTED
HANDOVER MEASUREMENTS
The present invention relates in general to base
radiocommunication systems and, in particular, to
channel allocation combined with power control in a
mobile radio communication system.
The concept of frequency reuse is at the heart of
cellular technology. In the conventional sense,
frequency reuse is a technique whereby groups of
frequencies are allocated for use in regions of limited
geographic coverage known as cells. Cells containing
equivalent groups of frequencies are geographically
separated to allow callers in different cells to
simultaneously use the same frequency without
interfering with each other. By so doing many
thousands of subscribers may be served by a system of
only several hundred frequencies. The design and
operation of such a system is described in an article
entitled Advanced Mobile Phone Service by Blecher, IEEE
Transactions on Vehicular Technology, Vol. VT29, No. 2,
May, 1980, pp. 238-244. Known commonly as the AMPS
system, this syste7n had allocated to it by the FCC a
block of the UHF frequency spectrum further subdivided
into pairs of narrow frequency bands called channels.
Pairing results from the frequency duplex arrangement
wherein the transmit and receive frequencies are offset
by 45 MHz. At present there are 832, 30 kHz wide,
channels allocated to cellular mobile communications in
the United States. A table of the frequencies
dedicated to mobile communications in the U.S. is shown
in Figure 1. It is worth noting at this point that of
the 832 available channels, there are 21 control
WO 95/24810 ~ ''~ ~~ PCTlUS95l02517
2
channels dedicated each to the A-carrier and the B-
carrier. These 42 control channels provide system
information and cannot be used for voice traffic. The
remaining 790 channels, known as voice or traffic
channels, carry the burden of voice communication and
are equally divided between the A-carrier and the B-
carrier. A particular user can access at least half of
the available channels, or 395. With regard to TDIrIA
systems, such as specified by the IS-54B standard,
these channels are further divided into 3 time slots.
In this instance, a given user can access 3x395, or
1185, nchannels".
Link quality is the benchmark of any radio
communication system. To provide high quality voice
communication the desired signal in a cellular system
must maintain a minimum signal strength above all other
interference. The ratio of the desired signal to the
interference is known as C/I. Aside from noise, which
is omnipresent, there are fundamentally two other types
of interference with which a designer must contend.
The first of these is interference arising from users
simultaneously operating on the same channel. This is
known as co-channel interference. The second source of
interference is from users operating on adjacent
channels. This' is known as adjacent-channel
interference. Adjacent channel interference is
controlled by selecting the frequencies in a given cell
to be separated by large frequency increments, e.g.,
200 kHz, and by using sharp cutoff in the channel
filters in order to obtain a high adjacent-channel
suppression. Co-channel interference is reduced by use
of a frequency reuse pattern which geographically
separates cells with the same frequency group. An
example of an ideal seven cell frequency reuse pattern
is shown in Figure 2(a).
WO 95/24810 ~ ~ ~ PCT/US95102517
3
Frequency . planning is the . process by which ~, .
individual channels are assigned to cells within the
network. Currently, most frequency planning is done a
priori; that is a fixed frequency plan is "hard-wired"
in place by each cellular system operator. This is
known as fixed channel allocation, or FCA. However, as
interference and traffic load are time varying, FCA is
not optimal. As shown in Figure 2(b), highways which
bisect cellular boundaries may have significantly
l0 differing traffic patterns depending on location and
time of day. Some roads may have significant
automobile traffic in the morning and very little in
the afternoon. As a result, most fixed frequency plans
are not very efficient; many channels in a fixed
frequency plan will have a much better link quality
than is necessary to achieve high quality voice
communication while many others in the same system will
suffer from poor link quality which might force them to
be dropped or blocked. A capacity increase could be
obtained by some form of channel allocation where all
of the links have equal quality. Because of the time
varying nature of the interference, an adaptive scheme
must be used.
Adaptive channel allocation, or ACA, is a method
of dynamically allocating frequencies throughout a
cellular system to maximize system capacity. Under an
ACA scheme, more frequencies would be allocated to busy
cells from more lightly loaded cells. In addition, the
channels can be allocated such that all links have
satisfactory quality.
The concept of ACA is well-known to those skilled
in the art. Many publications have illustrated the '
potential for ACA yet do not discuss specific
strategies. For example, "Capacity Improvement by
Adaptive Channel Allocation", by H$kan Eriksson, IEEE
a
wo 9s124810 ~ ~, ~ ~, ~, y .'~~ PGTIUS9sl02517
4
Global Telecomin. Conf . , Nov. 28-Dec. I; 1988, pp. 1355-
1359, illustrates the capacity gains associated with a
cellular radio system where all of the channels are a
common resource shared by all base stations . In the
above-referenced report, the mobile measures the signal
quality of the downlink and channels are assigned on
the basis of selecting the channel with the highest C/I
level.
Another approach is described by G. Riva,
"Performance Analysis of an Improved Dynamic Channel
Allocation Scheme for Cellular Mobile Radio Systems",
42nd IEEE Veh. Tech. Conf., Denver, 1992, pp. 794-797
where the channel is selected based on achieving a
quality close to or little better than a required C/I
threshold. Furuya Y. et al., "Channel Segregation, A
Distributed Adaptive Channel Allocation Scheme for
Mobile Communications Systems", Second Nordic Seminar
on Digital Land Mobile Radio Communication, Stockholm,
October 14-16, 1986, pp. 311-315 describe an ACA system
wherein the recent history of link quality is
considered as a factor in allocation decisions. In
addition several hybrid systems have been presented
where ACA is applied to a small block of frequencies on
top of an FCA scheme. Such an example is presented in
Sallberg, K., et al., "Hybrid channel assignment and
reuse partitioning in a cellular mobile telephone
system", Proc. IEEE VTC X87, 1987, pp. 405-411.
A common denominator for all of these ACA systems
is that they allocate a channel out of the set of
channels which fulfills some predetermined quality
criteria. The difference in each is how the channel is
chosen out of the set. Apart from increases in system
capacity, adaptive channel allocation obviates the need
for system planning. Planning is instead performed by
the system itself; this is particularly attractive when
c
v
~1~~~~
WO 95/24810 PCTJUS95/02517
system changes are implemented or new base stations are
added.
Adaptive power control, or APC, is also a known
art in cellular systems. See, for example, U.S. Patent
5 4, 485, 486 to Webb et al. With APC, the power of the
transmitter is varied according to the needs in the
receiver. In general, there are two types of adaptive
power control schemes: the C-based and the C/I-based.
In the C-based scheme, the signal strength level at the
receive side is maintained at a predefined level. As
soon as the (average) received signal strength deviates
from this level, the transmitter is ordered by the
receiver to increase or decrease its transmit power.
C-based APC only responds to changes in the path loss
whereas C/I-based APC tries to maintain a predefined
C/I level at the receiver. In addition to changes in
the path losses, changes in the interference condition
also result in the transmit power being adjusted.
The conventional allocation algorithms described
above base their decisions on the knowledge of which
channels are used by which base stations and then
attempt to optimize the quality in each link. However,
they do not take advantage of the possibilities offered
by adaptive power control in the mobiles and base
transmitters.
In a TDMA environment, the allocation decision
includes more than just a selection of the base and
channel combination. Since TDMA channels are broken up
into time slots, the allocation decision should also
take this into account. U.S. Patent 4,866,710 to
Schaeffer, for example, describes a method of
allocating frequency and timeslots to mobile stations
such that all the timeslots on a given frequency 'are
filled before allocating timeslots on another
frequency. Although seemingly efficient, this scheme
. ._ ~ ~1~~~~
wo 9snasio pcr~s9s~ozsm
6
does not consider the contributions to interference and
does not consider~the possibility of adaptive power
control.
BUl~RY
It is therefore an object of the present invention
to provide a method for dynamically allocating channels
in a communication system which maximizes system
capacity while minimizing the transmitted power of the
mobile radiotelephones. It is a further object of the
present invention to take advantage of measurements
made by the mobile radiotelephones in determining link
quality. It is yet another object of the present
invention to enable battery operated radiotelephones to
enjoy extended battery life by reducing the transmitted
power.
The cantrol scheme presented herein adapts to the
current traffic and interference situation in a
communication environment in order to optimize the
quality of each link and maximize the overall system
capacity. The status of the current traffic and
interference condition is derived from measurements
taken both by the mobile station and the base station.
The channel allocations are periodically updated to
ensure that, on average, the least amount of transmit
power is used on the channels. Once a channel has been
allocated, the adaptive power control scheme tries to
maintain a satisfactory link quality with the minimum
amount of radiated power. Since the adaptive power
control and the adaptive channel allocation form one
integrated process, the term adaptive channel
allocation and power control, or ACAPC, is hereinafter
adopted to describe the present invention. The present
invention is not restricted to a particular type of
CA 02162256 2003-12-30
7
access scheme and can therefore be applied equally, for
example, to FDMA, TDMA, CDMA or hybrid systems.
Accordingly, in one aspect, the invention provides a
method for assigning uplink radio channels in a radio
communication system, the method comprising the steps of
(a) measuring, in a mobile station, received signal
strength indications (RSSIs) of control signals broadcast
from at least one base station, (b) determining a path loss
between the mobile station and the at least one base
station using the RSSI measurements, (c) measuring, in the
at least one base station, an RSSI of interference signals
on a plurality of available traffic channels, (d)
determining transmit powers required for the mobile station
to produce a signal on each of the plurality of available
traffic channels at the at least one base station, wherein
a strength of the signal is a predetermined level above a
corresponding RSSI interference level measured on a traffic
channel taking into consideration the path loss, and (e)
assigning one of the plurality of available traffic
channels as an uplink channel based on the determined
transmit powers.
In another aspect, the invention provides a method for
assigning downlink radio channels in a radio communication
system, the method comprising the steps of (a) measuring,
in a mobile station, received signal strength indications
(RSSIs) of control signals broadcast from at least one base
station, (b) determining a path loss between the mobile
station and the at least one base station using the RSSI
measurements, (c) measuring, in the mobile station, an RSSI
of interference signals on a plurality of available traffic
channels, (d) determining transmit powers required of the
at least one base station to produce, at the mobile
station, a signal on each of the plurality of available
CA 02162256 2003-12-30
7a
traffic channels having a signal strength which is a
predetermined level above a corresponding RSSI interference
level measured on a traffic channel taking into
consideration the path loss, and (e) assigning one of the
plurality of available traffic channels as a downlink
channel based on the determined transmit powers.
The invention also provides a radiocommunication
system comprising means, disposed in a mobile station, for
measuring received signal strength indications (RSSIs) of
control signals broadcast from at least one base station,
means for determining a path loss between the mobile
station and the at least one base station using the RSSI
measurements, means, disposed in the mobile station, for
IS measuring an RSSI of interference signals on a plurality of
available traffic channels, means for determining transmit
powers required of the at least one base station to
produce, at the mobile station, a signal on each of the
plurality of available traffic channels having a signal
strength which is a predetermined level above a
corresponding RSSI interference level measured on the
traffic channel taking into consideration the path loss,
and means for assigning one of the plurality of available
traffic channels as a downlink channel based on the
determined transmit powers.
In yet another aspect, the invention provides a time
division multiple access (TDMA) radiocommunication system
comprising means, disposed in a mobila station, for
measuring received signal strength indications (RSSIs) of
control signals broadcast from a first base station, means
for determining a path loss between the mobile station and
the first base station using the RSSI measurements, means
for predicting RSSIs of interference signals emanating from
CA 02162256 2003-12-30
7b
a second base station on a plurality of available traffic
channels, means for determining a transmit power, on each
carrier having one slot open, needed for each base station
to produce, at the mobile station, a signal on the
plurality of available traffic channels having a signal
strength which is a predetermined level above a
corresponding interference level measured on that traffic
channel, means for determining a transmit power, on each
carrier having two slots open, needed for each base station
to produce, at the mobile station, a signal on the
plurality of available traffic channels having a signal
strength which is a predetermined level above the
interference level measured on that traffic channel, means
for determining a transmit power, on each idle carrier,
needed for each base station to produce, at the mobile
station, a signal on the plurality of available traffic
channels having a strength which is a predetermined level
above a corresponding interference level measured on that
traffic channel, and means for assigning a channel based on
the determined transmit powers.
In exemplary embodiments of the present invention,
systems and methods allocate a channel that minimizes the
average transmit powers. During updates, which are
periodically performed, it is checked whether the average
transmit power was indeed low, or that another channel can
be found on which the transmit power can be even lower.
Between channel updates, the APC tries to maintain the
required C/I level in the receiver. The scheme ensures
minimizing the transmit power which not only reduces
interference to other channels, but also serves the
laudable goal of extending the battery life of handheld
mobile radiotelephones.
CA 02162256 2003-12-30
7C
BRIEF DESCRIPTION OF THE DRAWINGS
These and other objects, features and advantages of
the present invention will be readily apparent to one
skilled in the art from the following written description,
read in conjunction with the drawings, in which:
Figure 1 is an illustration of the allocated frequency
spectrum as per the U.S. standard IS-54B;
Figure 2a is an exemplary illustration of a frequency
reuse pattern as employed in a fixed plan cellular system;
Figure 2b is an exemplary illustration of the time
dependent characteristics of cellular system loading which
illustrates the need for an adaptive channel allocation
according to the present invention;
IS Figure 3 is an illustration of uplink and downlink
interference;
Figure 4 is a flow chart illustrating the basic
operation of an exemplary embodiment of the present
invention;
25
WO 95/24810 ~ .~~, ? ~ PCTNS95/02517
8
Figure 5 is a flowchart illustrating the uplink
allocation according to an exemplary embodiment of the
present invention;
Figure 6 is a flow chart illustrating the downlink
allocation according to an exemplary embodiment of the
invention;
Figure 7 is an illustration of the IS-54H TDMA
frame structure;
Figure 8(a) is a graph illustrating an exemplary
power spectrum of a transmitted signal;
Figure 8(b) is an exemplary filter characteristics
of a receiver;
Figure 8(c) is a graph showing the adjacent
channel'interference for the examples of Figures 8(a)
and 8 (b) ;
Figure 9 is an exemplary illustration of the
downlink interference prediction;
Figure 10 is a block diagram of an exemplary
mobile station according to the present invention;
Figure li is a block diagram of an exemplary base
station according to the present invention; and
Figure 12 is a block diagram of a portion of an
exemplary base station controller according to the
present invention.
DETAILED DESCRIPTION
In the following description, for purposes of
explanation and not limitation, specific details are
set forth, such as particular circuits, circuit
components, techniques, etc. in order to provide a
thorough understanding of the invention. However it
will be apparent to one skilled in the art that the
present invention may be practiced in other embodiments
that depart from these specific details. In other
instances, detailed descriptions of well-known methods,
WO 95IZ4810 PCT/US95/02517
~1'~~~~~
9
devices,. and, circuits are omitted sows not to obscure
the description of the present invention with
unnecessary detail.
The implementation of the present invention will
vary depending on the particular requirements of the
communications system to which it is applied. In
frequency duplex systems (e.g., AMPS, IS-54B), for
example, the uplink and downlink channels are paired,
separated by a fixed 45 Mhz offset. The assignment of
a downlink channel thereby automatically fixes the
uplink channel and vice versa. However, a more general
case occurs when the downlink and the uplink can be
independently allocated.
An exemplary embodiment of the present invention
will therefore first be presented for the case in which
the uplink and downlink channels can be selected
independently, and for which the transmit power can be
independently controlled on any channel. Later, other
exemplary embodiments will be described for systems
where: the uplink and downlink channels occur in pairs,
for systems with limited measuring capabilities in the
mobile, and for systems where the APC in the downlink
operates on a group of channels as is the case in the
current TDMA system defined by IS-54B standard. As
will be discussed, each of the embodiments will be
influenced, in some degree, by the operating status of
the mobile station.
In most cases, the mobile station operates in one
of three possible modes. The first to be considered
occurs when the mobile station is in active
communication (i.e., a call is in progress) and where
the mobile station maintains active knowledge of the
radio environment. The second mode to be considered is
the stand-by mode. When in stand-by mode, the mobile
station is listening for pages by periodically scanning
._
wo 9snaslo pcrius95~o2sm
to
the available control channels and therefore has same
limited knowledge of the radio environment. Fast to be
considered is the power on mode. In this mode the
mobile station has been completely deactivated and has
no a priori knowledge of its environment when first
turned-on.
The channel allocation scheme described by this
invention makes use of periodic measurements performed
both by ane or more base stations and/or by one or more
mobile stations to determine the best channel under the
current radio environment. Note that the term mobile
station is used herein to refer to any remote station,
for example vehicular radiotelephones and hand-held
radiotelephones. In the following description, it is
assumed that each base station transmits a pilot signal
(i.e., a set-up or control channel) at a known
frequency and of a known power level.
As will be described in greater detail, the
allocation process according to this exemplary
embodiment of the present invention can be divided into
three general phases. As shown in Figure 4, these are:
1) acquisition 410 of measurement data reflecting the
current conditions of the local environment;
2) determination 420 of the most appropriate channel
consistent with the measurements made during
acquisition; and 3) channel allocation 430 according to
predetermined criteria.
The acquisition phase 410 begins with measurements
of the signal power of the pilot signals broadcast from
surrounding base stations. These measurements are
performed by the mobile station which, through a
procedure to be described in greater detail,
periodically measures the received signal strength
(RSSI) of a number of individual pilot signals
transmitted from surrounding base stations.
CA 02162256 2003-12-30
WO 95/24810 PCTIUS95102517
11
The RSSI of a signal can be measured,
for example, as described in U.S. Patent
5,048,059 to Paul W. Dent entitled "Log-Polar
Signal Processing". The RSSI of a channel
is, for this example, simply a measure of the amount of
signal power contained within a 30 Khz bandwidth
centered at a particular frequency. In addition to
signals intentionally broadcast on the channel, the
signal power may include co-channel signal power,
spillover from adjacent channels, noise, and any other
power which exists in the band. In determining the
RSSI of each individual downlink signal, the mobile
station averages a plurality of individual RSSI
measurements (e. g., over 1 second) to smooth out fast
fading phenomena. For example, a vehicle equipped with
a mobile radio operating at 869 l~iz and traveling at 40
I~H should ideally perform RSSI measurements at
approximately 7 millisecond intervals to have those
measurements uncorrelated to Rayleigh fading. As a
2o typical radio receiver can tune to a specific
frequency, perform an RSSI measurement and return to
within a few hundred Hz of its initial frequency on the
order of 3 milliseconds, such a requirement is not
difficult to achieve. To accommodate varying speeds
and frequencies, many such measurements are performed
and then averaged.
The acquisition procedure will be described with
regard to the typical example of a mobile radio
communication arrangement illustrated in Figure 3. It
will be appreciated by those skilled in the art that a
typical system can, and most likely will, include many
more base stations and many more mobile stations than
shown in this exemplary illustration. However, to
avoid obfuscating the present invention, the two
WO 95124810 ~ ~ PCTliTS95102517
12
mobile, three base station arrangement of Figure 3 is
presented. ~~ .
The acquisition operation, indeed operation of the
entire network, is controlled via the mobile telephone
switching office (MTSO) 300. The MTSO 300 is
connected, either directly or indirectly, to each base
station controller 310, 320, and 330. This description
will first consider the acquisition procedure when a
call is presently in progress. In this example, mobile
station A (numbered 3?0) is currently connected to base
station A (numbered 350).
As well-described in the IS-54B specification,
when a call is in progress, the base station may issue
instructions to a mobile station via the FACCH. The
FACCH, or fast associated control channel, is a "blank
and burst" data transmission protocol whereby messages
may be sent over the voice channel. Included in the
list of FACCH messages is the "Channel Scan Message".
This instruction prompts the mobile station to perform
a series of RSSI measurements on a list of frequencies
included with the message. Often, this list includes
the control channel frequencies broadcast from
surrounding base stations. For specific details of
FACCH signaling and the Channel Scan Message, the
interested reader is referred to the IS-54B document.
Note that this particular method of controlling the
mobile station is presented for the purposes of
illustration only. It will be appreciated by those
skilled in the art that other signaling methods and
formats could also be used in the execution of the
present invention.
In response to instructions generated by the MTSO
300, mobile station A monitors the signals transmitted
by surrounding base stations, including that to which
it is currently connected. With regard to the
~lb~~~~
WO 95124810 PCT/US95I02517
13
situation illustrated in Figure.4,_ mobile station A is
located so that it receives signals not only from the
base station to which it is connected (i.e., signal 345
transmitted from base station A) but from several other
surrounding base stations: signal 305 transmitted from
base station B (numbered 340), and signal 315
transmitted from base station C (numbered 360).
According to the present invention, mobile station A
periodically receives a channel scan message broadcast
from base station A. This message includes
instructions to scan the pilot signals from base
stations located in the vicinity of base station A.
Note that the actual base station location andJor
identification need not be transmitted to the mobile
station; the base stations need only be identified by
the frequency of the pilot signal broadcast therefrom.
After making RSSI measurements (as described
above) of the three downlink signals 305, 315, and 345,
mobile station A reports these results to base station
controller A (numbered 320), via the slow associated
control channel (SACCH) of an uplink traffic channel
335. As with the FACCH, the SACCH is a signaling
format specified by IS-54B.
The aforementioned RSSI measurements can now be
used to calculate the path losses between the base
stations whose signals were measured and the mobile
station performing the measurements. The RSSI
measurement together with the base transmit power
(which is either fixed, or otherwise known by the
network 300 and reported to the base station controller
of base station A) provide all the information required
to calculate the signal path loss (PL) for each signal.
Path loss, calculated as the RSSI divided by the base
transmit power, is an expression of the attenuation
that a signal will experience as it propagates between
. ~ ~~.~'~~~~
WO 95/24810 PCT/US95I02517
14
the. base station and the mobile station. For example,
assume that mobile station A has measured the RSSI of
signal 315 broadcast from base station C to be -
125 dBm. It is also known that signal 315 is broadcast
at a power level of 0 dBm. Therefore, the calculation
of path loss is straightforward:
Known Transit Power of Base Station C = p dgm
Measured RSSI of Base Station C by mobile station = -125 d8m
Path Loas between mobile station A and base station C =125 d8
In general, the path loss between a base station
of arbitrary index J and the mobile station performing
the measurement will be represented by PL(J). Since
the pilot RSSI measurements are assumed to be
uncorrelated to Rayleigh fading, the path loss from the
base to the mobile can, by the principle of
reciprocity, be assumed to be identical to the path
loss from the mobile station to the base. PL(J)
therefore represents the link path loss irrespective of
link direction (i.e., uplink or downlink). Thus
calculation of the path losses between the mobile
station and surrounding base stations completes the
first stage of the acquisition process.
The second acquisition stage involves the
measurement of RSSY on idle traffic channels (i.e., the
potential voice channels which may be allocated) to
obtain an estimate of the interference levels. If no
restrictions are set on the selection of channels, and
if the mobile station and the base station have the
same scanning capabilities, the process according to
the present invention operates independently for the
uplink and the downlink channels. First, the uplink
routine is described.
Under instructions from the MTSO 300, base station
A continuously monitors the interference levels on its
- . _ - ~16a2~~
wo 9snasio pcrius9s~o2sm
uplink and the downlink channels. First, the~uplink
routine is described.
Under instructions from the MTSO 300, base station
A continuously monitors the interference levels on its
5 idle traffic channels (i.e., voice channels not in
use). The base station scans through all the idle
channels, takes RSSI measurement samples and, as
before, averages the samples (e.g., over a period of 10
seconds) to make the samples uncorrelated to Rayleigh
10 fading. Note that as with the pilot signal
measurements, these RSSI measurements include any
interfering power that falls within the measuring
bandwidth - irrespective of where the interference
comes from. The energy can come from, for example:
15 noise (which is ubiquitous), co-channel users, the
spill-over of users on adjacent channels,
intermodulation products that happen to fall into the
measuring bandwidth, and from non-cellular emissions
(both licensed and unlicensed.) With regard to Figure
3, this process is repeated (either sequentially or
simultaneously) by base station B and base station C.
The average uplink interference level on channel
K at base J is represented by I~(J,K). The I~ values
of all idle channels, K, at base J are now sent via the
base station controller of base station J to the base
station controller of the base station to which the
considered mobile station is currently connected. For
- the example shown in Figure 4, this means that the
values measured on idle channels by base stations A, B
and C (i.e., I~(A,K) , I~(B,K) , and I~(C,K) ,
respectively) are sent to base station controller 320
of base A. Index K, of course, represents the voice
channels and can, in this example, be any number from
1 to 395.
W0 95124810
PCT/US95/02517
16
Buring the acquisition process, the. path loss
information calculated from the pilot signal
measurements sent by the mobile station and the
interference information sent by the surrounding base
stations provide the base station controller with the
information necessary to calculate all the power levels
required on all of the channels that can be allocated.
The acquisition phase 410 is then completed and the
next step is the determination of the best channel to
allocate.
The determination phase 420 includes finding the
base station, J, and channel, R, combination that
requires the least amount of mobile station transmit
power in order to maintain satisfactory quality on the
link. For each base station, J, and channel, R, the
required mobile station transmit power P~,~ (J, R) can be
calculated as:
Pxs, zec ( J, K) _ ( i ) o +PL ( J) +I~ ( J, K) dB ( 1 )
As previously described, all the information
required~to solve equation (1) now resides in base
station controller 320. (C/I)o is the target C/I value
that the system tries to maintain on a link and can be
defined by the network operator. For example, if it is
assumed that the path loss between the mobile station
in question and a base station J, PL(J), is calculated
as before to be 125 dB and the uplink interference on
channel R at base station J is -150 dBm, and the
desired C/I is 25 dB, then the required transmit power
can be calculated to be:
P~,,~(J,K) - 25 dB + 125 dB - 150 dBm = 0 dBm (1.0 m
Watts)
Wo ~~4slo 216 2 2
PCT/US95l02517
17
If the desired object is to minimize~the~transmit
power of the mobile station then that channel R on base
J with the lowest P~~ is chosen as the best base and
channel combination. Notice that, since the exemplary
process according the present invention takes
interference into account, this process does not
necessarily choose the strongest base (i.e.,
geographically closest to the mobile, or lowest path
loss) as the one with which to establish a link. If
the interference level of an available voice channel on
a more distant (i.e., in this context correlating to
higher path loss) base is significantly lower than the
interference levels on any of the idle channels of a
closer base station, the required P~ to connect to the
more distant base might actually be lower.
For example, assume the path loss, PL, on downlink
345 from base A to mobile station A is 80 dB and the
path loss on downlink 305 from base B to mobile station
A is 90 dB. Assume further that the lowest
interference measured at base A is on channel 32
(825.96 l~iz) at a level of -100 dBm. Assume also that
the lowest interference level measured at base B occurs
on channel 245 (832.35 I~iz) at a level of -120 dBm. If
the target (C/I)o is 25 dB, then the minimum required
P~ for a link to base A on channel 32 is calculated
according to equation (1) to be:
P~.~ (A,32) - 25 dB + 80 dB - 100 dBm = +5 dBm
For a link to base B on channel 245 the minimum
required transmit power would be:
3 0 P~"s,,~ ( B , 2 4 5 ) - 2 5 d8 + 9 0 dB - 12 0 dBm = -5 dBm
If both base stations transmit at the same power level
the downlink signal received at the mobile station from
base B is weaker (i.e., higher path loss and ostensibly
further away). Assuming reciprocity, the signal
WO 95/24810 ~ PCTlUS95/02517
18
received at base station B from the mobile' station
would be weaker than that received at base station A.
However, because of the consideration of interference,
even with an additional 10 dB of path loss, less
transmit power is required by the mobile station to set
up a satisfactory link to base station B. Therefore,
for the uplink, the mobile station would be handed-off
to channel 245 on base station B. This stands in
marked contrast to the earlier-described systems in
which the mobile station is handed-off to the base
station with the strangest signal.
In addition to the power levels required on
potential channels, the base station controller is
provided with the prevailing transmit power on the
current traffic channel. The mobile can, for example,
send this information over the SACCH to the base to
which it is connected. This information can then be
directed to the base station controller of base
station A.
If the best channel and base station combination
is not identical to the current channel and base
station combination, a re-allocation is considered.
However, in order to avoid channel hopping and Ping-
Pong effects (the mobile station being repeatedly re-
allocated between channels A and B), a hysteresis is
built in. The new channel is only allocated if the
required transmit power on the new channel is at least
x dB smaller than the transmit power on the current
channel. The hysteresis value, x, is called the hand-
over margin and can be chosen freely. According to
exemplary embodiments, x typically lies somewhere in
the range of 3 to 6 dB. If the difference between new
and current power levels is smaller than the hand-over
margin, no re-allocation is made, and the current link
is maintained.
z~~'a~~~
WO 95124810 PGT/US95102517
19
The above discussion considers the allocation
decision in which a call is already in progress and
during which MAIiO, or mobile assisted hand over,
measurements are sent over the SACCH (or FACCH) of the
current traffic channel. In case of a call setup, no
traffic channel has been allocated, and the MAHO
measurements are sent over the control/calling channel
of the base station to which the mobile station is
locked on. This is because even if a mobile is in
stand-by (i.e., sleep mode) it still keeps track of the
surrounding bases during brief periods of wakeup when
it performs measurements of pilot signal strength and
listens for pages. From this information, the
preferred channel and base combination can be
determined as described above, and the required
transmit power on this new channel is calculated.
However, since no current link exists, there is no
current transmit power to which a comparison can be
made; that is, there is no hand-over margin. Instead
the P~.",q on the new channel is compared with the power
control range (set by the minimum and maximum power the
mobile station can transmit) or with a set-up threshold
as being the maximum transmit power allowed at set-up.
If the P~,,e,q of the best channel is below the
predetermined maximum transmit power allowed, this
channel can be allocated right away. If not, the call
must be blocked. The set-up threshold, which may be
lower than the maximum available output power, will
prevent users that need a lot of power (and would
therefore produce a high level of interference) from
entering the system at the expense of users who would
require lower transmit powers. p
In the third operational mode (i.e., power on) no
MAHO measurement data is available at all. At power
on, the mobile station first scans through all the
WO 95/24810 PCTIUS95102517
calling/control channels of the base stations and locks
on to the strongest. The corresponding base will then
download the channel numbers of the surrounding bases
to be measured into the mobile station. From here on,
5 the procedure is as described above. A flowchart
summarizing the uplink channel allocation procedure for
all three operational modes is shown in Figure 5.
In Figure 5, superblock 500 indicates a first
general step of determining in which of three modes a
10 mobile station is currently operating. If the mobile
station is operating in the power on mode at block 501,
the flow proceeds to block 504 where the mobile station
scans the control channels to receive scan codes.
Otherwise, the mobile station is operating in the
15 standby mode 502 or the call in progress mode 503, and
the flow proceeds to the acquisition phase at 510. The
acquisition phase 510 begins with the step of having
the mobile station make RSSI measurements of pilot
signals transmitted from surrounding base stations at
20 511. Next, these RSSI measurements are sent back to
the base station to which the mobile station has an
active link at block 512. At block 513, the base
stations make their own RSSI measurements of the idle
traffic channels. The following steps relate to
determining which of the channels should be assigned
for this link with the mobile station at 520. Thus,
the f low proceeds to block 521 wherein the base station
controller calculates a path loss between the mobile
station and each of the base stations whose signals
were measured at block 511. The base station
controller then calculates the required transmit power
levels for the mobile station for each combination of
base station and traffic channel using, for example,
equation (1) and determines the combination which gives
1
wo 9sr2as
PCTIUS95/02517
21
the minimum required transmit power for the mobile
station at block 522.
Having determined which channel should be
allocated for this link with the mobile station, the
5 flow then proceeds to the handover phase indicated by
superblock 530. At decision block 531, it is
determined whether or not the mobile station currently
has a call in progress. If so, the flow moves then to
block 532 where it is then determined whether or not
10 the minimum required transmit power calculated in block
522 exceeds a current transmit power minus the handover
margin. If the answer to this determination is no,
then the flow proceeds to block 533 where the handoff
to the' determined channel/base station occurs.
Otherwise, the flow loops back to the beginning of the
acquisition phase 510 and no handover occurs at this
time.
Looking back again at decision block 531, if it is
determined, on the other hand, that there is no call in
ZO progress, then the flow proceeds to block 535 where it
is determined if the minimum required transmit power
exceeds a maximum allowable power as discussed above.
If not, the process then moves to block 534 where the
call is set up on the channel/base station combination
determined in block 522. If the minimum required
transmit power does exceed the predetermined maximum
allowable power, then the call is blocked at 536.
Assuming that the uplink and downlink channels can
be independently allocated, the procedure for assigning
the downlink is similar to that described above for
assigning the uplink. The path loss values are found
the same way as before by using the mobile station
measurements of the pilot RSSIs. In fact, these can be
made directly without the assumption of reciprocity as
was necessary with the uplink estimate. However, now
WO 95124810
PCT/US95/02517
22
interference measurements on the idle channels at the
mobile are made. This assumes that the mobile station
has the same ability as the base station to scan the
full range of voice channels. Since no particular base
is involved, the interference levels only depend on the
channel number K: I~(R). These measurements are sent
back to the base to which the mobile station is
connected, and are directed to its base station
controller. The required transmit power of base
station (J) is calculated as:
P~,Z~(J, I4 = C I) a+PL (Jj +Ina,,r,(K) dB (2 )
To reduce the number of computations, only those
channels R on base station J are considered that are
idle and free to be used. Before the actual re-
allocation is made, a comparison between the new P~ on
the best channel and the average P~ on the current
channel can be made. If smaller than the hand-over
margin, then no re-allocation should be made. The
downlink procedure according to this exemplary
embodiment will now be described with reference to
Figure 6.
In Figure 6, superblock 600 denotes the step of
determining the mobile station's current operating
mode. If the mobile station is in the power on mode
601, the flow then proceeds to block 604 where the
mobile station scans the detected control channels to
receive scan codes. Otherwise, in standby mode 602 or
call in progress mode 603, the flow proceeds to block
610 where the acquisition phase occurs. At block 611,
the mobile station makes RSSI measurements of the pilot
signals transmitted from surrounding base stations.
Next, the mobile station at block 612 makes RSSI
WO 95124810 ~ ~ PCT/US95/02517
23
measurements of the idle~traffic channels. These RSSI
measurements are sent back to the base station with
which the mobile station has an active link at block
613. The determination phase 620 begins with the base
station controller calculating a path loss between that
base station and the mobile station at block 621. The
base station controller then calculates the required
transmit power levels for the base station for each
combination of base station and traffic channel using
equation (2), above, and then determines the minimum
required transmit power for the base station at block
622.
Next, during the handover phase 630, there is an
initial determination as to whether or not the mobile
station has a call in progress at block 631. If the
result of this determination is affirmative, the flow
moves next to decision block 632 where it is determined
whether the transmit power calculated at block 622
exceeds a current transmit power minus a handover
margin. If not, then the handoff can be made to the
channel base station combination requiring the lowest
base station transmit power at block 633. Otherwise,
no handoff occurs and the flow loops back to the
acquisition phase 610. Looking back again at block
631, if there is no call in progress then it is
determined at block 635 whether or not the minimum
required transmit power exceeds a predetermined maximum
allowable power. If not, then the call can be set up
at block 634 on the channel/base station combination
which requires the lowest base station power.
Otherwise, the call will be blocked at 636 and the flow
proceeds back to the acquisition phase 610.
The present invention has been discussed in terms
of an exemplary embodiment wherein uplink and downlink
channels can be selected independently, and wherein
WO 95/24810 s PGT/US95/02517
~~.~N~J°J
24
each channel is provided with its own-APC. Another
exemplary embodiment of the present invention is now
presented wherein the uplink and downlink channels are
paired in such a manner that there is a f fixed of f set
(e. g., 45 MHz in the IS-54B standard.) In this case,
the uplink and downlink information is combined.
Instead of selecting the best uplink or downlink
channel, one selects the best channel pair which
minimizes the transmit power in both the uplink and in
the downlink. Thus, a method to minimize the weighted
sum of the P~,,~ and the P~ is given by:
. min (b ( Pte, rte) + ( 1-b) ( Pte, r.Q) }
The parameter b can be chosen with respect to the
system conditions and the desires of the operator. For
example, if attention is placed on maximizing the
mobile station battery life (i.e., minimizing the
transmit power of the mobile), greater weight should be
placed on minimizing P~,~,,q, so b should approach one.
However, if for example, capacity is limited by the
downlink (~e.g., when the base applies receive diversity
with two receive antennas or has other means to reduce
its C/I while maintaining satisfactory link quality),
most weight should be placed on P~,~, so b should
approach zero. It should be noted that if the
interference situations in the uplink and downlink are
highly correlated, then minimizing the sum comes close
to minimizing either P~,,,$ or P~ individually. In all
other respects, the path loss, PL, and interference
measurements, Ice, I~,~,I," which affect P~ or P~ are
performed as discussed for the earlier described
exemplary embodiment.
WO 95124810 d'' ~ '~ PCT1LTS95102517
In the above, it was assumed that the mobile
station had the same capacity as the base station for
scanning the range of voice channels. To scan, the
mobile locks on a frequency, settles, and then performs
5 a measurement. Clearly for a cellular systems with
over 1000 traffic channels, such as the system defined
by the IS-54B specification, it is impractical for the
mobile to monitor every channel. In addition to the
large scanning load in the mobile station, TDMA
10 protocols further exacerbate the scanning problem in
that it is impossible to determine the downlink
interference on each individual slot. This is a result
of the continuous transmission from the base station.
As is well-known, a TDMA system divides the 30 kHz wide
15 channels used in AIDS into time slots. IS-548, for
example, divides the channel into three time slots
resulting in a three-fold increase in traffic capacity.
Far reasons beyond the scope of this disclosure, the
base station transmits on all three downlink slots even
20 if only one slot is active. Therefore, even if only
one time slot is being used, all three time slots on
the same carrier carry the same transmit power (with
filler information on the idle slots) . An illustration
of the IS-54B time slot structure is shown in Figure 7.
25 Note that, although six slots are shown, the current
system allocates to each channel two slots per frame
(i.e., TSO = TS3, TSl = TS4, TS2 = TS5). In order to
more clearly describe the function of the invention in
a TDMA system, a pedagogic description of the problem
of measuring downlink interference in a TDMA system is
provided.
With regard to Figure 7, assume that a base
station uses time slot TSO on carrier 1 to communicate
with user A. The slots TS1 and TS2 on carrier 1 are
assumed to be idle. Now user B is also close to this
:~~ ~~~~~
WO 95!24810 ~ ~ ~ ~'' PCT/US95/02517
26
base and is interested in startup, or hand-over, of a
call on TS1 or TS2 on carrier 1, which, as stated, are
both available. The mobile station would therefore
like to determine the downlink interference on TS1. In
our notation, the downlink interference on base station
J, channel R, and time slot, TSx, would be written as
I(J,R,TSx). Therefore the measurement made by the
mobile station of the downlink interference emanating
from base station A, on channel 1, time slot 1 would be
I~,,,(A,1,TS1) according to this notation. Since base
station A does not shutdown its power after TSO, but
continues to transmit (with the same power as used in
TSO) on TS1 and TS2 as well, user B measures a large
RSSI (which would actually be its carrier strength when
selecting TSl). This large signal completely
overwhelms the interference signal which is likely much
weaker than the carrier. In this case a measurement of
the downlink interference strength on TS1 is
impossible. (Note that the idle power transmitted in
TS1 would only be an interference to user B if it would
select TS1 on carrier 1 on a different base.) Observe
that, in the previous exemplary embodiment, downlink
interference measurements were made on idle channels.
That is, the mobile station would only measure the RSSI
of downlink channels that were currently not being used
by the base station to which a link was being
considered.
Downlink interference measurements could be
avoided altogether if one can rely on a high
correlation between uplink and downlink interference
conditions. In this case, by choosing the best uplink
channel using the method summarized by equation (1), .
the corresponding downlink channel (offset by 45 MHz)
will be acceptable as well. Uplink interference
measurements using TDMA are possible because, unlike
i
WO 751iY010 ~' Mr
PCTlUS95102517
27
the base stations,. the mobiles transmit on only, for
example in IS-54B, one slot of three. Measurements
made on unoccupied slots would therefore only include
interference. If the correlation is poor, as it most
often would be, one must make a prediction of the
downlink interference. Therefore, according to another
exemplary embodiment of the present invention, a method
of predicting the downlink interference levels is
included to deal with situations where direct
measurement thereof is impossible.
To predict the downlink interference, one should
identify the base stations in a wide range around the
mobile station that are transmitting on the same
channel. The location of these bases and the power
levels they are currently using on the considered
channel, must be directed to the base station
controller doing the ACAPC processing (e.g., in the
example of Figure 3, this is base station controller
320.) This can be accomplished, for example, directly
through the MTSO 300. Adjacent channel interference
can also be included if it is corrected with the
adjacent channel rejection factor.
Adjacent channel interference results because of
the non-ideal filtering operations in the transmitter
and receiver. The power spectrum of a transmitted
signal is not zero in adjacent bands, but instead falls
off as a function of the frequency offset, see Figure
8(a). On the other hand, the receive filter
characteristic is not rectangular, and some power
outside of the receive band is taken in as well, see
Figure 8(b). The total adjacent channel interference
that user A introduces to user B can be seen in Figure
8(c) and is.determined by, for example:
(a) the shape of the power spectrum transmitted
by user A, e.g., the curve in Figure 8(a);
w _~ ' ~~.6~~~
WO 95/24810 PCT/US95I02517
28
(b) the absolute value of. the power transmitted
by user A, e.g., the area under curve of
Figure 8(a);
(c) the filter characteristic of user B's
receiver, e. g. , the curve in Figure 8 (b) ; and
(d) the frequency offset between user and
interf erer .
Items (a) and (c) can be derived from the system
specifications (or from the specification of the
transmitter and receiver equipment). Item (b) is the
transmit power used by A which information can be
passed on to the system and item (d) is known as well.
The adjacent channel rejection factor can be determined
using items (a), (c) and (d). Together with the
absolute transmit power of the adjacent interferer, the
adjacent channel interference can then be determined.
If the spectral shape indicated in item (a) and the
filter characteristic, item (c), are not exactly known,
at least there exist system specifications which give
worst case conditions for allowable spectral shape and
filter response. Using this information, the worst-
case adjacent channel interference can be calculated.
Unlike the previous situation where path loss
could be calculated from pilot signal data, the path
loss from each of these bases to the mobile station is
predicted. Because of the frequency plan adopted for
the control channels, only the nearest (i.e., the
closest 21) base stations can be measured. For these
nearby bases which are included in the measurement
list, the path losses can accurately be determined, as
previously discussed from information contained within
the base station and measurements made at the mobile
station.
The interfering signals are however more likely to
come from distantly located base stations. The control
wo 9snasio rcrrt~s9sio2.sm
29
channels from these base stations oannot~be measured
directly since their signals are overpowered by those
base stations in close proximity which have already
been measured. Therefore only a crude prediction of
the interference emanating from these base stations can
be made using known propagation loss models and/or
geographical data of the local surroundings which can
be located in a database connected to the base station
controller. An example of known propagation loss
models can be found, for example, in Y. Okumura,
et al., "Field Strength and it Variability in UHF and
VHF Land-Mobile Radio Service", Review of Elec. Comm.
Lab., ,Vol. 16, Sept.-Oct. 1968, pp. 825-873 and
M. Hata, "Empirical Formula for Propagation Loss in
Land Mobile Radio Services", IEEE Trans. on Veh. Tech. ,
vol. VT-29, no. 3, Aug. 1980, pp. 317-325. These data
can be obtained during installation using transportable
test transmitters at proposed base locations, and the
use of a specially equipped vehicle to measure RSSI in
a local area. Once the transmit powers and the path
losses are known, the signal received in the mobile
from each base can be calculated. By adding all these
power levels together, a prediction of the interference
level is obtained. Alternatively, the path losses
between all of the base stations can be measured when
the system is installed and stored within the network
database. These data can be periodically updated as
characteristics of the terrain evolve.
An example of interference prediction will now be
described with respect to Figure 9. Therein mobile
station A is close to base station A and wants to
initiate a call over channel K on base station A. The
network scans the region around base station A for
other base stations that use channel R. It finds three
bases L, M, and N. In most cases, these base stations
' _..
WO 95124810 PCT/US95/02517
are not normally scanned _ by a mobile station located
near base station A since they are remotely located.
(The mobile usually scans closely located base stations
such as B through G inclusive). It is known that base
5 stations L, M, and~N use a transmit power on channel K
of 0 dBm, 10 dBm and 15 dBm, respectively. In our
notation this is : P.~ ( L, K) = 0 dBm, P.~ (M, K) = 10 dBm, and
P.~(N,K)= 15 dBm.
These numbers are sent to the base station
10 controller of base A, together with the location of
bases L, M, and N. The network knows, a priori, the
distances between base station A and the L, M, and N
base stations. In addition, the network may have extra
information about the terrain between base station A
15 and the other base stations (e. g., mountains, hills,
high buildings). From this information the path loss
between base station A and the interfering bases can be
determined. Although the location of the mobile
station is not exactly known (unless the system
20 contains a locating feature which can determine the
location of the mobile within the cellj, if the
distance is large it can safely be assumed that the
path losses from the interfering base stations to the
mobile station does not deviate too much from the path
25 losses to base station A.
Alternatively, a so-called beacon may be employed
from which these path losses may be directly obtained.
From the total number of voice channels available, one
channel may be selected for use as a radio beacon.
30 That is, a particular frequency can be chosen to be
transmitted sequentially from several distant base
stations. To obtain the path losses between any given
mobile station and a distant base station the mobile
station can be instructed to scan a beacon frequency.
Each base station then, in turn, can broadcast
WO 95124810 ~ ~ PCT/US95I02517
31
information identifying the particular base station and
its power level on this frequiency. The mobile station
can measure the RSSI of this signal and demodulate the
identification information according to well-known
methods. Determination of the path loss can then be
obtained as previously described. Indeed, the mobile
station can simply report raw RSSI data back to the
base station controller. Since only one base station
would broadcast the beacon at a time, the raw RSSI data
can be correlated to an individual base station by the
network.
Assume that he path losses (predicted or
measured) from base stations L, M, and N to the mobile
station are 160 d8, 155 dB and 170 dB, respectively.
Again using our notation to express the path loss
between the mobile station and base station J, we have:
PL(L)= 160 dB, PL(M)= 155 dB, and PL(N)= 170 dB. By
using the known transmit powers of base stations L, M,
and N (i.e., P.~(L,R)=0 dBm, P~(M,R)=10 dBm, and
P~(N,R)=15 dBm) and the path losses, the predicted
RSSIs received in the mobile station are calculated as:
I(J, R) - P .~ (J, R) - PL (J)
Using this equation with the values selected for this
example, the interference from each base station is:
I(L,K) - OdBm - 160dB = -160 dBm
Ice,, (M,R) - lOdBm - 155dB = -145 dBm
I~~,, (N,R) - lSdBm - 170dB = -155 dBm
The total predicted interference Im~,,(R) on channel K
results from the addition of these values which amounts
to approximately -144.5 dBm. Note that the powers must
be added since the voltages are uncorrelated. The a
predicted I~~(K) can then be used as described before.
In this manner the interferences can be predicted
whenever the mobile station, whether due to the
WO 95J24810 PCT/US95J02517
32
signaling format or limited scanning capability,. is
unable to measure these values directly.
In all of the embodiments presented so far, it was
assumed that the APC controlled the power level on each
channel individually. Not all systems provide this
capability. For example in the IS-54B system, this is
only true in the uplink: the mobile station only
transmits in its own slot on its own carrier, and can
freely adjust its power level. This is not so in the
downlink. Because of the absence of guard times in the
signaling format, power cannot be ramped up or down
between slots. Therefore, all slots sharing the same
carrier all use the same power level in the downlink.
As we have seen before, this is even true if a slot is
idle. Only if all three time slots are idle can the
base shutoff the transmit power on that channel. It is
also clear that to avoid wasting power (and to reduce
overall interference), idle slots on active carriers
should be avoided. This means that the allocation
scheme should attempt to maximize the number of calls
on as few channels as possible. This so-called time
slot packing is desirable in IS-54B and can be part of
the allocation process wherein priority can first be
given to placing calls on those active carriers with
idle slots.
An exemplary embodiment of the present invention
for a system with downlink transmit power restrictions
as described above is now discussed. To distinguish
the carrier from its constituent slots (a channel is
made up of a carrier frequency and a time slot) we will
represent the carrier number by F and the slot number
by TS. The APC in the downlink for this exemplary
embodiment will be assumed to be limited to variations
in carrier power only. If it is assumed that the power
is controlled with respect to the user with the lowest
_. ~~6~~~
WO 95/24810 PCTlUS95/02517
33
C/I, all other users using the same carrier will then
.' have excess quality in the downlink (i.e., a better C/I
than is necessary). To minimize transmit power, the
allocation process should also organize slots such that
those mobile stations needing a similar p~~ will be
placed on the same carrier.
The acquisition phase is the same as before. The
path loss values are derived from measurements in the
mobile 'station, and the downlink interference I~~,,(F)
is either measured or predicted. Note that in the
downlink, the interference levels on all slots sharing
a carrier F are the same. In the previously developed
notation:
I~~,I,,(F,TSl)=I~~(F,TS2)=I~~(F,TS3)=Im~(F)
Then the required base transmit power P~~ (J, F) on
base J, carrier F can be calculated using equation (2)
repeated below with the variable R replaced by F.
Pas. zeQ t aT, F') _ ( I ) a +PL ( a» +I~ ( F) dB ( 3 )
Only those bases, J, are taken into account that are in
the measurement list, and only those carriers, F, are
taken into account that have at least one idle time
slot.
The process for the downlink considers first the
carriers which are already active, but have at least
one idle slot. These carriers are currently
transmitting at a transmit power P~ that is controlled
by the user with the lowest C/I. In accordance with
this exemplary embodiment of the invention, all
carriers which have all slots idle are inactive and are
considered at the same time as active carriers with
open slots. The carrier which is currently used should
also be included; it is marked inactive if only one
WO 95124810
PCT/US95/02517
34
slot is ~ . .occupied ( i . e. , ' by the user under
consideration), or otherwise marked active (then X
other users are sharing this carrier) with X slots
occupied not including the user under consideration.
The ACAPC process subsequently calculates the
dif f erence dP between the required power P~.,~ and the
actual transmit power P~ on the active carriers:
dP ( J, F') =Pte, z~ ( J, F') -P~ ( J, F~
From this information three ordered lists are
generated: one for channels with one slot open, one for
channels with two slots open and one for inactive
carriers. For each list the process sorts the dP
values from strongest to weakest (most positive to most
negative). Subsequently the process scans the dP
values of each list separately starting with the
strongest (most positive) first, until dP becomes
negative ( i . a . , dP<0 ) f or the f first time or the last
element in the list is reached (when all dP > 0.) The
P~,"~ for each of the three lists corresponding to the
first negative dP value, or to the smallest positive dP
value are thereby obtained. The corresponding carriers
are optimal in the sense that their current transmit
powers are just sufficient (largest dP < 0) or require
the least amount of increase (smallest dP > 0) to
satisfy the C/I requirements of the user under
consideration. The P~~ values of the three lists are
identif led as P~,,~ ( 2 ) , P~.,~ ( 1 ) , P~,,~ ( 0 ) f or channels
with one slot open, two slots open, or all three slots
open respectively. Notice that the nomenclature refers
to the number of occupied, rather than open, slots.
These three values are then compared with regard to the
absolute power level and occupancy.
WO 95/24810 PCT/US95/02517
Although, it is generally preferred to allocate a
carrier with as many occupied slots as possible, this
will not occur if it results in transmitted power
levels being near maximum. To facilitate the decision,
5 hysteresis values are established. Specifically,
Hys2,1 is the required difference, in dB, between
choosing to allocate the mobile station to a carrier
with one slot occupied and a carrier with two slots
occupied. Hysl,O is then, similarly, the required
10 difference, also in dB, between choosing to allocate
the mobile station to a carrier with one slot occupied
and an inactive carrier (i.e., zero slots occupied) and
Hys2,0 ,is the required difference in dB between
choosing to allocate the mobile station to a carrier
15 with two slots occupied and an inactive carrier. The
carrier with one slot open is chosen if both of the
following inequalities, with Hys2,1 less than Hys2,0
are true:
P~.r~(2) <P~.r,Q(1) +Hys2,1
and
PBS, rsQ ~ 2 ) <P~. raQ ~ 0 ) +Hys2, 0
Otherwise, the carrier with two open slots is selected
20 provided that:
P~.r~(1) <P~,r~(0) +HySl, 0
If the selection criteria has not yet been satisfied,
then the inactive carrier with the lowest required
transmit power is selected.
As may be gleaned from the above, the greater the
25 hysteresis values, the greater the slot packing. It
can also be observed that there is a trade-off between
maximizing slot-packing and taking the channels with
the lowest required transmit power. According to an
exemplary embodiment, Hys2,1 is equal to 3 dB, Hys2,0
30 is equal to 9 dB, and Hys 1, 0 is equal to 6 dB. In
w
WO 95124810 ~ ~ ~ ~ PCTIUS95/OZ517
36
other words, to activate an inactive carrier, the
required power level must be at~least 6 dB lower than
that required on an active carrier with two slots idle
and at least 9 dB lower than that required on an active
carrier with 1 slot idle. Similarly, an active carrier
with one slot idle is preferred over an active carrier
with two slots idle if the former does not require a
power level which is more than 3 dB of that required on
the carrier with two slots idle. Note that these
values are offered for the purposes of illustration and
are not meant to be limitative.
In the downlink, for this exemplary system, the
C/I ratios on all slots sharing a common carrier are
the same, thus, it is,~~immaterial which slot should be
selected if more than one is idle. Therefore a random
selection could be made, or the selection can be
determined by the uplink characteristics as described
next.
Again the base station which is selected is not
necessarily that with the lowest path loss (i.e.,
ostensibly closest) since another base station with a
higher path loss (i.e., ostensibly, further away) may
offer better slot packing possibilities. In the
description above, for the purposes of illustration,
the number of time slots in the TDMA frame was chosen
to be three. Those skilled in the art will observe
that the present invention may be easily extended to
other TDMA systems using more than three (or less than
three) slots. For example, there will be an equivalent
number of ordered lists and time slots.
An exemplary embodiment of the ACAPC process for
the uplink is as follows. Since the mobile station
only transmits in its own time slot and shuts down the
transmission in the other slots, the difficulties in
determining interference found in the downlink are not
wo 95124810 ~ ~ ~ ~ ~ ~ ~ PCT/US95l02517
37
experienced in the uplink. In addition, the mobile
station itself determines the APC on the ~.nplink:
Therefore, the uplink C/I is usually better than that
of the downlink, and the performance of the system is
limited by the downlink. When the uplink and the
downlink can be selected independently, the best uplink
can be found by using equation (i) repeated below with
I~(J,R) replaced by I~(J,F,TS) and selecting that
channel base/carrier/slot combination that gives the
lowest PI"g,~ (J, F, TS ) .
Pxs, zeQ ( J, F, TS) _ ( I ) a +PL ( J) +Ipp ( J, F, TS) dB ( 4 )
When the uplink and downlink channels are paired (e. g.,
with 45 I~iz offset), then the uplink should accept the
base/carrier combination (J, R) found by the downlink
process as described before (since this is the best
combination for the downlink which is limiting the
performance). If more than one idle time slot is
present on this base/carrier combination, the upiink
can now be optimized by selecting that time slot that
gives the best uplink performance with the lowest
P~.,~(J,F,TS). Exemplary embodiments of a mobile unit
and a base station in which the foregoing exemplary
channel allocation schemes can be implemented will now
be described in conjunction with Figures 10 and 11,
respectively.
In Figure 10, the mobile station 900 has an
antenna 902. A transmitter 904 is connected to the
antenna 902 and is controlled by a transmitter control
unit 906 which, among other functions, is able to
effect channel allocation in conjunction with control
logic 916. The transmitter is also connected to a
signal processing unit 908. A receiver 910 is also
W0 95/24810 ~ PCT/US95/02517
38
connected to the antenna and is used in time multiplex .
together with the transmitter 904. The receiver.910 is
also connected to the signal processing unit 908.
Radio equipment for modulating, demodulating and
equalizing purposes is included in the blocks 904 and
910. The signal processing unit 908 includes, for
example, circuitry for channel coding, channel decoding
and signal processing of incoming and outgoing speech.
The signal processing unit 908 is also connected to a
microphone and speaker in block 914, and to control
logic 916. In turn, the control logic 916 is connected
to the transmitter control unit 906 and to I/O-block
918 which processes the I/O signals from a keyboard
(not shown) and to a display 919.
Figure 11 is a block diagram illustrating an
exemplary base station. Although the block diagram of
Figure 11, is illustrated as a single system, those
skilled in the art will readily appreciate that the
hardware shown in Figure 11 can also be distributed
over several units, for instance over a base station
and a base station controller.
The base station, generally referred to by
reference numeral 1000, has three antennas, of which
two, 1002 and 1004, are used for receiving signals,
whereas only one antenna 1006, is used to transmit
signals. A transmitter 1008 is connected to the
antenna 1006 and is controlled by a transmitter control
unit 1010. The transmitter 1008 is also connected to
the signal processing unit 1012. A receiver 1014 is
also connected to the antennas 1002 and 1004 and the
signal processing unit 1012. Radio equipment for
modulating and demodulating and equalizing purposes is
included in the blocks 1008 and 1014. The signal
processor unit 1012 provides for channel coding and
decoding and processing speech in the incoming and
CA 02162256 2003-12-30
WO 95!24810 PCT/US95/OZ517
39
outgoing directions. The signal processor unit 1012 is
also connected to the PCM-link adaptor block 1016 and
to the control logic 1018. In turn, the control logic
IOI8 is connected to the transmitter control unit 1010.
Figure 12 is a block diagram which illustrates a
portion of an exemplary base station controller which
handles the ACAPC routine according to the present
invention. The CPU 1100 will receive the measurement
data, perform averaging, make ordered lists and perform
the other decisions described above regarding the
allocation of channels. The memory 1101, in addition
to storing current channel allocation assignments, can
contain a database including information regarding the
geographical characteristics of the surrounding area
which can be used to make predictions. Alternatively,
such a database could be located at the MTSO. The I/O
unit 1102 can connect this base station controllers to
other BSCs, the MTSO and the base station that it
controls.
Those skilled in the art will appreciate that the
foregoing exemplary ,mobile and base station
descriptions are intended simply to illustrate
apparatuses which can be used to implement the channel
allocation schemes according to the present invention
and that any other type of base station or mobile
station can be used. For example, those systems
disclosed in U.S. Patents Nos. 5,230,082 entitled
"Method and Apparatus for Enhancing Signaling
Reliability in a Cellular Mobile Radio Telephone
System" to Ghisler et al. and 5,109,528 entitled
"Handover Method for a Mobile Radio System" to
Uddenfeldt can also be used.
wo 9snamo ~ ~ ~ ~ ~ ~ ~ Pcr~s~o2sm
The. above-described exemplary embodiments are
intended to be illustrativew in all respects, rather
than restrictive, of the present invention. Thus, the
present invention is capable of many variations in
5 detailed implementation that can be derived from the
description contained herein by a person skilled in the
art. All such variations and modifications are
considered to be within the scope and spirit of the
present invention as defined by the following claims.
