Note: Descriptions are shown in the official language in which they were submitted.
CA 02340644 2001-02-13
WO 00/10258 - 1 - PCT/EP99/05619
Method and device for a full-duplex-ca able radio trans-
mission system with CDMA access
The invention relates to a method and a device
S for a full-duplex-capable radio transmission system with
CDMA access, having a central radio base station and a
plurality of subscriber stations which are independent
of one another.
In the field of radio-supported information
systems which operate with a central radio base station
and a plurality of external stations or subscriber
stations which are independent of one another and which
permit information to be transmitted in full duplex form
in both directions, the information in the downlink
which is intended for the individual users is frequently
multiplexed into a telecommunications channel and
Transmitted organized as an access system in the uplink.
Examples of such systems are mobile radio systems,
public trunked mobile radio systems, point-to-multipoint
microwave radio systems and wireless local loop systems.
Orthogonal signal domains which differ from one another
are used in each case for the multiplexing or multiple
access, these signal domains being, for example,
- frequency division multiplex or access systems FDMA
(frequency division multiple access)
- time division multiplex or access systems TDMA (time
division multiple access)
- code division multiplex or access systems CDMA (code
division multiple access) or SSMA (spread spectrum
multiple access)
- space-division multiplex or access systems.
The systems differ in that the transmission of
information from and to the individual users takes place
in separate frequency, time, code or spatial segment
3S positions. Interleaved, coupled or respectively differ-
ent multiplex and access technologies within one system,
so-called hybrid methods, have become known. Depending
on the use and implementation, different transmission
CA 02340644 2001-02-13
- 2 -
parameters and transmission quality criteria can be
obtained with these methods.
In CDMA systems, the user signal is coded by
gating it with a spread function using logic operations,
a separate spread function which is orthogonal to the
other spread functions being selected for each
subscriber station. The logic operation is carried out
here in each case by means of an X-OR gate, for example.
At the receiver end, the coded signal can be demodulated
~0 through knowledge of the associated spread function, the
ceded user data for other subscriber stations becoming
zero during the demodulation process owing to the
orthogonality. It is particularly advantageous with CDMA
systems that all the users can operate in the same
I5 frequency band and a relatively high degree of
interference power in the band can be tolerated.
Furthermore, under certain conditions it is possible
that adjacent radio cells can operate on the same
frectuency band. It is generally a disadvantage that the
20 mufti-user interference, which arises in practice as a
result of implementation problems such as band
limitation, level differences between the individual
tra.~.smissions, mufti-path propagation etc. and which
leads to a loss of orthogonality. In the radio systems
25 under consideration, it is to be noted basically that
because of the different signal transmit times owing to
different distances between the external stations and
the central station an asynchronous reception situation
is normally produced in the base station receiver, which
30 situation considerably aggravates this interference to
such an extent that under ideal conditions code orthogo-
nality is then no longer produced in the uplink. In this
case, the maximum number of simultaneous transmissions M
within a frequency band in the uplink of, by
35 approximation, a DS-CDMA system can be estimated as
follows:
M = PG/ (Eb/No) ,
CA 02340644 2001-02-13
- 3 -
PG being the process gain or spread factor and ED/No
being the ratio of bit energy to interference power,
necessary for t:-~e aimed-at bit error rate, at the
demodulator. The spread factor is the ratio of tbit to
t~nip and is typicall y between 101 and 10~.
Assuming the ratio Eb/No is, for example, 3,
which corresponds to approximately 5 dB, only
approximately 1/3 of the transmission capacity, based on
the same bandwidth being seized, is available in the
uplink in comparison with the downlink or in comparison
with TDMA or FDMA systems if orthogonal signals are
assumed for the latter.
Various methods are known which reduce the
previously described disadvantage of the asynchronous
CDMA methods, for example the synchronization of the
external station in such a way that its transmission can
be processed in synchronism with the chip in the
receiver of the base station. In addition, it has been
proposed to implement interference concelers which, by
means of mathematical algorithms, subsequently eliminate
the interference component of the parallel transmissions
on the basis of different a priori or a posteriors know-
ledge. Furthermore, it has also been proposed to use
mufti-user detectors. A disadvantage of all these known
methods is that they are very costly to implement.
The invention is therefore based on the
technical problem of providing a method and a device for
synchronization in a radio transmission system with CDMA
access, by means of which method and device the multi-
user interference in the radio base station in the
uplink mode can be reduced with low cost in terms of
implementation.
The technical problem is solved by means of
the features of patent claims 1 and 9. It is necessary
for the radio transmission system to be operating in
time division duplex mode in which transmission and
reception are separated from one another in terms of
time within one telecommunications channel, which
significantly simplifies the sequence control. In order
CA 02340644 2001-02-13
- 4 -
to synchronize all the subscriber stations, the radio
base station transmits a maximum sequence or gold
sequence, specific to the radio transmission system, in
the form of a preamble for all the subscriber stations
before the actual data transmission. Since the
information on the direct subscriber-specific system
control, such as, for example, call setup and the like
is transmitted in a central service channel, a common
preamble can be used for all the subscriber stations.
This preamble can be detected without restricting other
system parameters with a significantly better
signal/noise ratio, since mufti-user interference is not
present and the subscriber-specific signal powers can be
transmitted in an additive, coherent fashion, which
brings about a high level of detection reliability in
Lhe subscriber stations. The preamble which is received
there is fed to a matched or correlation filter whose
output signal serves as a trigger criterion when a
defined amplitude threshold value is exceeded. Further
advantageous refinements of the invention emerge from
the subclaims.
The averaging over time of the synchronous
information which is determined, and the evaluation of
the kncwledge of the precise value between two
successive preambles, makes it possible to achieve
substantially greater precision, given sufficient clock
stability in the subscriber stations, since an
uncertainty in terms of timing of up to 0.5 x chip
duration t~nip can occur with simple detection using a
matched filter.
The transmission of the synchronous
information in the uplink parallel to the transmission
of user data is made more difficult by the fact that the
synchronous information is a priori not known, or not
known sufficiently precisely, as a result of which its
acquisition would lead to an asynchronous interference
situation with respect to the actual user data
transmission. In order to avoid this, in each case only
one item of synchronous information for all the
CA 02340644 2001-02-13
subscriber stations wh-~ch are active in parallel is
transmitted simultaneously in the delay time between the
transmission cycle and reception cycle, as a result of
which the time information can be detected more reliably
because the transmission is subject to significantly
less interference. For this purpose, if appropriate the
delay time must be extended somewhat, but this is
acceptable in order to achieve improved detection.
As a result of the transmission-end shifting of the
symbols by one sample value in each case, but [lacuna]
symbol-based matched filtering, with fixed timing, in
the receiver of the base station, the time resolution or
precision of the synchronization information within only
one burst is improved up to a sample val ue tS~,Dler which,
in the case of conventional detection, can be up to
0.5 x chip duration trip.
In order to avoid data collisions, the radio
base station transmits to the subscriber station via the
central service channel a status signal specifying which
2C subscriber stations are to transmit their
synchronization sequence consecutively. After the
evaluation of the signal transmit time by the radio base
station, said station transmits via the service channel
the subscriber-specific starting times for the uplink
transmission.
In a further preferred refinement of the
method, orthogonal gold sequences of the length of one
symbol in each case are used for the code spreading of
the data both in the uplink and in the downlink, said
sequences being relatively easy to generate.
Furthermore, the orthogonal gold sequences have defined
cross-correlation properties, the result of which is the
subscriber stations, in which the synchronization
mechanism fails, do not cause any significant faults in
the other parallel transmissions. Furthermore, in
comparison with Walsh sequences and similar sequences,
these have the advantage of a uniform spectral power
distribution, which is significant in particular in the
case of short sequences.
CA 02340644 2001-02-13
- 6 -
It is advantageous for the design of cellular
structures if all the radio base stations which lie in
the range of mutual radio influence are synchronized in
terms of the transmission/reception cycle. In particular
radio base stations or subscriber stations which are in
line-of-sight with respect to one another owing to an
exposed geographical position could otherwise give rise
to considerable interference at the receivers of the
respective other radio cells. The synchronization may be
carried out, for example, by means of GPS or beacon
signals which are passed on within the radio network.
The method can be particularly advantageously
implemented in wireless local loop systems, since in
these systems the stationary nature of the subscriber
stations with relatively small changes of the properties
of the radio channel over time can be exploited.
The invention is explained in more detail
below with reference to a preferred exemplary
embodiment.
Fig. 1 shows a signal profile of a transmission
in the downlink for a subscriber station,
Fig. 2 shows a signal profile of a transmission
in the downlink for n subscriber
stations,
Fig. 3 shows a schematic signal profile at the
output of a matched filter in a sub-
scriber station,
Fig. 4 shows an illustration of the polling
method for the synchronization in the
uplink,
Fig. 5 shows a structure of an uplink synchron-
ization sequence,
Fig. 6 shows a detailed illustration of the
structure according to Fig. 5, and
Fig. 7 shows a schematic signal profile at the
output of a matched filter of a radio
base station.
Fig. 1 illustrates a schematic signal profile of
a transmission of a radio base station to a subscriber
CA 02340644 2001-02-13
7 -
station over time t. The signal comprises a preamble 1
and a data item 2 which are transmitted with an
amplitude PT. The preamble 1 is in this case a radio-
system-specific maximum sequence or gold sequence which
is generated by the radio base station. The data item 2
constitutes the actual user data for the subscriber
station. Since the information is transmitted for the
purpose of direct subscriber-specific system control in
a central service channel, a common preamble 1 can be
used for all the subscriber stations.
Fig. 2 illustrates the signal profile of the
transmission in the downlink for all n subscriber
stations. Since the radio base station transmits
simultaneously to all n subscriber stations, a
corresponding superimposition of the signal profiles
occurs. Owing to the transmission of a common preamble 1
for all the subscriber s to Lions into a service channel,
a coherent addition occurs and the amplitude is
P, ~ n'P.:. The superimposition of the user data takes
place i:~ accordance with the code modulation which is
used, and varies correspondingly in amplitude, on
average approximately the following is true P2 ~ nPT.
In order to determine a first item of
synchronization information, the preamble 1 which is
received by each subscriber station is fed a matched
filter [sic] by means of which the reception quality can
be determined. A typical signal profile at the output of
the matched filter of a subscriber station is
illustrated in Fig. 3. In order to determine the
reception time of the transmission from the radio base
station to the respective subscriber station, the output
signal at the matched filter is evaluated by means of an
amplitude threshold value switch. If the output signal
exceeds a predefinable threshold Trl, the amplitude
threshold value switch produces a trigger signal, that
[sic] represents the starting time for the reception of
the preamble.
Fig. 4 illustrates the signal profiles for the
synchronization in the uplink. In order to avoid inter-
CA 02340644 2001-02-13
ference, the transmission of synchronization sequences 3
by the individual subscriber stations takes place here
in the form of a polling method, i.e, in the first burst
only the first subscriber station transmits its
synchronization sequence 3 to the radio base station.
Subsequently, all n subscriber stations then transmit
their user data 4 simultaneously to the radio base
station. In the second burst, only the second subscriber
station then transmits its synchronization sequence 3,
until finally in the n-th burst the n-th subscriber
station transmits its synchronization sequence 3.
A more precise structure of the synchronization
sequence 3 is illustrated in Fig. 5. The synchronization
sequence 3 comprises, for example, four identical
symbols 5 which are transmitted successively, the
distance between the symbols 5 being increased
successively by one clock pulse ts~,ple of the system
clock, and the first symbol 5 serving as preamble.
An exemplary profile of a symbol 5 is
illustrated in Fig. 6, and it corresponds to the second
symbol 5 with the transition to the third symbol 5
according to Fig. 5.
Figure 7 illustrates an exemplary signal profile
at the output of a matched filter in the radio base
station when a synchronization sequence 3 according to
Fig. 5 is received. Here, each of the four symbols 5
produces an output signal with an amplitude P which is
larger than a predefined threshold value Tr2 of a down-
stream amplitude threshold value switch. The first
symbol 5 produces an output signal with the amplitude
Pb. The second symbol 5 which is transmitted directly
after the first symbol 5 also produces an amplitude Pb.
The third symbol 5 which is delayed by a system clock
pulse ts,~Ple produces an amplitude Pa, and the symbol 5
which is correspondingly delayed by 2 X ts~ple produces
an amplitude Pc. The optimum reception is therefore that
of the third symbol 5 so that the signal transit time
has to be corrected correspondingly by one system clock
pulse ts~ple. In this way, the transit time between a
CA 02340644 2001-02-13
- 9 -
subscriber station and the radio base station car. be
determined with corresponding precision so that the
synchror_ization can also be performed in the order of
magnitude of ts~,ple.
CA 02340644 2001-02-13
1~ -
List of reference numerals
1) Preamble
2) Data item
3) Synchronization sequence
4) User data
5) Symbol
