DESCRIPTION
A MULTIUSER RECEIVER DEVICE WITH A MINIMUM BIT ERROR RATE IN THE TIME
VARYING AND FREQUENCY SELECTIVE FADING DS-CDMA CHANNELS THEREOF
Technical Field
5
The present invention relates to a multiuser receiver device minimizing the bit error rate
(BER) in the time varying and frequency selective fading channels for the base stations used
in the Direct Sequence Code Division Multiple Access (DS-CDMA) or GSM systems.
Prior Art
In the CDMA systems, different users communicate using the same frequency band
10
simultaneously. In the CDMA systems, a special spreading code is assigned to each user,
simultaneous utilization of the whole bandwidth by all users is provided. DS-CDMA is the
most widely used type of CDMA.
The traditional multiuser receiver structures being the most basic part of the DS-CDMA
systems, used in the receiver or transceiver chip sets of the base stations have a linear
15
structure, wherein said traditional multiuser receiver structures are intended for destroying
the interaction among the users. The two most classical multiuser receiver structures,
'Decorrelating' (DEC) and the 'Linear Minimum Mean Square Error' (LMMSE) receivers,
respectively, minimize the multiple access interference and mean square error. However,
since these types of linear receiver structures are not designed for minimizing the BER value,
20
the BER performances of the base stations with said receiver structures decrease with the
increasing number of users and problems are encountered in the communication.
The best known multiuser receiver structure is the 'Maximum Likelihood' (ML) receiver
structure. This receiver structure has a very complex structure, wherein the high
mathematical process requirement thereof prevents said structure from being used in the
25
real time applications in today's technology. Therefore, multiuser receiver structures with a
BER performance at least as good as the ML receivers and a less complex structure are
needed.
Similar studies carried out previously are realized only for the channel models provided with
either the 'Additive White Gaussian Noise' (AWGN) or frequency selective fading. In the
30
present invention, a time varying frequency selective fading channel model being more
appropriate to the real world and representing the moving users (mobile terminals) is
created, an algorithm minimizing the BER value in said channel model is generated and said
1
algorithm is designed for the multiuser receiver device having the best BER performance by
means of using the Field Programmable Gate Array (FPGA) integrated circuit.
Objects of the Invention
The object of the present invention is to realize a multiuser receiver always having a better
5
BER performance instead of the classical and traditional multiuser receivers, BER
performance of which decreases with the increasing fixed or mobile users.
Another object of the present invention, in order to achieve the same BER performance as
the existing receivers, is to realize a 'multiuser receiver' requiring less signal strength (Signal
to Noise Ratio, SNR) and being affected less from the interference (Near-Far Ratio, NFR)
10
emitted by the high strength users in the environment.
Description of the Figures
The channel, system model and performance curves of the multiuser receiver structure
realized for achieving the objects of the present invention are shown in the appended figures.
Figure 1: System model describing the time varying and frequency selective fading DS-
15
CDMA channel model.
Figure 2: FPGA integrated circuit multiuser receiver with a minimum bit error rate.
Figure 3: Flow diagram of the algorithm providing the minimum bit error rate.
Figure 4: Comparative performance curve of the realized multiuser receiver with the existing
classical (DEC and LMMSE) and with the best known BER performance (ML) receivers .
20
Figure 5: Comparative performance curve showing the amount of signal strength required for
providing the same BER performance (BER=10-2), when the number of users increases in
the environment.
Figure 6: Comparative performance curve of the realized receiver and the other receivers
(DEC and LMMSE), when the users in the environment have higher strength (NFR).
25
References in the Figures
1. Mobile terminal (mobile phone, etc.).
2. Base station
3. Multiuser receiver device
2
4. FPGA integrated circuit diagram
5. Flow diagram of the algorithm providing the minimum bit error rate
DEC: Performance curve of the multiuser decorrelating receiver
LMMSE: Performance curve of the multiuser Linear Minimum Mean Square Error receiver
5
MBER: Performance curve of the multiuser receiver according to the present invention with a
minimum BER value
ML: Performance curve of the multiuser Maximum Likelihood receiver with the best known
BER performance
BER: Bit error rate
10
SNR: Signal to noise ratio
dB: Decibel
NFR: Near-far ratio
Description of the Invention
A constrained optimization algorithm is developed for realizing the multiuser receiver with a
15
minimum bit error rate according to the present invention and the FPGA integrated circuit, on
which said algorithm is run, is used together with a digital signal processor (DSP).
For this purpose, BER cost function (Equation 1) of the time varying and frequency selective
fading channel DS-CDMA system model (Figure 1) is generated, said expression is
transformed into a constrained optimization problem (Equation 2) and filter coefficients
20
conferring the minimum value to the BER cost function are determined by means of the
iterations performed using Golden-Section and Newton methods along with a barrier
parameter.
,
(1)
(2)
The determined filter coefficients have the minimum bit error rate, thus the best BER
25
performance for the said channel and the system model.
Said optimization algorithm is as follows;
3
 Step 1
o 1.1. Users' spreading code, signal strength, data and channel matrices are determined.
o 1.2.
barrier value is selected such that
o 1.3. An appropriate starting filter coefficient,
5
o 1.4. An error parameter,
.
, is selected.
,to stop the iteration is determined.
o 1.5. Iteration number is set to
.
 Step 2
o
is set.
 Step 3
10
o 3.1.
is set and the next filter coefficient,
o 3.2. Iteration step value,
, is calculated.
, is calculated with the advancing direction,
,
through the Golden-Section method.
o 3.4.
The
next
value
of
the
expression
is
calculated
using
.
15
 Step 4
o If the condition,
, is satisfied, the expression
coefficient being sought. Iterations are stopped and
=
is the filter
is set. Otherwise,
is set and iterations are repeated starting from step 2.
20
Implementation of the Present Invention in the Industry
The multiuser receiver device realized by means of the FPGA integrated circuit, on which
said algorithm mentioned above in detail is executed, is provided with a structure that can
be used in the chip sets of the receiver or transceiver units of the base stations, which are
the basic components of the GSM or DS-CDMA communication systems.
25
4
CLAIMS
1) A multiuser receiver (3) for base stations (2) of the DS-CDMA/GSM systems,
characterized in that it comprises the FPGA integrated circuit (4) minimizing the bit
error rate in the time varying and frequency selective fading DS-CDMA channels.
5
2) The multiuser receiver (3) for base stations (2) of the DS-CDMA/GSM systems
according to Claim 1, characterized in that the constrained optimization algorithm (5)
minimizing the bit error rate is executed on the FPGA integrated circuit (4) provided
therein.
10
3) The multiuser receiver (3) for base stations (2) of the DS-CDMA/GSM systems
according to Claims 1 and 2, characterized in that the constrained optimization
algorithm (5) executed on the FPGA integrated circuit (4) provided therein minimizes
the bit error rate of the multiuser receiver (3) by using Newton and Golden-Section
15
methods.
5
ABSTRACT
A multiuser receiver device minimizing the bit error rate (BER) in the time varying and
frequency selective fading channels for the base stations used in the Direct Sequence Code
Division Multiple Access (DS-CDMA) or GSM systems.
6
Figure 1
Channel frequency
response
Multiuser
Receiver
k. User specific
code
k. User data
Mobile Terminal
k. Determined user
data
Channel
Base station
7
8/4
Figure 3
Users' Signal Strength, Spreading
Code, Data and Channel
Parameters are determined.
barrier value and 𝜀 error
parameter are selected.
Starting coefficient,
, is selected.
İteration number is set to i=0.
A
L
𝒘0 (𝒊) = 𝒘𝑖 is set.
G
μ = μl
O
is set and 𝒘𝑘 (𝒊) is calculated.
R
Iteration step value, 𝜍𝑘 (𝒊) , is calculated
I
with the advancing direction Ω𝒌 (𝒊) .
T
H
The next filter coefficient is
calculated.
M
𝒘𝑘+1 (𝒊) = 𝒘𝑘 (𝒊) + 𝜍𝑘 (𝒊) Ω𝑘 (𝒊)
Yes
𝒘𝑖+1 = 𝒘∗ is the filter
coefficient being sought.
Stop the process.
‖𝒘𝑖 − 𝒘𝑖+1 ‖ < 𝜀
No
Set 𝑖 = 𝑖 + 1 and
continue from step 2.
5
9/4
Figure 4
SNR required for BER=10-2 (dB)
Figure 5
Number of Active Users
10/4
Figure 6