METHOD AND RECEIVING APPARATUS FOR MINIMUM BIT ERROR RATE LINEAR MULTI-USER DETECTION OVER TIME- VARYING AND

FREQUENCY-SELECTIVE FADING DS-CDMA CHANNELS Technical Field

The present invention relates to a method and a multi-user receiving unit which minimizes the bit error rate (BER) over 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, all users communicate by using the same frequency band at the same time. In the CDMA systems, unique spreading code is assigned to each user, and therefore it is provided for all the users to use the whole band width commonly at the same time. DS- CDMA is the most commonly used CDMA type. The conventional multi-user detectors used in the receiver or transceiver chip sets of the base stations that are among the most fundamental parts of the DS-CDMA systems are the linear detectors, and they aim to suppress the interference between the users. Two of the known most classical linear multi-user detectors that are "Decorrelating" (DEC) and "Linear Minimum Mean Square Error" (LMMSE) minimize the multiple access interference and the mean square error, respectively. However, as such types of classical linear detectors have not been designed for minimizing the BER value. Therefore the BER performance of the base stations having these detectors decreases while the number of active users increases and some problems occur in the communication.

The best multi-user receiver structure that is known is the "Maximum Likelihood" (ML) receiver. The fact that this receiver has a complicated structure and requires a high ability of mathematical operation prevents it from being used in the real-time applications in the current technology. Therefore new multi-user receiver structures ar e needed which has a BER performance as best as in the ML receiver and which is less complicated.

Similar studies conducted previously are made only for the channel models where "Additive White Gaussian Noise" (AWGN) is available or for the frequency- selective fading channel models. In the present invention, on the other hand, a multi-user receiving units have been designed that are more suitable to the real world, better represent the mobile terminals, and have the best BER performance by creating a time-varying and frequency-selective fading channel model and making an algorithm reducing the BER value to the minimum level in this channel model and by using the FPGA (Field Programmable Gate Array) integrated circuit.

Object of the Invention

The object of the present invention is to compose a multi-user receiver with a better BER performance all the time instead of classical and conventional multi-user receivers having a BER performance that has been decreased with the increase in the number of fixed or mobile users.

Another object of the present invention is to create a "multi-user receiver" requiring less Signal to Noise Ratio (SNR) so as to achieve the same BER performance with the current receivers, and being affected less from the Near-Far Ratio (NFR) that the high-power users send out in the medium. Description of the Figures

The channel and system model, and the performance curves regarding the multi-user receiver structure that has been created for achieving the objective of the present invention are shown in the accompanying figures.

Figure 1 : It is the system model explaining the time-varying and frequency-selective fading DS-CDMA channel model.

Figure 2: It is the multi-user receiver having minimum Bit Error Rate and FPGA integrated circuit.

Figure 3: It is the flow chart of the algorithm providing the minimum Bit Error Rate.

Figure 4: It is the performance curve of the multi-user receiver having the best BER performance ever which is compared with the current classical DEC and LMMSE and the ML receivers having the best BER performance ever known.

Figure 5: It is the comparative performance curve showing the signal power ratio necessary for providing the same BER performance (BER=10 ^-2 ) when the number of users in the medium increases. Figure 6: It is the curve comparing the performance with the other receivers (DEC and LMMSE) when the other users in the medium have more power (NFR).

Description of the References in the Figures

1 : Mobile terminal (Mobile phone etc.) 2: Base station

3: Multi-User Detector (Receiver)

4: FPGA integrated circuit diagram

5: Algorithm flow chart providing minimum bit error rate

DEC: The performance curve of the Decorrelating multi-user detector. LMMSE: The performance curve of the Linear Minimum Mean Square Error multi-user detector.

MBER: The performance curve of the multi-user receiver according to the present invention having minimum BER value.

ML: The performance curve of the Maximum Likelihood multi-user receiver which has the best BER performance ever known.

BER: Bit Error Rate.

SNR: Signal to Noise Ratio. dB: Decibel.

NFR: Near - Far Ratio. Description of the Invention

In order to create a multi-user receiver having minimum bit error rate according to the present invention, an algorithm of constrained optimization has been developed and the FPGA integrated circuit on which the proposed algorithm is working has been used together with a Digital Signal Processor (DSP). Therefore, a BER cost function has been created (Equation 1) for the DS-CDMA system model (Figure 1) having time- varying and frequency-selective fading channel, this equation is converted into a constrained optimization problem (Equation 2), and the filter coefficients have been found which will make the BER cost function to assume the minimum value by means of the iterations performed by using the Golden-Section Method and the Newton method together with a barrier parameter.

The found filter coefficients have the minimum bit error rate and therefore the best BER performance for the said channel and system model.

The said optimization algorithm comprises;

• 1st Step.

o 1.1. The spreading code of the users, signal power, data vector and channel matrix are determined.

o 1.2. barrier value is selected so as to achieve .

o 1.3. A filter coefficient is chosen which is appropriate for the start.

o 1.4. A error parameter is determined which will stop the iteration.

o 1.5. The Iteration number is set as

• 2nd Step.

o Determined initial coefficient vector as

• 3rd Step.

o 3.1. It is set as and the next filter coefficient which is value is calculated.

o 3.2. The iteration step value is calculated with the progressing direction by means of the Golden-Section method.

o 3.4. is used and the next value of the expression is calculated.

• 4th Step.

o In the case that the condition is achieved, the expression is the filter coefficient that is sought. The iterations are stopped and it is set as

Otherwise, it is set as and the iterations are repeated starting from 2nd Step.

Type of Application to the Industry The multi-user receiver configured together with the FPGA circuit on which the above- mentioned algorithm is working can be used in the chip sets of the receiver or transceiver units of the base stations which are the fundamental parts of the GSM or DS-CDMA communication systems.