readme.txt

Space-Time Processing for CDMA Mobile Communications.Written for research students and design engine

WCDMA Uplink/Downlink Simulation Environment


NOTICE
These M-files are User Contributed Routines which are being redistributed
by The MathWorks, upon request, on an "as is" basis.  A User Contributed
Routine is not a product of The MathWorks, Inc. and The MathWorks and the authors assume
no responsibility for any errors that may exist in these routines.


1. Introduction


This 'readme.txt' provides a brief description of an integrated 3G CDMA simulation
platform for the comparison of the performance of the different systems' performance
under similar signalling conditions. The simulation environment has developed written
under MATLAB 5.2, running on a WINDOWS 95/98/NT platform. The complete MATLAB
simulation environment is available on the floppy disk included with the book,
as well as on the Internet at


    web: ftp://ftp.mathworks.com/pub/books/vanrooyen


or by anonymous FTP to


    Unix login: ftp ftp.mathworks.com
    Name: anonymous
    Guest login ok, send your complete e-mail address
    Password: (type in e-mail address)
    cd/pub/books/vanrooyen


2. Link Level Simulation


The simulation software is capable of simulating both the uplink and downlink of a
typical 3G CDMA systems, similar to that proposed for UMTS. Figures H.1 and
H.2 illustrate the general block diagrams of the transmission
links, respectively.


As has been described in Appendix G, the UMTS frame consists of multiple slots.
In the simulation, the key modules rely on frame and slot based processing for both
the receive and transmit functions. In the case of the transmitter, the frame based
processing consists of frame encoding and interleaving.


For the uplink a single transmitter anttena is assumed, while the downlink may include
M_T transmit antennas, using either CDTD or TDTD signalling. In the simulation,
the in-phase and quadrature components of the transmitted signal are multiplied by
a random segment of a pre-generated fading channel complex envelope. The channel models
that may be used, include the UMTS indoor, outdoor-to-indoor/pedestrian and vehicular
models. The resulting signals are then summed and finally AWGN of known power is
added as per a SNR specified.


For each user the data transmitted on the physical channel results from the encoding
of an information sequence with the control information randomly generated. The
length of the information sequence and the encoding rate sets the number of binary
symbols to be transmitted on the I and Q branches of the modulator. This in turn sets
the processing gain of each user. As defined in the preliminary UMTS standard, users
with higher information rates will have correspondingly lower spreading gains.
In addition to specifying the processing gain, frame interleaver sizes used in the
convolutional- and turbo encoding and decoding are also determined by the information
data rate.


In the receiver, the received signal is first processed by a chip matched filter.
Thereafter, a RAKE receiver consisting of a number of correlators (or fingers),
operating in parallel is used to process the received signal. Each finger correlates
a shifted version of the received signal with the spreading sequence for the user of
interest. The different shifts correspond to the different excess delays for each
multipath component received by the mobile terminal.  The outputs of the RAKE fingers
must be combined (once per symbol period) to obtain an estimate of the received
symbol. In addition to the standard operation of the RAKE receiver, in the case of
transmit diversity, channel estimation is performed on each resolved path, and used
in a pilot symbol assisted (PSA) RAKE combiner to resolve each of the transmitted
streams from the multiple transmit antennas.


Perfect power control is assumed on the dedicated channels to reduce the imbalance
in received power (near-far effect). Ideally, the base station adjusts the transmit
power of mobiles such that the base station observes a prescribed  (SIR). Both pilot
and data symbols are used in measuring the instantaneous received signal power, with
pilot symbols being used in the measurement of instantaneous interference plus
background noise power.


2.1 Monte-Carlo Simulation Technique


To enable statistically valid simulation results to be obtained in reasonable
simulation times, Monte Carlo methods are used.


For a given number of users, channel model and link configuration the aim of the
simulation is to produce a bit error rate curve as a function of the SNR. For each
SNR value, the UMTS uplink or downlink is simulated until a reliable estimate of
the bit error rate at the output of the multi-user detector can be obtained.
A simulation ``loop'' is defined as the transmission and reception of a 10 ms frame.
Simulation loops are continued until both of the following conditions are satisfied:
    - the number of bit errors detected by the receiver is greater than a
      specified minimum number of errors.
    - the number of simulation loops performed is greater than a specified
      minimum number of loops.


With the above two conditions met, the simulation will continue until one of the
following conditions is true;
    - the number of simulation loops reaches a specified maximum number of
      loops.
    - the current bit error rate is less than a specified minimum rate.


2.2 General Simulation Assumptions


- Inter-cell interference is not modeled.
- Narrowband interference is not modeled as a component of the channel model.
- Linear power amplifiers are assumed for both the transmitter and receiver.
- The UMTS uplink and downlink simulation described is based the preliminary
  specifications of the UTRA FDD system.
- For all receiver types, it is assumed that the receiver can synchronize to
  the received signal. No synchronisation errors are taken into account by the
  simulation.
- The RAKE receiver is provided with the excess delays for each multi path
  component processed. Thus, the simulation does not perform MPC delay
  estimation.
- The IC based receivers are provided with ideal channel estimates. It is true
  that the performance of the channel estimator in an IC environment will be
  better than other cases since the multiuser interference is greatly reduced
  in the signal supplied to the channel estimator.
- The AS-TDTD transmitter and turbo MAP decoder are provided with estimated
  channel estimates.


2.3 Simulation Cases


Three types of users, each having different service requirements, may be considered.
These are indicated in Table H.1.


Table H.2 provides a summary of the implemented receivers, transmit diversity, and
coded techniques and their corresponding labels.


3. MATLAB Simulation Software


3.1 Getting Started


To run the MATLAB simulation platform, the following steps should be followed:


    [Step 1]
    Create a suitable working directory to which the software will be copied.
    For example: `c:\wcdmasim'.


    [Step 2]
    Copy the downloaded `p-code' files (`*.p') to the working directory.


    [Step 3]
    Create the simulation data directory to which 'error' and 'log_file' results
    will be stored.  This directory should be created on the `C' drive as follows:
    `c:\data'.


    [Step 4]
    Start MATLAB, and add the directory created under Step 1 to the MATLAB path.


    [Step 5]
    You should be able to run the simulation platform. Type in `wcdmasim' at the
    MATLAB command line.


3.2 Main Simulation Window


By invoking 'umts_sim.p' at the MATLAB command line, the main GUI from which different
simulation engines are called from will be opened. A screen capture of this GUI window
is depicted in Figure H.3.


3.3 Simulation Environment Configuration


Figure xyz shows a screen capture of the simulation environment configuration window.
By selecting `Transceiver/Channel Setup' button, the configuration window, shown in
Figure H.4 will be displayed.


The DS/CDMA transceiver and environment parameters controlled through this GUI are
given below:


General transceiver parameters:
-> Number of simultaneous users, K.
-> Users' load in a mixed throughput environment, given as percentage of number of
   simultaneous transmitting users.
-> SNR range and step increments.


Channel environment parameters:
-> Type: AWGN, UMTS Indoor, UMTS Outdoor-to-Indoor and Pedestrian, and UMTS Vehicular.
-> Average speed and log-normal shadowing variance.


Monte-Carlo simulation parameters:
-> Minimum number of bit errors to detect.
-> Minimum and maximum number of frames to receive.


Parameters common to uplink and downlink:
-> Number of RAKE fingers available.
-> Power control algorithm selection.


Parameters specific to uplink:
-> Number of receiving antennas.
-> Choice of receiver (choice between single- and multi-user detectors):
    Iterated SIC, No clip.
    Iterated SIC, Clip.
    Iterated SIC, Hard.
    Iterated PIC, No clip.
    Iterated PIC, Clip.
    Iterated PIC, Hard.
    Estimated Matched Filter (EMF).
    Normalized LMS (NLMS).
-> Choice of FEC technique:
    No coding.
    Convolutional encoder with soft-input Viterbi decoder.
    Turbo encoder with iterative MAP decoder (8 Iterations).


Parameters specific to downlink:
-> Number of transmitting antennas.
-> Transmit diversity selection:
    No transmit diversity.
    Orthogonal Code-Division Transmit Diversity (CDTD).
    Round-Robin Time-Division Transmit Diversity (RR-TDTD).
    Antenna-Selection Time-Division Transmit Diversity (AS-TDTD).
-> Choice of receiver:
    Estimated Matched Filter (EMF).
    Normalized LMS (NLMS).
-> Choice of FEC technique:
    No coding.
    Convolutional encoder with soft-input Viterbi decoder.
    Turbo encoder with iterative MAP decoder (8 Iterations).


3.4 Example


The following example can be used as a reference describing the operation of the
simulation.


    [Step 1]
    Type `wcdmasim' at the MATLAB command line. This will bring up the main
    interface window (figure), shown in Figure H.3.


    [Step 2]
    Click on the `Transceiver/Channel Setup' button. This will open the
    configuration window, as shown in Figure H.4.


    [Step 3]
    Select the number of users, K.


    [Step 4]
    Change the 'E_b/N_o range in dB' entry to the desired range. Note that
    this parameter is entered in typical MATLAB style for vectors which is
    'start:step:end' with step defaulting to 1.0 if not specified.


    [Step 5]
    Select the desired channel environment.


    [Step 6]
    Select the vehicle speed and log-normal shadowing variance.


    [Step 7]
    Set up the users' loads as a percentage.  Upon exit the entries will be
    normalized to a total load of 100 %.


    [Step 8]
    Change the simulation control parameters.


    [Step 9]
    Set up the parameters common to both the uplink and downlink.


    [Step 10]
    Set up the uplink specific parameters.


    [Step 11]
    Set up the downlink specific parameters.  Note that when only a single
    transmit antenna has selected, that the transmit diversity scheme will be
    defaulted to the 'No Transmit Diversity (TD)' selection.


    [Step 12]
    Click on the `Continue' button. This causes the configuration window to
    close. The theoretical curve for selected uncoded DS/QPSK system will be
    plotted over the E_b/N_o range in dB.


    [Step 13 (Optional)]
    Click on the `Clear' button to remove the plotted curves.


    [Step 14]
    Click on either the `UPLINK Simulation' or `DOWNLINK Simulation' button to
    start the simulation. The simulation continuous for each E_b/N_o value
    specified in the 'E_b/N_o range in dB' entry. Information on 'Simulation
    Completion' will be displayed. The plot will be updated as each simulation
    loop is completed.  At completion of the simulation, a legend is added and
    results are displayed on the graph. At this stage, the result can be copied
    to the clipboard to paste in some other application for recalling purposes.


    [Step 15]
    The main figure window can now be exited from by clicking on the `Exit'
    button or alternatively, more simulations can be performed.


Contact information:
Please contact Danie van Wyk at [e-mail] in case of inquiries