Top Banner
MIMO TRANSMISSION USING PARASITIC ARRAY ANTENNA by Chamath Divarathne [email protected] B. Sc. Eng, University of Peradeniya, 2007 A THESIS submitted in partial fulfillment of the requirements for the degree MASTER OF SCIENCE IN INFORMATION NETWORKING (MSIN) Information Networking Institute (INI) The College of Engineering CARNEGIE MELLON UNIVERSITY 2010 Co-supervised by: Prof. Constantinos Papadias [email protected] Prof. Antonis Kalis [email protected]
60
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Thesis Report Chamath

MIMO TRANSMISSION USING PARASITIC ARRAY ANTENNA

by

Chamath Divarathne

[email protected]

B. Sc. Eng, University of Peradeniya, 2007

A THESIS

submitted in partial fulfillment of the requirements for the degree

MASTER OF SCIENCE IN INFORMATION NETWORKING

(MSIN)

Information Networking Institute (INI)

The College of Engineering

CARNEGIE MELLON UNIVERSITY

2010

Co-supervised by:

Prof. Constantinos Papadias [email protected]

Prof. Antonis Kalis [email protected]

Page 2: Thesis Report Chamath

ii

DECLARATION

I Chamath Divarathne, declare that the work presented in this thesis is original and no

part of it (including the document, the implementation code, etc.) has been copied

from other sources. Work related to this one is cited appropriately.

Chamath Divarathne

30.04.2010

The work contained in this thesis MIMO Transmission using Parasitic Antenna

Arrays by Chamath Divarathne has been carried out under my supervision.

Constantinos Papadias

30.04.2010

Athens Information Technology

Antonis Kalis

30.04.2010

Athens Information Technology

Page 3: Thesis Report Chamath

iii

ABSTRACT Multiple-Input-Multiple-Output (MIMO) transmission is one of the promising

antenna technologies used for wireless communication. When the transceiver uses

more than one antenna, the antennas must be placed at least half of the carrier

wavelength apart, in order to transmit/receive uncorrelated signals. As a result size

and cost sensitive devices like mobile terminals cannot facilitate MIMO transmission.

Parasitic antenna on the other hand, uses only one active RF front end along with two

or more parasitic elements which can be placed much closer resulting in a compact

design. The parasitic elements can be electronically switched to form different

antenna beams for the active RF feed, in each symbol period. Both theory and

computer simulations have shown that, these different beams can be used to have

MIMO transmission with traditional uniform linear array (ULA) antennas placed at

the receiver. This work involves configuring and customizing a classical MIMO

testbed to use a 3-element electronically steerable parasitic array radiator (ESPAR) at

the transmitter and traditional two ULA of antennas at the receiver, using BPSK

modulation. Vertical Bell labs Layered Space Time (V-BLAST) architecture was used

together with zero forcing equalizer at the receiver, to verify over-the-air MIMO

transmission using the parasitic antenna.

Page 4: Thesis Report Chamath

iv

ACKNOWLEDGEMENT

Take this opportunity with the great pleasure in repressing the appreciation and

profound personal regards for my co-supervisors Prof. Constantinos Papadias and

Antonis Kalis. I would like to pay a special gratitude to Stanford University for

donating the MIMO testbed to the B-WiSE laboratory of Athens Information

Technology. Finally the assistance and guidance from Mr. Frank McCarthy, Mr.

Osama Al Rabadi, Mr. Philippos Tragas and Mr. Keshav Singh throughout the work

are highly appreciated.

Page 5: Thesis Report Chamath

v

TABLE OF CONTENT

ABSTRACT........................................................................................................... III ACKNOWLEDGEMENT..................................................................................... IV

TABLE OF CONTENT ..........................................................................................V TABLE OF FIGURES .......................................................................................... VI

1 INTRODUCTION ........................................................................................... 1 2 LITERATURE SURVEY................................................................................ 3

2.1 MIMO TRANSMISSION................................................................................ 3 2.2 PARASITIC ANTENNA ARRAYS .................................................................... 5

3 MIMO TESTBED OVERVIEW................................................................... 10 3.1 TRANSMITTER HARDWARE UNIT ............................................................... 12

3.1.1 DSPTX chip main board....................................................................... 13 3.1.2 DAC40 daughter card. ......................................................................... 14 3.1.3 I/O Board............................................................................................. 16 3.1.4 MMDS RF modules. ............................................................................. 18 3.1.5 Thunderbolt GPS synchronization unit. ................................................ 21 3.1.6 RF Filters............................................................................................. 22 3.1.7 Power units. ......................................................................................... 22 3.1.8 IP Engine. ............................................................................................ 22

3.2 RECEIVER HARDWARE UNIT ..................................................................... 23 3.2.1 MMDS RF modules. ............................................................................. 24 3.2.2 ADC daughter card. ............................................................................. 25 3.2.3 IP Engine. ............................................................................................ 26 3.2.4 Time and Sync Card (TAS). .................................................................. 29

4 MIMO TRANSMISSION EXPERIMENT................................................... 30 4.1 HARDWARE CONNECTIONS ....................................................................... 30 4.2 TRANSMITTER UNIT CONFIGURATION......................................................... 32 4.3 EXPERIMENT SPECIFICATIONS ................................................................... 35 4.4 DOWNLOADING A CODE TO THE DSP ......................................................... 35 4.5 TRANSMITTER SIDE INTERFACE ................................................................. 36 4.6 RECEIVER SIDE INTERFACE........................................................................ 38

5 RESULTS ...................................................................................................... 41 6 CONCLUSION AND FUTURE WORK ...................................................... 45

REFERENCES...................................................................................................... 46 APPENDIX A ........................................................................................................ 47

APPENDIX B ........................................................................................................ 50

Page 6: Thesis Report Chamath

vi

TABLE OF FIGURES

FIGURE 1: MAJOR COMMUNICATION BLOCKS USED IN MIMO TRANSMISSION ............... 1 FIGURE 4: NORMALIZED 3-ELEMENT ESPAR ANTENNA BASIS FUNCTIONS .................. 6 FIGURE 5: MIMO TESTBED SETUP............................................................................ 10 FIGURE 6: TRANSMITTER UNIT BLOCK DIAGRAM........................................................ 12 FIGURE 7: SNAPSHOT OF THE TRANSMITTER UNIT ..................................................... 13 FIGURE 8: DSP MAINBOARD BLOCK DIAGRAM ......................................................... 13 FIGURE 10: SNAPSHOT OF THE IO BOARD.................................................................. 16 FIGURE 11: CABLE CONNECTION OF THE IO BOARD................................................... 17 FIGURE 12: TRANSMITTER RF MODULES .................................................................. 18 FIGURE 13: GPS SYNCHRONIZATION UNIT ................................................................ 21 FIGURE 15: RECEIVER UNIT BLOCK DIAGRAM ............................................................ 23 FIGURE 16: SNAPSHOT OF THE RECEIVER UNIT........................................................... 24 FIGURE 17: ADC BLOCK DIAGRAM .......................................................................... 25 FIGURE 18: SNAPSHOT OF THE IP ENGINE.................................................................. 26 FIGURE 19: CABLE CONNECTION BETWEEN DSP & IP ENGINE ................................... 27 FIGURE 20: BLOCK DIAGRAM OF IP ENGINE.............................................................. 27 FIGURE 21: HIGH LEVEL SYSTEM DIAGRAM OF THE MIMO TESTBED ........................ 30 FIGURE 22: LO CONNECTIONS .................................................................................. 31 FIGURE 23: TRANSMITTER SIDE PROCESSING ............................................................. 32 FIGURE 24: SCHEMATIC DIAGRAM OF THE COMPARATOR CIRCUIT USED...................... 33 FIGURE 26: COMPARATOR INPUT AND OUTPUT SIGNAL .............................................. 34 FIGURE 27: SNAPSHOT OF THE OSCILLOSCOPE SCREEN FOR COMPARATOR INPUT AND

OUTPUT ............................................................................................................ 34 FIGURE 28: DSP MAINBOARD................................................................................... 35 FIGURE 29: TRANSMITTER SIDE INTERFACE ............................................................... 36 FIGURE 30: INPUT PARAMETERS FOR BASEBAND SIGNAL GENERATION........................ 37 FIGURE 31: RECEIVER SIDE INTERFACE SNAPSHOT..................................................... 38 FIGURE 32: ADC POWER MEASUREMENT UTILITY INTERFACE.................................... 39 FIGURE 33: IN-PHASE AND QUAD-PHASE SIGNALS OF ONE OF THE RECEIVING ANTENNA

........................................................................................................................ 40 FIGURE 35: RECEIVED CONSTELLATIONS AT THE TWO ANTENNAS.............................. 42 FIGURE 36CONSTELLATION DIAGRAM FOR THE TWO SUB STREAMS............................ 43 FIGURE 37: BER COMPARISON FOR THE THREE SCENARIOS ....................................... 44

Page 7: Thesis Report Chamath

1

1 Introduction

Mobile communication is well popular among the world in day to day life. With the

ever increasing sophisticated mobile applications, the requirement of having high data

rate communication is a challenge. There are several technologies which caters this

requirement such as W-CDMA, and WiMax.

Multiple-input-multiple-output (MIMO) transmission is one of the promising antenna

technologies used for wireless communications. Through spatial multiplexing, MIMO

achieves high capacities. The only limitation is that, the transmitting and receiving

antennas should be placed at least half the wave length of the carrier signal in order to

transmit or receive uncorrelated signals. Apart from that, each of transmit or receive

antenna requires a separate circuit which means, higher the no of antennas used higher

the cost.

Modulation Up Conversion

Modulation Up Conversion

DemodulationDown Conversion

DemodulationDown Conversion

Transmitter Receiver

… …

Modulation Up Conversion

Modulation Up Conversion

DemodulationDown Conversion

DemodulationDown Conversion

Transmitter Receiver

… …

Figure 1: Major communication blocks used in MIMO transmission

Due to these reasons, implementation of traditional MIMO transmission with multi-

element (MEA) arrays on cost and size sensitive devices like mobile terminals is not

feasible. A novel approach to MIMO transmission using a single RF front end was

introduced in [1] and [2]. In [1], a novel methodology for using electronically

steerable parasitc array radiator (ESPAR) antennas as a compact, cheap and less

power-hungry MEA, for spatial multiplexing BPSK symbols. The simulated results

have shown that this methodology still provides comparable spectral efficiency and

performance as traditional MEA.

Page 8: Thesis Report Chamath

2

This work involves verifying over-the-air MIMO transmission using a 3-element

ESPAR at the transmitter and a traditional 2 MEA using at the receiver. A traditional

MIMO testbed was used and customized to facilitate the ESPAR at the transmitter.

The received data is analyzed using Matlab to evaluate performaces.

Page 9: Thesis Report Chamath

3

2 Literature Survey 2.1 MIMO Transmission In the never-ending quest for capacity increment in wireless communication, MIMO

system architecture has shown the possibility of increasing the capacity substantially.

Fading is considered as a challenge in wireless communication. But in MIMO

channels it uses the fading to obtain high spectral efficiency.

Spectral efficiency is the capacity of the channel per unit bandwidth, which is

measured in bits per second per Hertz (bps/Hz). MIMO transmission achieves high

spectral efficiency using the spatial multiplexing. In spatial multiplexing different

signals are transmitted from different antennas at the same time, so that receiving

antenna array receives a superposition of all the transmitted signals. The receiver

solves a linear equation system to demodulate the message. This provides an

additional degree of freedom, which increases the spectral efficiency. Figure 2

illustrates the idea behind spatial multiplexing.

One of the common receiver processing schemes used for solving these linear

equation system is called Vertical Bell lab LAyered Space Time (V-BLAST) [3]. V-

BLAST together with zero forcing equalizer is briefly explained in terms of linear

equations.

Tx RxTx 1

Tx 2

Rx 1

Rx 2Data Stream Data Stream

Figure 2: Spatial Multiplexing

Page 10: Thesis Report Chamath

4

Assume,

NR = No. of receiving antennas

NT = No. of transmitting antennas

L = Burst size

H = Channel matrix with dimensions (NRxNT)

H# = Pseudo inverse of H (NRxNT)

X = Transmitted BPSK matrix (NTxL)

X^ = Decomposed BPSK matrix (NTxL)

W = Decomposing matrix (NTxNR)

Z = Received signal matrix (NRxL)

The received signal mixed by the channel is given by,

Z = H . X

Zero forcing equalizer yields,

W = H#

Finally, the decomposed BPSK symbols were given by,

X^ = W.Z

There are other receiver processing schemes such as Maximum Ration Combining

(MRC), MRC with Successive Interference Cancellation (MRC-SIC) and Maximum

Likelihood decoder.

Page 11: Thesis Report Chamath

5

2.2 Parasitic Antenna Arrays ESPAR is a smart antenna system which is able to control the beam patterns

dynamically while being implemented using a single active antenna and number of

parasitic elements placed around the active element. The parasitic elements are short

circuited and loaded with variable reactors (varactors), to control the input impedance.

By adjusting the varactors’ response beams can be formed in certain directions either

in an adaptive or predefined fashion. [4]

When the ESPAR antenna is used for the uplink transmission scenario, where the

receiver is equipped with traditional ULA, the system is physically a single input

single output (SIMO) system as there is only one active element at the transmitter.

However instead of sending symbols using uncorrelated antennas like in the

traditional approach, they are mapped onto uncorrelated set of patterns as follows.

Figure 3: 3-element ESPAR Antenna

Page 12: Thesis Report Chamath

6

si is the ith symbol to be spatially multiplexed and bi is the ith basis function

composing the ESPAR pattern

The far field pattern of a three-element ESPAR antenna with an inter-element spacing

of d can be written as,

B1 and B2 are orthogonal to each other theoretically; they can be simplified as

follows.

B1 = 1 (omni directional)

B2 = cos(θ)

Following figure shows two basis functions B1 and B2 with normalized power, when

the inter-element spacing d = λ/16.

Figure 4: Normalized 3-element ESPAR Antenna Basis functions

Page 13: Thesis Report Chamath

7

By estimating response from each receiving antenna to each basis function

symbols are de-multiplexed and later demodulated. For a system with 2 receiving

antennas and 2 basis functions at the ESPAR antenna, then the radiation pattern would

look like

G = [g0 g1 … gL-1]T

where gl is the sample of the array pattern. The patter is decomposed into a

linear combination of two basis functions such that they are orthogonal to each other.

b1 = [b1,0 b1,1 … b1,l-1]

b2 = [b2,0 b2,1 … b2,l-1]

let hl be the complex gain of the path l. Then the responses of the two receive

antennas to the basis function b1 + b2 (corresponding to the symbol pair [1 1]T) will

be,

Equivalently,

which yields to,

hij is the response from ith receiving antenna to the jth basis function. It is

evident that linear combination of field patterns gives a linear combination of antenna

responses.

Page 14: Thesis Report Chamath

8

Now the receiver can simply decompose the vector of the receiving antenna

responses,

that denotes the SIMO channel, into a MIMO channel as follows.

Possible radiation patterns

B [1 1] = b1 + b2

B [1 -1] = b1 – b2

B [-1 1] = –(b1 – b2)

B [-1 -1] = –(b1 + b2)

These four combinations can be generalized as,

s1 (b1+(-1)sb2)

s = b1xorb2

That’s why the active element of the parasitic antenna is always fed with the

first symbol stream and the depending on the ‘xor’ed result of the two streams, the

parasitic elements are switched using pin diodes.

A training sequence can be used to train the receiver as follows.

Sending BPSK symbol pairs [1 1]T and [1 -1]T by transmitting (b1+b2) and (b1-b2)

respectively, gives the response first receiving antenna to the two beams as follows,

So that, the response from the first receiving antenna to the basis function b1 and b2

can be obtained as follows.

Page 15: Thesis Report Chamath

9

The same procedure applies to the second antenna as well. Finally the

coefficients of H are fed to the V-BLAST algorithm as discussed in the previous

section to demultiplex and demodulate the original symbols.

Page 16: Thesis Report Chamath

10

3 MIMO Testbed Overview MIMO Test bed at a glance,

No. of MMDS Transmitter RF modules – up to 2

No. of MMDS Receiver RF modules – up to 2

Carrier Frequency – 2.5 to 2.7 GHz (MMDS Band)

Signal Bandwidth – up to 1 MHz

GPS synchronization feature.

Stores gathered data into a computer, connected via 10BaseT Ethernet.

Interfaces for tapping the Intermediate Frequency (IF) signal from the RF

modules (only at receiver RF modules).

Rx Antennas

Host Computer

Tx Unit

Rx Unit

Tx Antennas

Figure 5: MIMO Testbed Setup

Above figure shows the overall setup of the MIMO testbed including transmitter and

receiver units, transmitting and receiving antennas together with the host computer.

The MIMO test bed consists of two major prototype units.

1. Transmitter Unit.

2. Receiver Unit.

Page 17: Thesis Report Chamath

11

Transmitter Unit

This unit consists of following major sub-units.

DSPTX chip main board.

DAC daughter board (attached to DSP main board).

I/O board.

MMDS RF modules. (2 Nos)

Thunderbolt GPS synchronization unit.

RF Filters.

Power units (2 Nos)

IP Engine.

Receiver Unit

This unit consists of following major sub-units.

DSPRX chip main board.

ADC daughter card (attached to DSP main board).

I/O board.

Thunderbolt GPS synchronization unit.

MMDS RF modules. (3 Nos)

RF Filters.

Power units (2 Nos)

IP Engine.

Time and Sync Card (TAS)

Page 18: Thesis Report Chamath

12

3.1 Transmitter Hardware Unit The transmitter hardware unit is responsible for converting bits to modulated

baseband signals and then finally to transmit with a RF carrier using the available

antennas. The control signal path and the data signal path is illustrated in figure ??.

DSP Mainboard

IO Board

Power UnitipEngine

DAC

Master

Slave

GPS Unit

RF Modules Front-end Filters

Reference Signal

Control signal pathData signal path

Figure 6: Transmitter unit block diagram

First the code writing for the Digital Signal Processor (DSP) is downloaded to the

DSP using RS232 serial cable from the host computer. Then the DSP cofigures the

two RF modules via the IO Board. The master RF module takes a 10 MHz reference

signal from the GPS unit, which is used to generate the internal local oscillator

frequencies.

Once the RF module initialization is completed successfully, DSP sends Digital to

Analog Converter (DAC) with the digitized data generated for the baseband signal

depending on the modulation scheme. Each RF module has to be fed with in-phase

and quadrature-phase baseband signals. So DAC uses 4 independent channels to

generate these baseband signals for the two RF modules namely, master and slave.

Finally the output RF signals produced by RF modules are passed through the front-

end RF filter before transmitting using the transmitting antenna.

Page 19: Thesis Report Chamath

13

RF Modules

DSP Mainboard

ipEngine

Power Unit

IO Board

GPS Unit

RF Filters

Figure 7: Snapshot of the Transmitter Unit

Description of each sub module found in the transmitter unit is given below. 3.1.1 DSPTX chip main board.

Figure 8: DSP Mainboard Block Diagram

DSP chip – Innovative DSP 320C6201. (Clock to 200 MHz)

Digital I/O port – a 40-pin connector with 32 data pins. (JP1)

JTAG Port – connect via JTAG debugger to the host. (JP 8)

Serial Port – communicate with the RS232 port of a PC

Page 20: Thesis Report Chamath

14

Omnibus – DAC40 daughter card is connected using this. (JP12 &

JP17)

There are two possible ways to download code from the PC host to the target

(DSPTX). First method is using JTAG debugger. The C program is written in Code

Composer Studio and downloaded to the DSP using the JTAG debugger. The second

method is to RS232 Serial port to connect the PC to the relevant port in the DSPTX.

For that the software program called ‘Uniterminal’ is used. The COFF file is

generated using the Code Composer Studio (ex. Hello.out) and downloaded to the

DSP.

First of all generation of the analog baseband signal is described. The C code

generates the baseband signal and uses OMNIBUS to transfer those data to the Digital

to Analog converter (DAC40).

More details can be found in the CH11 of [11].

3.1.2 DAC40 daughter card.

DAC40 OMNIBUS module provides 4-channel, high-speed (40MHz)

14bit resolution analog-to-digital conversion facility.

Used to generate two (2) In-phase and two (2) Quad-phase signals for

the two MMDS RF modules. (Master and slave)

The DAC40 OMNIBUS module gives the target processor card four channels

of very high speed 40 MHZ, 14-bit resolution digital input to analog output

conversion (D/A). This makes the DAC40 module well suitable for use in high-speed

signal and arbitrary waveform generation, communications, and control systems. The

DAC40 module uses four Analog Devices AD9764 D/As (one for each channel)

along with the necessary output buffering, allowing for independent channel control.

The AD9464 features 72 dB total harmonic distortion and a spurious free dynamic

range of 75 dB, which is ideal for communications applications. The output is +/- 2V

(max) into a 50-ohm load, which is DC accurate and can be calibrated.

Page 21: Thesis Report Chamath

15

The DAC40 module has special memories (SARAM) placed between the DSP

and the D/As. This allows the module to load waveform tables into memory for

uninterrupted repeated playback independent of the DSP. This relieves the DSP from

servicing the D/A at the high rates generally required for waveform generation. The

DSP simply loads and playback table into RAM after which it may be played back

continuously at full rate. The DSP may update the contents of the SARAM while

playback is occurring without interrupting the current playback. Additionally, the

DSP may bypass the SARAM and communicate directly with the DAC when

necessary.

These four DAC channels are used to generate In-phase and Quadrate-phase

baseband inputs to the each RF modules. So with one DAC40 module two RF

modules can be fed with corresponding baseband signals.

After the baseband signal is generated successfully with the help of DAC40,

RF module configuration is described. DSP uses 32 bit Digital port to communicate

with RF module. First control and data signals are transferred from DSP main board

to the I/O board.

More details on DAC40 can be found in CH8 of [9]

Figure 9: DAC40 Black Diagram

Page 22: Thesis Report Chamath

16

3.1.3 I/O Board.

Acts as a facilitator to configure and check the status of the MMDS RF

modules.

Communicate with DSP, using the W7A cable, connecting JP 1 of

DSP main board and J-15 of IO board.

Communicate with RF modules, using W2A and W1B cables.

Figure 10: Snapshot of the IO Board

The I/O Board for the Gigabit Wireless Prototype is a custom PCB that is

meant to provide power and signal interconnects between internal components. A top-

level block diagram of the I/O Board is shown below.

Page 23: Thesis Report Chamath

17

Figure 11: Cable connection of the IO Board

So I/O board receives control and data signals from the DSP main board and

forward them to the corresponding RF module which has been already configured as a

master or slave.

More details on I/O Board can be found in [10]

Page 24: Thesis Report Chamath

18

3.1.4 MMDS RF modules.

MMDS RF moduels operates in the MMDS frequency bands which is from

2.5 GHz to 2.7 GHz.

Figure 12: Transmitter RF Modules

Currently there are two RF modules namely master and slave.

Each module takes in in-phase and quad-phase from the DAC40.

Master RF module takes a 10MHz carrier signal from GPS

synchronization unit (using W18B cable) and according to the LO

configuration given in the following reference, master shares LO1 and

LO2 output carrier signals with the slave RF module.

Finally the modulated RF signals are produced and they are connected

to the RF filters using W14D and W14E RF cables.

Following table describes the connections and different interfaces of RF

modules.

Page 25: Thesis Report Chamath

19

Page 26: Thesis Report Chamath

20

Apart from that each RF module contains a 25-pin control connector. One

analog input signal provides for control of the transmitter output level (by controlling

the IF gain). The remaining signals are digital and provide for radio configuration,

control, and status monitoring. The pin numbers, signal names, and functions for the

DE25P connector are described in the following table.

Page 27: Thesis Report Chamath

21

More details on MMDS RF modules can be found in [11] and Cable

configuration for MMDS RF modules can be found in [12]

3.1.5 Thunderbolt GPS synchronization unit.

Figure 13: GPS Synchronization Unit

Takes in the output of the GPS antenna and produces a 10MHz carrier

signal.

That signal is connected to the MMDS RF master module.

The status of the GPS unit is transferred to the front panel LED using

the RS 232 cable shown.

Page 28: Thesis Report Chamath

22

3.1.6 RF Filters.

Located between MMDS RF modules and antennas.

A19 and A18 are the RF filters for master and slave MMDS RF

modules respectively.

3.1.7 Power units.

These power units provide +5 VDC, +9 VDC, +24VDC and +/- 15VDC to I/O

board. Then I/O board is responsible for distributing power among each module.

More details can be found at Figure 1: in [10]

3.1.8 IP Engine.

Provides an interface with a 10BaseT Ethernet to store the data

collected from DSP into a Secondary device like hard disk.

More important at the receiver end than the transmitter end.

Detailed description for IP Engine is provided under the Receiver Unit section.

Figure 14: Front-end RF Filters

Page 29: Thesis Report Chamath

23

3.2 Receiver Hardware Unit

Receiver unit is responsible for demodulating the original baseband signal and

sending the digitized version of the received baseband signal for further signal

processing. The control signal path and the data signal path of the receiver unit is

illustrated in figure ??.

DSP Mainboard

IO Board

Power UnitipEngine

ADC

Master

Slave

GPS Unit

RF Modules Front-end Filters

Reference Signal

Control signal pathData signal path

Figure 15: Receiver unit block diagram

Like in the transmitter unit, the code written for the DSP should be downloaded using

the RS232 serial cable. Then the DSP configures the two RF modules namely master

and slave, via the IO Board. Again the master RF module expects a synchronized

reference signal of 10 MHz from the GPS unit.

Once the configuration is successful, the received signals which already passed

through the frond-end filters are fed to the RF modules. Using the internal local

oscillators, they are down-converted to IF stage and then to baseband. Each module

produces the two baseband signals like in the transmitter unit, and those are fed to a

Analog to Digital Converter (ADC). DSP can read those digitized baseband signals

and can be used for further signal processing. If the signal processing has to be taken

at the host computer, they can be transferred to the host using the RS232 cable.

Page 30: Thesis Report Chamath

24

RF Modules

Power Unit

IO Board

GPS UnitDSP Mainboard

ipEngine

Figure 16: Snapshot of the receiver unit

Receiver unit is almost similar to the transmitter except for few hardware modules.

Those will be discussed in detail here. First the received RF signal is passed through a

RF filter and then fed to each RF module.

3.2.1 MMDS RF modules.

There are two RF modules namely master (M) and slave (S).

Each module takes a RF signal from the receiving antenna.

Master RF module takes a 10MHz carrier signal from GPS synchronization

unit (using W18B cable) and according to the LO configuration given in the

following reference, master shares this carrier signal with the slave RF

module.

Each MMDS RF module provides an interface to access IF signal.

Finally, it produces the appropriate In-phase (I-OUT) and Quad-phase (Q-

OUT) signals (in baseband) for RF signals in each MMDS RF modules. This

same signal can be accessed with two different interfaces. (I-OUT 1, I-OUT 2,

Q-OUT 1, Q-OUT 2).

These demodulated analog baseband signals are then connected to the Analog-to-

Digital Converter (ADC).

More details on MMDS RF modules can be found in [13] and cable configuration for

MMDS RF modules can be found in [14]

Page 31: Thesis Report Chamath

25

3.2.2 ADC daughter card. A4D1 OMNIBUS module provides 4-channel, high-speed (10MHz) 14-bit resolution

analog-to-digital conversion facility. Plus a channel for high-speed (10MHz) 14-bit

resolution digital-to-analog conversion.

Used to digitize In-phase and Quad-phase analog signals from the three MMDS RF

modules.

Figure 17: ADC Block Diagram

The A4D1 OMNIBUS module provides the target card processor with four channels

of very high speed 10 MHz, 14-bit resolution analog input to digital output conversion

(A/D). In addition to a single channel of very high speed 10 MHz, 14- bit resolution

digital input to analog output conversion (D/A). This makes it well suited for high-

speed data acquisition applications, transient capture, data processing, and control

systems. The A4D1 module uses two pairs of Analog Devices AD9240 A/Ds and one

AD9774 D/A to provide for excellent dynamic range over a wide input bandwidth.

The A/D’s use a novel four stage pipelined architecture as well as a wideband sample-

and-hold amplifier making them well suited for direct IF down conversion extended

to 45 MHz. The A/D delivers 10 MHz data from a pipeline, which is only four

samples deep, resulting in low data latency. In addition, each A/D channel has

gain/offset error adjustment for accurate measurements.

Page 32: Thesis Report Chamath

26

A one (1) kword FIFO on each channel separates the A/Ds and D/A from the data bus

allowing the DSP time for other tasks while the FIFOs are filling or emptying. The

FIFOs allow the DSP to collect and transport the data from the A/D’s as single points

or as a data set of up to 1K sample size. This reduces the interrupt rate to the host DSP

allowing for highly efficient data connection.

More details on A4D1 can be found in CH2 of [9] 3.2.3 IP Engine. Provides an interface with a 10BaseT Ethernet to store the data collected from DSP

into a Secondary device like hard disk.

The DSP has to be programmed so that it will construct the required Protocol Data

Unit (PDU) with the gathered data.

Figure 18: Snapshot of the IP Engine

Control Signals are sent from the DSP to the IP Engine with the help of the cable

shown in the following figure. IP Engine is at the left side of the picture.

Page 33: Thesis Report Chamath

27

Figure 19: Cable connection between DSP & IP Engine

The following diagram shows the functional block of IP Engine. In fact this unit is

mounted on the IP Engine Motherboard. So these interfaces are accessible via the

connectors available in the motherboard.

Figure 20: Block Diagram of IP Engine

Bright Star Engineering’s ipEngine credit-card sized processor module provides an

ideal core for a network enabled product. The ipEngine utilizes the Motorola

PowerPC MPC823 processor with an array of on-chip peripherals including Ethernet,

USB, LCD/Video, I2C and serial controllers. The on-board flash memory file system

Page 34: Thesis Report Chamath

28

provides storage for the operating system as well as OEM application software and

data. The ipEngine has the following significant features:

66 MIPS Power PC CPU

16 MB DRAM

2 MB FLASH

10Base-T Ethernet

On-board 16- Watt Power Supply

16,000 Gate FPGA

Virtual I/O Interface

USB Host/Slave Controller

LCD/Video Controller

Dual RS-232

The external interface to the ipEngine hardware is via an FPGA-based "virtual

interface" which can be configured on the fly to adapt to the OEM's needs. The FPGA

can be configured to emulate a variety of bus architectures as well as to implement

peripheral functions like UARTs, PWM control, memory emulation, data capture and

synthesis, and interfaces to a variety of input devices. The ipEngine provides a

firmware boot loader programmed into the flash memory "bootblock" at the factory.

The bootloader is used to initialize the ipEngine and load the primary operating

system from flash memory, from the network or from the serial console. The ipEngine

includes a run-time license for BSE's pKernel real-time network operating system.

The ipEngine can also be readily used with other operating systems like VxWorks,

Windows CE, PSOS, QNX, eCos or Linux.

When combined with BSE’s pKernel software environment, the ipEngine can

significantly reduce the time, cost and complexity required to "network enable" a

product by providing a vertically integrated "network engine" complete with the

required hardware, software, and development environment. By employing POSIX

and ANSI C standards, the ipEngine's pKernel software environment leverages a

familiar programming environment and facilitates migration of software from UNIX

and Windows 95/NT platforms. The pKernel SDK provides an integrated

development environment based upon the industry standard GNU tool chain. The

Page 35: Thesis Report Chamath

29

software developer’s toolkit includes the gcc cross-compiler, a linker, archive

librarian and other utilities. Source level over-the-network debugging of code running

on the target system is accomplished via a thread aware version of the GNU GDB

debugger.

Using the IP Engine the received data samples are saved in the host PC. Then with the

help of a Matlab code, receiver processing is done in order to extract the transmitted

digital data.

3.2.4 Time and Sync Card (TAS).

Responsible for synchronization at the receiver.

TAS card provides signals from two time bases, either 10MHz reference

signal’s time base or sample clock reference signal’s time base.

Mounted on top of the RF filters at the receiver unit.

Currently this card is not used.

More details on TAS card can be found in,

“C:\Chamath\Documents\Selected Docs\Time_Sync\ Design_Reqs_TAS_2.doc”

Page 36: Thesis Report Chamath

30

4 MIMO Transmission Experiment In this experiment the MIMO testbed was configured to facilitate the 3-element

ESPAR antenna array at the transmitter and the traditional 2-element array of

antennas at the receiver.

4.1 Hardware Connections

Following figure illustrates the high level system diagram of the MIMO testbed.

Figure 21: High Level System Diagram of the MIMO Testbed

Internal cable connections of TX and RX units are described in [12] and [14]

Both transmitter and receiver units should be supplied with a 10 MHz reference signal

with an amplitude of 0-10 dBm. Each Tx and Rx unit has a ‘Trimble Thunderbolt’

GPS synchronization module which provides this 10 MHz reference signal.

At the moment instead of using the GPS synchronization, two replicas of a 10 MHz

tone was generated with the help of the frequency synthesizer available at the

transmitter side GPS synchronization module. This was possible as both transmitter

TX

RX

GPS Sync Unit

PC

10 MHz Ref. Signal Download

the code to DSP

Upload a snapshot

Page 37: Thesis Report Chamath

31

and the receiver modules are placed physically nearby. Those two synchronized signal

outputs are connected to the ‘10 MHz Ref In’ channel of the ‘Master RF Module’ of

each transmitter and receiver units.

Master RF Module is responsible for transferring the LO output signal to the slave RF

modules. Following figure illustrates the cable connection at the transmitter side RF

modules.

Figure 22: LO connections

The MIMO testbed configuration at the receiver is the same as the traditional 2x2

MIMO case. So the transmitter unit configuration is discussed in detail first.

Page 38: Thesis Report Chamath

32

4.2 Transmitter unit configuration

Just like in the traditional 2x2 MIMO transmission, the BPSK modulated first bit

stream is fed to the master RF module in the transmitter unit and its output is

transferred to the RF feed in the 3-element ESPAR antenna array. The major

modification is that, the second bit stream is not modulated like in traditional 2x2

MIMO transmissions. Instead the bits in the two streams are XORed in each symbol

duration and depending on the result, a switching signal is generated at one of the

available DAC channels. This procedure is illustrated in figure 23.

BPSKModulation

XOR Comparator

BPSKModulation

XOR Comparator

Bit Stream

RF Feed

Switching Signal

Figure 23: Transmitter side processing

With the DAC specifications, the maximum levels that can be produced are, +2 Volts

and -2 Volts. There has to be at least 6 Volts in order to drive the pin diodes in the

ESPAR antenna array. So an additional comparator circuit is used at transmitter side

to amplify this switching signal produced by the DAC channels to a level which is

large enough to drive the pin diodes. The function used for DSP programming can be

found at Appendix A.

A schematic diagram of the comparator circuit is shown in figure 24.

Page 39: Thesis Report Chamath

33

Figure 24: Schematic diagram of the comparator circuit used.

A snapshot of the actual switching circuit setup is shown in figure 25.

The simulated performances of the comparator circuit is shown in figure 26.

Figure 25: Snapshot of the switching circuit setup

Page 40: Thesis Report Chamath

34

Figure 26: Comparator input and output signal

After implementing the comparator circuit, the DAC output and the comparator

output was compared using a oscilloscope. The figure 27 shows that the switching

signal conforms to the design specification.

DAC Output

ComparatorOutput1 µS

5 V

Figure 27: Snapshot of the oscilloscope screen for comparator input and output

Page 41: Thesis Report Chamath

35

4.3 Experiment Specifications

Following are the specification used for the MIMO transmission experiment used for

3-element ESPAR antenna array.

BPSK Modulation

Baseband Frequency - 500 kHz

Baseband signal - Raised cosine

Training sequence - 8 symbols

Burst size - 410x2 symbols

Switching signal level - 8 V

V-BLAST with Zero Forcing Equalizer at the receiver

4.4 Downloading a code to the DSP

The C code is written with the help of the Code Composer Studio IDE, and the

output file (xxx.out) is generated. Then ‘Uniterminal’ application developed by TI is

used to download the COFF (xxx.out) file to the DSP using the RS232 serial

interface. The Ribbon serial cable is connected to the DSP motherboard as shown in

the following figure.

Figure 28: DSP Mainboard

Page 42: Thesis Report Chamath

36

The code used can be found at [9]. The same code has to be compiled and

build with debug variables DSPTX and DSPRX for the transmitter and receiver unit

respectively. Setting up the Code Composer Studio, project is explained in [10].

4.5 Transmitter side Interface Once the transmitter side .out file (readpow.out) is dumped successfully using

uniterminal, the screen of the uniterminal should be as follows.

Figure 29: Transmitter side interface

Center frequency can be configured with the option 2. It is configured to be at 2.6

GHz by default. Then option 8 is selected in order to generate a baseband signal at the

transmitter.

Page 43: Thesis Report Chamath

37

Figure 30: Input parameters for baseband signal generation

Parameter C is the baseband frequency in kHz s. N is the no of samples per symbol. A

and B are the amplitudes of the baseband symbol for Master and Slave RF modules

respectively. Currently in the code the baseband signal level is made to be fixed and

as long as the amplitude is not zero and within the range, it generates a baseband

signal for that channel with fixed amplitude hard-coded at the ‘vfun.c’ file. For this

experiment, amplitude for both channel A and B has to be given a non zero value.

Channel A is used to feed the master RF module and one DAC channel from channel

B is used to generate the switching signal. That’s why amplitude for both channels

has to be given a non zero value. Otherwise the free DAC channels used for channel

B will not be able to use.

RF output power is given at last. If the antennas are much closer, it’s better to use a

less power (with the currently used antennas it’s about 5 dB) if not received signal

should be connected to the receiver unit via an attenuator. Once the input parameters

are provided, the summary page would look as figure 30.

Page 44: Thesis Report Chamath

38

4.6 Receiver side Interface At the receiver side, after the code downloading completed successfully the

uniterminal screen should look as follows.

Figure 31: Receiver side interface snapshot

In order to run the ‘ADC Power Measurement Utility’ option 5 has to be selected. All

the features of the utility are illustrated in the following figure.

Page 45: Thesis Report Chamath

39

Figure 32: ADC Power measurement utility interface

By pressing ‘S’ followed by ‘U’ a snapshot of the received baseband signals can be

uploaded to the computer.

A snapshot of the in-phase and quad-phase signals from one of the receiving antenna

is shown in figure 33.

Page 46: Thesis Report Chamath

40

500 mV

5 µS

Figure 33: In-phase and quad-phase signals of one of the receiving antenna

Page 47: Thesis Report Chamath

41

5 Results The experiment was carried out as described in the chapter 4. The received baseband

signals for each antenna is sampled and digitized from the ADC. Those digitized data

samples are then transferred to the host computer. Those received data samples are

then analyzed using a Matlab code for decomposing. The used matlab code is

available at Appendix B.

First the peak of the raised cosine function is detected from the received baseband

data.

The received constellation for each antenna is shown in the following figure.

Figure 34: Peak Detection for further signal processing

Page 48: Thesis Report Chamath

42

Figure 35: Received Constellations at the two antennas

Then the channel is estimated using the 8 training symbols. Once the channel is

estimated, the receiver side processing is carried out to decompose the original

symbols. The transmitted and estimated symbols are shown in figure 36.

Page 49: Thesis Report Chamath

43

Figure 36Constellation Diagram for the two sub streams

Finally the Bit Error Rate (BER) is calculated by comparing the bit streams. BER

performances are evaluated with different SNR levels. Figure 37 shows a comparison

of BER performances with SNR for three scenarios.

Beam Space (BS) MIMO Simulation

Traditional MIMO Transmission

OTA MIMO Transmission with 3-Element ESPAR

Page 50: Thesis Report Chamath

44

Figure 37: BER Comparison for the three scenarios

Page 51: Thesis Report Chamath

45

6 Conclusion and Future Work

Firstly the traditional 2x2 MIMO transmission was verified using the MIMO

testbed. Thereafter, 2x2 MIMO transmission with antenna switching was tested and

verified with two omni directional antennas placed at the transmitter. Finally a three-

element ESPAR antenna is used at the transmitter and the 2x2 MIMO transmission

was validated. The BER performance of over-the-air MIMO transmission using the

three-element ESPAR antenna is similar to the traditional 2x2 MIMO and Beam

Space MIMO simulation results. So that it can be concluded that, MIMO transmission

with three-element ESPAR antenna placed at the transmitter and traditional two

elements ULA placed at the receiver is possible. This can make a good impact on the

future mobile transmission.

The immediate next step would be to make the testbed work for real-time data

processing. So that the received constellation, BER, and other characteristics can be

observed dynamically. The testbed can even be used to verify other phenomenon such

as transmitter diversity and aerial modulation. Due to the high capacity available for

each channel, the testbed can be used for experiments with multimedia

communication as well.

Page 52: Thesis Report Chamath

46

REFERENCES

1. A Novel Approach to MIMO Transmission Using a Single RF Front End ,

Antonis Kalis, Athanasios G Kanatas and Constantinos B Papadias, IEEE Journal on Selected Areas in Communications, Vol 26, Issue. 6, August 2008, Digital Object Identifier 10.1109/JSAC.2008.080813

2. MIMO Transmission and Reception Techniques Using Three-Element ESPAR

Antennas Osama N. Alrabadi, Constantinos B. Papadias, Antonis Kalis, Nicola Marchetti, and Ramjee Prasad. IEEE COMMUNICATIONS LETTERS, VOL. 13, NO. 4, APRIL 2009. Digital Object Identifier 10.1109/LCOMM.2009.082190

3. Foschini, G.J “Layered space-time architecture for wireless communication in

a fading environment when using multiple antennas”, Bell Lab. Tech. J.,vol. 1,N. 2, pp. 41–59, 1996

4. Spatial multiplexing by decomposing the far-field of a compact ESPAR

antenna, Osama N Alrabadi, Antonis Kalis, Constantinos B Papadias and Athanasios G Kanatas, 19th IEEE International symposium on Personal, Indoor and Mobile Radio Communications, 2008, 1-5 pg, Digital Object Identifier 10.1109/PIMRC.2008.4699495

5. The OMNIBUS User’s Manual was prepared by the technical staff of Innovative

Integration on October 30, 2002.

6. I/O Board Interface Control Document, Gigabit Wireless September 29, 1999

7. The SBC6x Development Package Manual was prepared by the technical staff of Innovative Integration, April 3, 2002

8. IP Engine Hardware Reference Manual, Bright Star Engineering November 1998

9. Readpow Code composer studio project, at

C:\Windows\Desktop\work_ait\work_deliver_tx\readpow\

10. Build_Option_for_Code_Composer_Studio.pdf found at C:\chamath\documents\selected documents\

11. Transmitter Interface Doc 1.pdf found at C:\Chamath\Documents\Selected

Docs\MMDS_radios\ MMDS\Transmitter\

12. DSO_Tx_Unit_interconnect_2006_10_14.pdf found at C:\Chamath\Documents\Selected Docs\MMDS_radios\ MMDS\Transmitter\

13. Receiver Interface Doc 1.pdf found at

C:\Chamath\Documents\Selected Docs\MMDS_radios\ MMDS\Receiver\

14. DSO_Rx_Unit_interconnect_2006_10_14.pdf found at C:\Chamath\Documents\Selected Docs\MMDS_radios\ MMDS\Receiver\

Page 53: Thesis Report Chamath

47

APPENDIX A

The function used at the transmitter DSP to generate the required baseband and the

switching signals. /* func: uFsOverN * * This utility function will fill an SCOMPLEX buffer with a * sinusoid that has a specified amplitude and period. The * sinusoid will have c periods in n samples (ie, a c/n f-sub-s * sinusoid). * * The value c/n is constrained in the following ways: * c must be < n, and n must be >= 4. * * Also, siglen is required to be an even multiple of n. * * use: * flag=uFsOverN(c,n,sigamp,sigbuf,siglen); * args: * input: c = number of periods in n samples * n = number of samples for an exact number of periods * sigamp = maximum amplitude to apply * sigbuf = buffer to place result * siglen = length of sigbuf // Chamath: ( = PROCSIZE = 2048) * output: sigbuf = this buffer is filled to siglen length. * flag = true if aok, false if error encountered. * */ int uFsOverN (int c, int n, int sigamp, SCOMPLEX* restrict sigbuf, int siglen) { register double angle; register double delta; register int i, b, training, trainingBy2, sampsPERsymb, flag; register SCOMPLEX *sigp; int rcosin[253] = {0,0,0,0,0,1,1,2,2,3,3,4,5,6,7,8,10,11,12,13,15,16,17,18,19,20,21,22,23,23,23,23,23,22,22,21,19,17,15,13,10,7,4,0,-4,-9,-13,-19,-24,-30,-36,-42,-48,-54,-61,-67,-74,-81,-87,-93,-99,-105,-110,-116,-120,-124,-127,-130,-132,-133,-134,-133,-131,-129,-125,-120,-114,-106,-98,-88,-77,-64,-50,-35,-18,0,19,40,62,85,110,135,162,189,218,247,278,309,340,372,405,437,470,503,536,568,600,632,663,694,723,752,779,806,831,855,877,897,916,933,949,962,974,983,990,996,999,1000,999,996,990,983,974,962,949,933,916,897,877,855,831,806,779,752,723,694,663,632,600,568,536,503,470,437,405,372,340,309,278,247,218,189,162,135,110,85,62,40,19,0,-18,-35,-50,-64,-77,-88,-98,-106,-114,-120,-125,-129,-131,-133,-134,-133,-132,-130,-127,-124,-120,-116,-110,-105,-99,-93,-87,-81,-74,-67,-61,-54,-48,-42,-36,-30,-24,-19,-13,-9,-4,0,4,7,10,13,15,17,19,21,22,22,23,23,23,23,23,22,21,20,19,18,17,16,15,13,12,11,10,8,7,6,5,4,3,3,2,2,1,1,0,0,0}; int bits[180] = {1,1,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,1,0,0,1,0,0,0,0,0,1,1,1,0,0,0,1,0,0,0,1,1,1,0,1,0,1,0,1,0,0,1,1,0,1,1,0,0,0,0,1,1,1,1,1,0,1,1,0,1,1,1,1,1,0,0,0,0,1,0,0,0,0,0,0,1,0,0,1,1,0,0,0,0,1,0,1,1,0,0,0,0,0,1,0,0,1,0,1,1,0,1,0,0,1,1,1,0,0,1,1,0,0,1,0,1,1,1,0,0,0,1,0,1,0,1,0,1,1,1,1,0

Page 54: Thesis Report Chamath

48

,1,0,0,1,1,0,1,1,1,1,1,1,0,0,1,0,0,0,1,1,1,0,0,1,0,0,1,0,1,1,1,0,1,1,1,1,1,0,0,1,0,0}; if ((c >= n) || (n < 4) || (n > siglen)) return(false); sigp=sigbuf; i = siglen / n; if ((i * n) != siglen) return(false); sampsPERsymb = (double)n / (double)c; delta = 253/sampsPERsymb; trainingBy2 = sampsPERsymb*4; training = trainingBy2*2; angle = 0.0; b = 0; flag = 0; for (i=0; i<siglen; i++) { if (sigamp==1) { // for antenna one (for DAC channels, J1 and J2) // training sequence if (i<training) { sigp->re = 8*rcosin[shortRound(angle)]; sigp->im = 0; } else { // transmit the first bits stream always (from the only available antenna) if (bits[b]==1) { sigp->re = 6*rcosin[shortRound(angle)]; sigp->im = 0; } else { sigp->re = -6*rcosin[shortRound(angle)]; sigp->im = 0;//-6*rcosin[shortRound(angle)]; } } } else { // for antenna two (for DAC channels, J3 and J4) if (i<trainingBy2) { sigp->re = 7500; sigp->im = 0; } else if (i<training) { sigp->re = -1*7500; sigp->im = 0; }

Page 55: Thesis Report Chamath

49

// beam switching starts for actual data bits else { if ((bits[179-b]^bits[b])==0) { sigp->re = 7500; sigp->im = 0; } else { sigp->re = -1*7500; sigp->im = 0; } } } angle += delta; if (angle > 255) { angle = 0.0; b++; flag ^= 1; } if (b>179) b=0; sigp++; } return(true); } /*end*/

Page 56: Thesis Report Chamath

50

APPENDIX B

MATLAB Code used for receiver side processing. close all A = 1700; % Amplitude of the tx raised cosine T = 2048; % No of samples in a snapshot at the receiver C = 410; % No of symbols in the snapshot, (the freq of the baseband rcosine signal in 500 kHz) L = 8; % L training symbols, so actual no of data symbols is C-L B = 283; % The index of the breaking point. (1-410) check for boundary conditions near 410. (if it is < 400 safe anyway) % Txed random bits pool bits = [1,1,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,1,0,0,1,0,0,0,0,0,1,1,1,0,0,0,1,0,0,0,1,1,1,0,1,0,1,0,1,0,0,1,1,0,1,1,0,0,0,0,1,1,1,1,1,0,1,1,0,1,1,1,1,1,0,0,0,0,1,0,0,0,0,0,0,1,0,0,1,1,0,0,0,0,1,0,1,1,0,0,0,0,0,1,0,0,1,0,1,1,0,1,0,0,1,1,1,0,0,1,1,0,0,1,0,1,1,1,0,0,0,1,0,1,0,1,0,1,1,1,1,0,1,0,0,1,1,0,1,1,1,1,1,1,0,0,1,0,0,0,1,1,1,0,0,1,0,0,1,0,1,1,1,0,1,1,1,1,1,0,0,1,0,0]; sample1 = zeros(1,T); sample2 = zeros(1,T); rxSymbol1 = zeros(1,C); rxSymbol2 = zeros(1,C); % Sampling the received signals for k=1:C q = k*5-4; if (q>1410) q=q-2; end sample1(q) = snap11(q); sample2(q) = snap22(q); rxSymbol1(k) = snap11(q); rxSymbol2(k) = snap22(q); end figure(1) subplot(2,1,1) plot(real(snap11)) hold on, stem(real(sample1),'r') title('Real Part of Rx Signal 1 and the Sampled Point for 500kHz') subplot(2,1,2) plot(imag(snap11)) hold on, stem(imag(sample1),'r') title('Imag Part of Rx Signal 1 and the Sampled Point for 500kHz')

Page 57: Thesis Report Chamath

51

figure(2) subplot(2,1,1) plot(real(snap22)) hold on, stem(real(sample2),'r') title('Real Part of Rx Signal 1 and the Sampled Point for 500kHz') subplot(2,1,2) plot(imag(snap22)) hold on, stem(imag(sample2),'r') title('Imag Part of Rx Signal 1 and the Sampled Point for 500kHz') figure(3) subplot(2,1,1) stem(abs(rxSymbol1)) title('Received Signal 1 Strengh') subplot(2,1,2) stem(abs(rxSymbol2)) title('Received Signal 2 Strengh') %% Channel Estimation % H = Y.X#((X.X#)^-1) % Constructing the training vector X = zeros(2,L); for ii=1:L/2 X(1,ii) = A;%+sqrt(-1)*A; X(2,ii) = 0; end for ii=(L/2+1):L X(1,ii) = 0; X(2,ii) = A;%+sqrt(-1)*A; end Xhash = (X'); % Constructing the received vector for training sequence Y = zeros(2,L); % Rx Antenna 1 Y(1,1:L) = rxSymbol1(B:B+L-1); % Rx Antenna 2 Y(2,1:L) = rxSymbol2(B:B+L-1); H = zeros(2,2); H = Y*Xhash*inv(X*Xhash); % Construct the received data matrix rxData = zeros(2,C-L); rxData(1,1:B-1) = rxSymbol1(1:B-1); rxData(1,B:C-L) = rxSymbol1(B+L:C); rxData(2,1:B-1) = rxSymbol2(1:B-1); rxData(2,B:C-L) = rxSymbol2(B+L:C);

Page 58: Thesis Report Chamath

52

figure(4) subplot(2,1,1), plot(real(rxData(1,:)), imag(rxData(1,:)),'*'), hold on, plot(A,0,'r*'), hold on, plot(-A, 0, 'r*') axis([-2*A, 2*A, -2*A, 2*A]) title('Constellation Diagram for Rxed signal at Rx1') subplot(2,1,2), plot(real(rxData(2,:)), imag(rxData(2,:)),'*'), hold on, plot(A,0,'r*'), hold on, plot(-A, 0, 'r*') axis([-2*A, 2*A, -2*A, 2*A]) title('Constellation Diagram for Rxed signal at Rx2') % Received SNR Calculation nn=1; oo=1; for mm=1:C-L, if (imag(rxData(1,mm))>100 && real(rxData(1,mm)>90)) snrDATA(1,nn) = rxData(1,mm); nn = nn+1; elseif (imag(rxData(1,mm))<100 && real(rxData(1,mm)<150)) snrDATA(2,oo) = rxData(1,mm); oo = oo+1; end end snr = zeros(1,4); S(1) = abs(mean(snrDATA(1,:)))^2; V(1) = var(snrDATA(1,:)); snr(1) = 10*log10(S(1)/V(1)); S(2) = abs(mean(snrDATA(2,:)))^2; V(2) = var(snrDATA(2,:)); snr(2) = 10*log10(S(2)/V(2)); % snr nn=1; oo=1; for mm=1:C-L, if (rxData(2,mm)>-225) snrDATA(1,nn) = rxData(2,mm); nn = nn+1; else snrDATA(2,oo) = rxData(2,mm); oo = oo+1; end end S(1) = abs(mean(snrDATA(2,:)))^2; V(1) = var(snrDATA(2,:)); snr(3) = 10*log10(S(1)/V(1)); S(2) = abs(mean(snrDATA(2,:)))^2; V(2) = var(snrDATA(2,:)); snr(4) = 10*log10(S(2)/V(2)); snr %% Decoding the two sub streams % Calculating W

Page 59: Thesis Report Chamath

53

% W = inv(H$) where, Hnote = (1/sqrt(2))*[H(1,1)+H(1,2) H(1,1)-H(1,2); H(2,1)+H(2,2) H(2,1)-H(2,2)] Hnote = (1/sqrt(2))*[H(1,1)+H(1,2) H(1,1)-H(1,2); H(2,1)+H(2,2) H(2,1)-H(2,2)]; W = inv(Hnote); Yhat = zeros(2,C-L); Yhat = W*rxData; figure(5) subplot(2,1,1), plot(real(Yhat(1,:)), imag(Yhat(1,:)),'*'), hold on, plot(A,0,'r*'), hold on, plot(-A, 0, 'r*') axis([-2*A, 2*A, -2*A, 2*A]) title('Constellation Diagram for Substream 1') subplot(2,1,2), plot(real(Yhat(2,:)), imag(Yhat(2,:)),'*'), hold on, plot(A,0,'r*'), hold on, plot(-A, 0, 'r*') axis([-2*A, 2*A, -2*A, 2*A]) title('Constellation Diagram for Substream 2') % Detection decSymb1 = zeros(1,C-L); decSymb2 = zeros(1,C-L); decBits1 = zeros(1,C-L); decBits2 = zeros(1,C-L); % for Substream 1 for k=1:C-L if (real(Yhat(1,k))>0) decSymb1(k) = A; decBits1(k) = 1; else decSymb1(k) = -A; decBits1(k) = 0; end end % for Substream 2 for k=1:C-L if (real(Yhat(2,k))>0) decSymb2(k) = A; decBits2(k) = 1; else decSymb2(k) = -A; decBits2(k) = 0; end end %% Compare with the original tx bits % Constructing the Tx Bits vector for BER Calculation txBits = zeros(2,C-L); % for Substream 1 txBits(1,1:52) = bits(129:180); txBits(1,53:232) = bits(1:180);

Page 60: Thesis Report Chamath

54

txBits(1,233:282) = bits(1:50); txBits(1,283:404) = bits(7:128); % for Substream 2 txBits(2,1:52) = flipud(bits(1:52)')'; txBits(2,53:232) = flipud(bits(1:180)')'; txBits(2,233:282) = flipud(bits(131:180)')'; txBits(2,283:404) = flipud(bits(53:174)')'; modTxBits = (txBits*2-1)*A; figure(6), subplot(2,1,1), stem(modTxBits(1,:)), hold on, stem(decSymb1,'r*') title('Zero Forcing 2x2 MIMO - Txed and Decoded symbols for sub streams 1') subplot(2,1,2), stem(modTxBits(2,:)), hold on, stem(decSymb2,'*r') title('Zero Forcing 2x2 MIMO - Txed and Decoded symbols for sub streams 2') % BER Calculation BER1 = sum(abs(txBits(1,:)-decBits1))/C BER2 = sum(abs(txBits(2,:)-decBits2))/C BER_Tot = (BER1+BER2)/2 SNR = mean(snr)