SoRaX-Page 1
Software-Defined Radio using Xilinx
(SoRaX)
Functional Requirements List and Performance Specifications
By:
Anton Rodriguez
&
Mike Mensinger
Project Advisors:
Dr. In Soo Ahn
&
Dr. Yufeng Lu
December 31, 2009
SoRaX-Page 2
Introduction
Software Defined Radios (SDR) are highly configurable hardware platforms that
provide the technology needed to realize the rapidly expanding third (and future)
generation digital wireless communication infrastructure. The concept for Software
Defined Radio (SDR) is not new. However, its capabilities and applications are ever
expanding. Digital communication systems adopt more sophisticated coding and
modulation technologies to meet ever-increasing demand of audio, video, and data
services. The SDR is a solution for rapid integration of various emerging wireless
standards and technologies.
The objective of this project is to design a communication radio system on the
FPGA board. The main focus will then lie on the carrier synchronization, and symbol
phase ambiguity correction based on the recovered signal.
Project Goals
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Gain an in-depth understanding about the FPGA implementation of carrier
synchronization.
Create a test signal of known hard-coded values (preamble) to estimate the
channel state.
Regenerate the carrier and symbol timing to decode the transmitted digital
data.
Achieve fast acquisition of carrier synchronization and symbol timing through
efficient Xilinx programming.
Construct a working Simulink model.
Implement the Simulink model on the FPGA board.
If time permits, implement different modulation schemes (i.e. QAM).
Block Diagram
The high level flowchart is shown in Figure 1. The input to the communication
system includes a stream of data, for example, sound, image, or computer-generated data.
The quadrature phase-shift keying (QPSK) will be used to modulate the data stream.
This modulated data stream is then transmitted wirelessly, using two antennas linking the
transmitter and receiver on the SignalWave Virtex II FPGA board. The channel will
introduce noise and distortions due to the multi-paths.
The QPSK constellation is shown in Figure 2. Due to channel imperfections,
phase error is introduced in the received signal causing an incorrect representation of the
transmitted data. In order to recover the data, a coherent detection is required and is
achieved through a digital phase lock loop. The coherent detection requires estimation of
the instantaneous phase offset of every data point in the constellation grid with respect to
where it should be. That information is then fed to a direct digital synthesizer (DDS),
which generates coherent sine and cosine carriers that correct the phase error. Upon
completing the carrier synchronization and symbol timing, the data can be decoded.
QPSK demodulation will be subject to phase ambiguity. This can be resolved by
transmitting a preamble sequence or differentially encoding the data. The overall system
will be implemented and verified using Simulink. Then it will be downloaded to an
FPGA board using Xilinx tools.
SoRaX-Page 3
Corrected
Signal
Noise
Starting
signal
Recover
carrier
signal
Send through
channel
Figure 1
Phase
detector
Phase error
correction
Direct
Synthesizer
High Level Flowchart
I (n) jQ(n)
Figure 2
(n)
I(n) jQ (n)
QPSK Constellation [1]
SoRaX-Page 4
Functional Requirements & Specifications
The success of this project lies heavily on the construction of an accurate and
implementable Simulink model. In order to do so, there are several different functional
requirements and specifications to follow:
The model shall operate with a system clock of 50 MHz.
There shall be an explicit sampling period of 1/8 throughout the Simulink model.
The sampling frequency for the model should be 12.5 MHz which is of the
system clock. (This provides a convenient clock signal).
We shall use a DDS compiler 4.0 to simulate our Voltage Controlled Oscillator
(VCO) for the Phase-Locked Loop.
The frequency offset provided should be no larger than 1 kHz.
The system will have a minimum data rate of 160 kbps.
We will attempt to achieve data encoding for both QPSK and 16-QAM types.
Also, we desire that phase-locking shall occur in fewer than 100 samples. This can be
accomplished by implementing proper gain values for the proportional controller in the
loop filter design.
Equipment List:
The heart of this project revolves around the development of the simulink model.
Thus very little equipment will be used, except for the Xilinx ISE 9.2 software, and the
Xilinx Virtex-II FPGA board. We are also implementing Windows remote desktop
feature to access Dr. Lus computer to run Xilinx ISE 9.2 from off campus.
Proposed Spring Semester Schedule:
Weeks
Tasks
(Winter Break)
Compile a functional Simulink model
Design Loop Filter
1-2
Load model onto the Virtex 4
2-5
Develop training sequence
5-9
Develop an algorithm to correct phase ambiguity of QPSK symbols
10 - 13
Implement 16 QAM modulation scheme
13 15
Prepare Final Report and Oral Presentation
The workload will be distributed as evenly as possible. Anton will focus on the website
design and layout as we achieve our project goals. Mike and Anton will collaborate over
the winter break to successfully compile a working phase-locked loop and
communication system. From there, the Mike will work on developing the symbol phase
ambiguity solution. Anton will focus on hardware implementation.
SoRaX-Page 5
Bibliography
[1]
Chris Dick, Fred Harris, and Michael Rice, FPGA Implementation of Carrier
Synchronization for QAM Receivers, Journal of VLSI Signal Processing, Copyright 2004
Kluwer Academic Publishers, Netherlands.
[2]
Stephens, Donald R. Phase-locked loops for wireless communications digital and analog
implementation. Boston: Kluwer Academic, 1998. Print.
