Implementing Adaptive Modulation in a Software-Defined Cognitive Radio - PowerPoint PPT Presentation

1 / 26
About This Presentation
Title:

Implementing Adaptive Modulation in a Software-Defined Cognitive Radio

Description:

An SDR is a flexible radio where programmable hardware is controlled by software. SDRs can tune to any frequency band or change to any modulation type. ... – PowerPoint PPT presentation

Number of Views:138
Avg rating:3.0/5.0
Slides: 27
Provided by: stude647
Category:

less

Transcript and Presenter's Notes

Title: Implementing Adaptive Modulation in a Software-Defined Cognitive Radio


1
Implementing Adaptive Modulation in a
Software-Defined Cognitive Radio
  • Brandon Bilinski
  • Computer Engineering Senior,
  • Clemson University

2
What is a Software-Defined Radio?
  • An SDR is a flexible radio where programmable
    hardware is controlled by software.
  • SDRs can tune to any frequency band or change to
    any modulation type.
  • I am employing adaptive modulation.

3
Previous Implementations
  • With the frequency bands getting more and more
    crowded, a solution is needed.
  • Software-Defined radios allow for the advent of
    cognitive radios.
  • Many cognitive radios sense busy channels and
    change to free channels
  • What if there are no free channels available to
    change between?

4
Why adapt modulations?
  • If a channel is noisy it is more beneficial to
    change modulations rather than adjust power.
  • The following modulation types had constant power
    and bandwidth.

5
First Steps of Adaptive Modulation
  • Construct a library of various modulation types
    in the FPGA chip.
  • My radio contains 32-ary Orthogonal, BPSK, QPSK,
    and 16-QAM
  • All modulation types designed in Simulink using
    the Xilinx FPGA blockset.
  • BPSK Binary Phase Shift Keying
  • QPSK Quadrature Phase Shift Keying
  • QAM Quadrature Amplitude Modulation

6
32-ary Orthogonal Modulation
  • This is the most robust modulation scheme used in
    my radio.
  • This scheme uses 32 chips to represent 5 bits of
    data, so it is also the slowest.
  • Each signal is orthogonal as a result of the
    orthogonality of the Hadamard matrix.

7
What are Hadamard Matrices?
  • A Hadamard matrix is a matrix of the form 2(n-1)
    x 2(n-1) where every row is orthogonal.
  • If you multiply the Hadamard matrix by the
    transpose of any row, you will get n in the
    appropriate row and 0s elsewhere.

8
32-ary Orthogonal Implementation
9
32-ary Orthogonal Modulation Cont.
  • Transmit two words 00000 and 00001
  • Corresponding Hadamard Rows
  • 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
    1 1 1 1 1 1 1
  • 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1
    1 -1 1 -1 1 -1 1 -1 1 -1 1 -1

10
BPSK Modulation
  • The transmitter modulates only one bit at a time.
  • A 0 is mapped to -1cosine while a 1 is
    mapped to 1cosine.
  • The BPSK constellation is mapped on the complex
    plane to the right

11
BPSK Implementation
12
BPSK Modulation Cont.
  • Transmit the following bits 11110101

13
QPSK Modulation
  • QPSK is less robust than BPSK, but it can send
    twice the bits in one transmission.
  • QPSK has four possible words 00, 01, 10,
    11 that are modulated using a combination of
    sine and cosine.
  • The constellation mapping is once
  • again shown on the right

14
QPSK Implementation
15
QPSK Modulation Cont.
  • Transmit the following bits 11011000

16
16-QAM Modulation
  • d3- 0 cosine is negative
  • 1 cosine is positive
  • d2- 0 1x cosine
  • 1 2x cosine
  • d1- 0 sine is negative
  • 1 sine is positive
  • d0- 0 1x sine
  • 1 2x sine
  • Unlike BPSK and QPSK, 16-QAM involves changes in
    phase and amplitude.
  • In this way, each word can be 4 bits long, which
    allows you to send twice the bits that are
    possible with QPSK.
  • This is the toughest scheme that I implement to
    demodulate due to close proximity of points in
    the constellation.

17
16-QAM Implementation
18
How do you get environment stats?
  • How do you know what the CENR is in the
    environment in order to adapt your modulation
    types?

19
Adapting 32-Orthogonal Modulation
  • As previously stated, multiplying an nxn Hadamard
    matrix by a transpose of one row of the matrix
    theoretically results in an nx1 matrix of n and
    all 0s.
  • To estimate the signal to noise ratio we can use
    the following equation 1 (second_largest/max_p
    roduct)
  • The following example is for 4-ary
    modulation

1 - (.7/3.9) .821
20
32-ary Ratio Statistic
  • When the ratio gets above .38, then it is
    desirable to switch to BPSK.
  • With additional orthogonal schemes included, a
    ratio lower than .38 would signify a switch to
    16-ary orthogonal.

21
BPSK Distance Statistic
  • If distance is greater than .6, switch to 32-ary
    orthogonal.
  • If distance is under .37, switch to QPSK.

22
QPSK Distance Statistic
  • If distance is greater than .37, switch to BPSK
  • If distance is less than .19, switch to 16-QAM

23
16-QAM Distance Statistic
  • If the distance is greater than .19, switch down
    to QPSK.

24
Future Plans
  • Finish 16-QAM demodulator and add timing and
    phase synchronization in all demodulators.
  • Implement modulation library in FPGA
  • Enable modulation scheme switching by programming
    the DSP chip in the radio.
  • Move the radio through different noise
    environments to ensure modulation switching
    occurs where expected.
  • Add Forward Error Correcting codes to improve
    robustness of system.

25
Acknowledgements
  • Faculty Advisor Dr. Michael Pursley
  • Graduate Student Advisors Joel Simoneau
    Tommy Royster
  • Dr. Noneaker, Dr. Xu and Josh Lawrence
  • Dr. Russell, Dr. Hubing, Dr. Baum, Dr. Hubbard
    and Dr. Fishman.

26
Questions?
  • Brandon Bilinski
  • Computer Engineering Senior,
  • Clemson University
Write a Comment
User Comments (0)
About PowerShow.com