Wireless Networking

1
Wireless Networking
2
Class organization

• Class web page
• http//www.cs.fsu.edu/zzhang/CIS5930_Spring_2009.
  htm
• Academic honor code
• Programs you submitted must be your own work
• While discussions of class materials and
  assignments are allowed, copying of solutions is
  strictly prohibited

3
Class Communication

• This class will use class web site to post news,
  changes, and updates. So please check the class
  website regularly
• Please also make sure that you check your emails
  on the account on your University record

4
Wireless Networking

• Wireless networks are everywhere cellular phone
  networks, wireless LANs, Bluetooth
• This class is intended to cover a very wide
  spectrum of topics related to wireless
  networking, including the physical layer, the MAC
  layer, and the network layer.
• After taking this class, you should be able to
• Understand basic wireless communication theory
  (BPSK, CDMA, OFDM, RS code, etc)
• Learn to implement wireless communication
  transmitters/receivers with GNU Software Defined
  Radio
• Understand the design of wireless networks
  (802.11 network, cellular phone network, wireless
  sensor network, etc)

5
How this course is designed

• This class is designed for CS majors who are
  interested in wireless networks.
• There are two groups of people studying wireless
  networks.
• The signal processing approach. Typically
  focusing on signal processing and deriving the
  channel capacity. Focusing on physical layer and
  cellular phone networks.
• The computer science approach. Typically treat
  the physical layer as a black box and focusing on
  MAC layer and network layer. Wireless LANs,
  wireless sensor networks.

6
How this course is designed

• Actually, both of these typical approaches are
  limited.
• Wireless medium and techniques are very different
  from wired medium and techniques.
• In wired medium, like Ethernet, what you send is
  likely what will be received. Limited noise,
  limited interference, large bandwidth.
• In wireless medium, what you send may be very
  different from what will be received. Substantial
  noise, substantial interference, limited
  bandwidth.
• Well, this is why it is so interesting!

7
How this course is designed

• Only focusing on physical layer wont be
  sufficient for computer networks where traffic is
  random.
• Simply treating it as a black box will lead to
  suboptimal solutions.
• What we need is a cross-layer approach.
• This will require you to understand everything
  from physical layer to network layer at least.
• This is why this course will take a
  non-traditional approach and will cover physical
  layer, MAC layer, and network layer, all in
  details.

8
How this course is designed

• This is a very challenging task (after all, this
  is a graduate level course! )
• Can we achieve this?
• To computer science majors, the MAC layer and the
  network layer are more familiar. The challenge is
  the physical layer.
• We will have to spend significant amount of time
  on the physical layer.
• I will teach the physical layer in a
  non-conventional way.
• Books about the physical layer are usually
  written by the signal processing people, and may
  be alien to the computer science majors.
• Our goal will be to understand the physical
  layer. We dont have to do things such as
  deriving channel capacity.
• We will also learn to implement the physical
  layer with GNU Software Defined Radio.

9
Materials that will be used in the class

• Fundamentals of Wireless Communication, by
  David Tse and Pramod Viswanath, downloadable
  at http//www.eecs.berkeley.edu/dtse/book.html
• GNU Software Defined Radio tutorial,
  by Dawei Shen, downloadable at http//www.nd.edu/
  jnl/sdr/docs/tutorials/
•  
• Other useful resources (not required)
• Computer Networks,'' by Andrew S. Tanenbaum,
  Prentice Hall, 4th edition, 2003
• Principles of Wireless Networks A Unified
  Approach, by Kaveh Pahlavan and Prashant Krishna
  murthy,  Prentice Hall, 1st edition, 2002.

10
Projects

• Physical layer projects will be implemented by
  GNU SDR (C and Python).
• Upper layer projects will be implemented by
  C/C.
• Projects will be in teams with maximum 3 persons.
• To work with GNU SDR, at least one of your team
  members should have access to a Linux machine as
  root.

11
Physical Layer

• Physical layer design goal send out bits as fast
  as possible with acceptable low error ratio
• Some simple schemes
• There is a wire between A and B. If A wants to
  send a bit 1, he connects the wire to the
  positive end of a battery. Otherwise he
  disconnects it from the battery.
• Or A can hold a radio, if 1, he sends at
  frequency f1 and if 0 he sends at frequency f2.
• Or there is an optical fiber between A and B and
  if 1 A lit up a light and if 0 A does
  nothing.

12
Wireless communications

• The fundamental fact is that if the sender sends
  a sine wave, the receiver will receive a sine
  wave at the same frequency. But with
• A different phase
• A new amplitude
• How do you design communication schemes based on
  that?

13
BPSK

• The simplest transmission scheme is BPSK, which
  is also widely used.
• Convert your information bits to a -1,1 square
  waveform. Let it be I(t). Multiply I(t) with
  cos(2 \pi ft), and send out.
• This is the basic idea. But to make it work, more
  work has to be done.

14
Bandwidth

• Bandwidth in wireless medium is limited.
• Check http//www.ntia.doc.gov/osmhome/allochrt.pdf
  
• 802.11g network, each channel has 22MHz
  bandwidth. Channel 1 is centered at 2.412GHz.
• The 2.412GHz is the f in the cos(2 \pi ft).
• What is bandwidth?

15
Bandwidth

• In very simple terms, it is how fast your signal
  can change.
• If you have an unlimited bandwidth, your signal
  can change infinitely fast.
• The frequency spectrum is shared, so you can use
  only a part of it.
• Means that your signal cannot change infinitely
  fast.
• I(t) changes infinitely fast at the transition
  points from -1 to 1 or from 1 to -1.

16
Baseband signals

• Roughly speaking, a signal can be represented as
  the summation of a series of sine waves (Fourier
  Transformation ).
• So you have to pass the bit stream to a Low Pass
  Filter to filter out the high frequency
  components.
• The filtered signal is called the baseband
  signal.

17
Low Pass Filter

• The simplest LPF is a RC circuit.
• In our projects, we will use RRC filter, or, the
  Root Raised Cosine filter. We will discuss it in
  details.
• So suppose you feed the bit stream to the RRC
  filter and gets the baseband signal, denoted as
  I(t).

18
The Transmitted Signal

• So what you actually send is I(t)cos(2\pi ft),
  where I(t) is band-limited to BHz.
• In 802.11g network, each channel has 22MHz
  bandwidth. What should B be?
• Assume you are given a bandwidth 2BHz centered at
  fHz. It means that all components higher than
  (fB)Hz and all frequency lower than (f-B)Hz will
  be (or should be) cut-off.

19
Receiver

• The receiver receives r(t) AI(t) cos(2 \pi ft
  \phi). Here, just for now, assume the receiver
  somehow magically finds the value of \phi and
  set it to be 0 (we will talk about this shortly).
  So he multiplies r(t) with cos(2 \pi ft), and
  gets AI(t)/2 AI(t)cos(4 \pi ft)/2.
• You apply the LPF again to get rid of the
  high-frequency components (AI(t)cos(4 \pi ft)/2),
  and what is left will be proportional to I(t).

20
A simplified wireless communication scheme
21
Complex representations

• You have to get used to representing the received
  signal as complex numbers.
• That is, r(t) Re(t) jIm(t).
• Why?
• Because if you will multiply the r(t) with both
  cos(2\pi ft) and sin(2\pi ft), and both will be
  sent to a LPF. The one corresponding to cos(2\pi
  ft) is regarded as the real part and the one
  corresponding to sin(2\pi ft) is regarded as the
  imaginary part.
• You will see why this is convenient later.

22
Here is an issue

• How to recover the original bits?
• I(t) is no longer the simple, clean square
  waveform.
• Solution sample I(t) at time instants and if the
  samples are taken correctly, you can get the
  correct bits.
• We will talk about this in details.

23
Here is another issue

• The oscillators are not perfect!
• The sender and receiver use local oscillators to
  generate cos(2\pi ft). There will be a slight
  difference between the sender and the receiver
  frequency.
• So, the receiver has to track the frequency
  difference, as well as the phase difference. Will
  be discussed later.

24
More issues

• Multi-path.
• In wireless communications, signals travel
  multiple paths to reach the destination. If you
  send I(t), the receiver will receive \sum a_i
  I(t-\tau_i).
• Solution
• 1. Ignore it. Valid if the symbol rate is low.
• 2. Use equalization.
• 3. OFDM.
• Will be discussed in details.

25
GNU Software Defined Radio

• http//www.gnu.org/software/gnuradio/
• A tool ideal for computer science majors to
  practice with wireless communications.
• You write signal processing blocks in C, and
  connect the signal processing blocks with Python.
• In Project 1, You will be asked to write signal
  processing blocks.

26
The code for a simple BPSK transmitter and
receiver

• http//www.cs.fsu.edu/zzhang/CIS5930_Spring_2009_
  files/OSMR_chest_snd.py