<?xml version="1.0" encoding="utf-8" standalone="no"?>
<!DOCTYPE document PUBLIC "-//CNX//DTD CNXML 0.5 plus MathML//EN" "http://cnx.rice.edu/cnxml/0.5/DTD/cnxml_mathml.dtd">
<document xmlns="http://cnx.rice.edu/cnxml" xmlns:m="http://www.w3.org/1998/Math/MathML" xmlns:md="http://cnx.rice.edu/mdml/0.4" xmlns:bib="http://bibtexml.sf.net/" id="m10173">

  <name>Channel Capacity</name>
  <metadata>
  <md:version>2.8</md:version>
  <md:created>2001/07/06</md:created>
  <md:revised>2005/10/26 20:30:46.927 GMT-5</md:revised>
  <md:authorlist>
      <md:author id="aaz">
      <md:firstname>Behnaam</md:firstname>
      
      <md:surname>Aazhang</md:surname>
      <md:email>aaz@ece.rice.edu</md:email>
    </md:author>
  </md:authorlist>

  <md:maintainerlist>
    <md:maintainer id="dinesh">
      <md:firstname>Dinesh</md:firstname>
      
      <md:surname>Rajan</md:surname>
      <md:email>dinesh@ece.rice.edu</md:email>
    </md:maintainer>
    <md:maintainer id="mohammad">
      <md:firstname>Mohammad</md:firstname>
      <md:othername>Jaber</md:othername>
      <md:surname>Borran</md:surname>
      <md:email>mohammad@ece.rice.edu</md:email>
    </md:maintainer>
    <md:maintainer id="rha">
      <md:firstname>Roy</md:firstname>
      
      <md:surname>Ha</md:surname>
      <md:email>rha@rice.edu</md:email>
    </md:maintainer>
    <md:maintainer id="mrshawn">
      <md:firstname>Shawn</md:firstname>
      
      <md:surname>Stewart</md:surname>
      <md:email>mrshawn@alumni.rice.edu</md:email>
    </md:maintainer>
    <md:maintainer id="aaz">
      <md:firstname>Behnaam</md:firstname>
      
      <md:surname>Aazhang</md:surname>
      <md:email>aaz@ece.rice.edu</md:email>
    </md:maintainer>
  </md:maintainerlist>
  
  <md:keywordlist>
    <md:keyword>channel</md:keyword>
    <md:keyword>channel capacity</md:keyword>
    <md:keyword>information theory</md:keyword>
    <md:keyword>transmission</md:keyword>
  </md:keywordlist>

  <md:abstract>A discussion of channels and how much information can be sent through a channel reliably.</md:abstract>
</metadata>

  <content>
    <para id="para1">
      In the previous section, we discussed information sources and
      quantified information.  We also discussed how to represent (and
      compress) information sources in binary symbols in an efficient
      manner.  In this section, we consider channels and will find out
      how much information can be sent through the channel reliably.
    </para>

    <para id="para2">
      We will first consider simple channels where the input is a
      discrete random variable and the output is also a discrete
      random variable.  These discrete channels could represent analog
      channels with modulation and demodulation and detection.
    </para>

    <figure id="fig1">
      <media type="image/png" src="Figure7-28.png"/>
    </figure>

    <para id="para3">
      Let us denote the input sequence to the channel as
      <equation id="eq01">
	<m:math>
	  <m:apply>
	    <m:eq/>
            <m:ci type="vector">X</m:ci>
            <m:vector>
              <m:ci>
                <m:msub>
                  <m:mi>X</m:mi>
                  <m:mn>1</m:mn>
                </m:msub>
              </m:ci>
              <m:ci>
                <m:msub>
                  <m:mi>X</m:mi>
                  <m:mn>2</m:mn>
                </m:msub>
              </m:ci>
              <m:ci>⋮</m:ci>
              <m:ci>
                <m:msub>
                  <m:mi>X</m:mi>
                  <m:mi>n</m:mi>
                </m:msub>
              </m:ci>
            </m:vector>
	  </m:apply>
	</m:math>
      </equation>
      where
      <m:math>
	<m:apply>
	  <m:in/>
          <m:ci>
            <m:msub>
              <m:mi>X</m:mi>
              <m:mi>i</m:mi>
            </m:msub>
          </m:ci>
          <m:ci>
            <m:mover>
              <m:mi>X</m:mi>
              <m:mo>¯</m:mo>
            </m:mover>
          </m:ci>
	</m:apply>
      </m:math>
      a discrete symbol set or input alphabet.
    </para>

    <para id="para4">
      The channel output
      <equation id="eq02">
	<m:math>
	  <m:apply>
	    <m:eq/>
            <m:ci type="vector">Y</m:ci>
            <m:vector>
              <m:ci>
                <m:msub>
                  <m:mi>Y</m:mi>
                  <m:mn>1</m:mn>
                </m:msub>
              </m:ci>
              <m:ci>
                <m:msub>
                  <m:mi>Y</m:mi>
                  <m:mn>2</m:mn>
                </m:msub>
              </m:ci>
              <m:ci>
                <m:msub>
                  <m:mi>Y</m:mi>
                  <m:mn>3</m:mn>
                </m:msub>
              </m:ci>
              <m:ci>⋮</m:ci>
              <m:ci>
                <m:msub>
                  <m:mi>Y</m:mi>
                  <m:mi>n</m:mi>
                </m:msub>
              </m:ci>
            </m:vector>
	  </m:apply>
	</m:math>
      </equation>
      where
      <m:math>
	<m:apply>
	  <m:in/>
          <m:ci>
            <m:msub>
              <m:mi>Y</m:mi>
              <m:mi>i</m:mi>
            </m:msub>
          </m:ci>
          <m:ci>
            <m:mover>
              <m:mi>Y</m:mi>
              <m:mo>¯</m:mo>
            </m:mover>
          </m:ci>
	</m:apply>
      </m:math>
      a discrete symbol set or output alphabet.
    </para>

    <para id="para5">
      The statistical properties of a channel are determined if one finds
      <m:math>
	<m:apply>
	  <m:ci type="function">
	    <m:msub>
	      <m:mi>p</m:mi>
	      <m:mrow>
		<m:mi fontweight="bold">Y</m:mi>
		<m:mo>|</m:mo>
		<m:mi fontweight="bold">X</m:mi>
	      </m:mrow>
	    </m:msub>
	  </m:ci>
	  <m:ci>
	    <m:mrow>
	      <m:mi fontweight="bold">y</m:mi>
	      <m:mo>|</m:mo>
	      <m:mi fontweight="bold">x</m:mi>
	    </m:mrow>
	  </m:ci>
	</m:apply>
      </m:math>
      for all
      <m:math>
	<m:apply>
	  <m:in/>
          <m:ci type="vector">y</m:ci>
          <m:apply>
            <m:power/>
	    <m:ci>
	      <m:mover>
		<m:mi>Y</m:mi>
		<m:mo>¯</m:mo>
	      </m:mover>
	    </m:ci>
	    <m:ci>n</m:ci>
          </m:apply>
	</m:apply>
      </m:math>
      and for all
      <m:math>
	<m:apply>
	  <m:in/>
          <m:ci type="vector">x</m:ci>
          <m:apply>
            <m:power/>
	    <m:ci>
	      <m:mover>
		<m:mi>X</m:mi>
		<m:mo>¯</m:mo>
	      </m:mover>
	    </m:ci>
	    <m:ci>n</m:ci>
          </m:apply>
	</m:apply>
      </m:math>.
      A discrete channel is called a <term>discrete memoryless channel</term>
      if
      <equation id="eq03">
	<m:math>
	  <m:apply>
	    <m:eq/>
            <m:apply>
              <m:ci type="function">
                <m:msub>
                  <m:mi>p</m:mi>
                  <m:mrow>
                    <m:mi fontweight="bold">Y</m:mi>
                    <m:mo>|</m:mo>
                    <m:mi fontweight="bold">X</m:mi>
                  </m:mrow>
                </m:msub>
              </m:ci>
              <m:ci>
                <m:mrow>
                  <m:mi fontweight="bold">y</m:mi>
                  <m:mo>|</m:mo>
                  <m:mi fontweight="bold">x</m:mi>
                </m:mrow>
              </m:ci>
            </m:apply>
            <m:apply>
              <m:product/>
	      <m:bvar>
		<m:ci>i</m:ci>
	      </m:bvar>
	      <m:lowlimit>
		<m:cn>1</m:cn>
	      </m:lowlimit>
	      <m:uplimit>
		<m:ci>n</m:ci>
	      </m:uplimit>
	      <m:apply>
		<m:ci type="function">
		  <m:msub>
		    <m:mi>p</m:mi>
		    <m:mrow>
		      <m:msub>
			<m:mi>Y</m:mi>
			<m:mi>i</m:mi>
		      </m:msub>
		      <m:mo>|</m:mo>
		      <m:msub>
			<m:mi>X</m:mi>
			<m:mi>i</m:mi>
		      </m:msub>
		    </m:mrow>
		  </m:msub>
		</m:ci>
		<m:ci>
		  <m:mrow>
		    <m:msub>
		      <m:mi>y</m:mi>
		      <m:mi>i</m:mi>
		    </m:msub>
		    <m:mo>|</m:mo>
		    <m:msub>
		      <m:mi>x</m:mi>
		      <m:mi>i</m:mi>
		    </m:msub>
		  </m:mrow>
		</m:ci>
	      </m:apply>
            </m:apply>
	  </m:apply>
	</m:math>
      </equation>
      for all
      <m:math>
	<m:apply>
	  <m:in/>
          <m:ci type="vector">y</m:ci>
          <m:apply>
            <m:power/>
	    <m:ci>
	      <m:mover>
		<m:mi>Y</m:mi>
		<m:mo>¯</m:mo>
	      </m:mover>
	    </m:ci>
	    <m:ci>n</m:ci>
          </m:apply>
	</m:apply>
      </m:math>
      and for all
      <m:math>
	<m:apply>
	  <m:in/>
          <m:ci type="vector">x</m:ci>
          <m:apply>
            <m:power/>
	    <m:ci>
	      <m:mover>
		<m:mi>X</m:mi>
		<m:mo>¯</m:mo>
	      </m:mover>
	    </m:ci>
	    <m:ci>n</m:ci>
          </m:apply>
	</m:apply>
      </m:math>.
    </para>

    <example id="example1">

      <para id="para6">
	A binary symmetric channel (BSC) is a discrete memoryless channel with
	binary input and binary output and
	<m:math mode="inline">
	  <m:apply>
	    <m:eq/>
            <m:apply>
              <m:ci type="function">
                <m:msub>
                  <m:mi>p</m:mi>
                  <m:mrow>
                    <m:mi>Y</m:mi>
                    <m:mo>|</m:mo>
                    <m:mi>X</m:mi>
                  </m:mrow>
                </m:msub>
              </m:ci>
              <m:ci>
                <m:mrow>
                  <m:mi>y=0</m:mi>
                  <m:mo>|</m:mo>
                  <m:mi>x=1</m:mi>
                </m:mrow>
              </m:ci>
            </m:apply>
            <m:apply>
              <m:ci type="function">
                <m:msub>
                  <m:mi>p</m:mi>
                  <m:mrow>
                    <m:mi>Y</m:mi>
                    <m:mo>|</m:mo>
                    <m:mi>X</m:mi>
                  </m:mrow>
                </m:msub>
              </m:ci>
              <m:ci>
                <m:mrow>
                  <m:mi>y=1</m:mi>
                  <m:mo>|</m:mo>
                  <m:mi>x=0</m:mi>
                </m:mrow>
              </m:ci>
            </m:apply>
	  </m:apply>
	</m:math>.
	As an example, a white Gaussian channel with antipodal signaling and 
	matched filter receiver has probability of error of
	<m:math>
	  <m:apply>
	    <m:ci type="function">Q</m:ci>
	    <m:apply>
	      <m:root/>
              <m:apply>
                <m:divide/>
		<m:apply>
		  <m:times/>
		  <m:cn>2</m:cn>
		  <m:ci>
		    <m:msub>
		      <m:mi>E</m:mi>
		      <m:mi>s</m:mi>
		    </m:msub>
		  </m:ci>
		</m:apply>
		<m:ci>
		  <m:msub>
		    <m:mi>N</m:mi>
		    <m:mn>0</m:mn>
		  </m:msub>
		</m:ci>
              </m:apply>
	    </m:apply>
	  </m:apply>
	</m:math>.
	Since the error is symmetric with respect to the transmitted bit, then
	<equation id="eq04">
	  <m:math>
	    <m:apply>
	      <m:eq/>
              <m:apply>
                <m:ci type="function">
                  <m:msub>
                    <m:mi>p</m:mi>
                    <m:mrow>
                      <m:mi>Y</m:mi>
                      <m:mo>|</m:mo>
                      <m:mi>X</m:mi>
                    </m:mrow>
                  </m:msub>
                </m:ci>
                <m:ci>
                  <m:mrow>
                    <m:mn>0</m:mn>
                    <m:mo>|</m:mo>
                    <m:mn>1</m:mn>
                  </m:mrow>
                </m:ci>
              </m:apply>
              <m:apply>
                <m:ci type="function">
                  <m:msub>
                    <m:mi>p</m:mi>
                    <m:mrow>
                      <m:mi>Y</m:mi>
                      <m:mo>|</m:mo>
                      <m:mi>X</m:mi>
                    </m:mrow>
                  </m:msub>
                </m:ci>
                <m:ci>
                  <m:mrow>
                    <m:mn>1</m:mn>
                    <m:mo>|</m:mo>
                    <m:mn>0</m:mn>
                  </m:mrow>
                </m:ci>
              </m:apply>
              <m:apply>
                <m:ci type="function">Q</m:ci>
                <m:apply>
                  <m:root/>
		  <m:apply>
		    <m:divide/>
		    <m:apply>
		      <m:times/>
		      <m:cn>2</m:cn>
		      <m:ci>
			<m:msub>
			  <m:mi>E</m:mi>
			  <m:mi>s</m:mi>
			</m:msub>
		      </m:ci>
		    </m:apply>
		    <m:ci>
		      <m:msub>
			<m:mi>N</m:mi>
			<m:mn>0</m:mn>
		      </m:msub>
		    </m:ci>
		  </m:apply>
                </m:apply>
              </m:apply>
              <m:ci>ε</m:ci>
	    </m:apply>
	  </m:math>
	</equation>
      </para>

      <figure id="fig2">
	<media type="image/png" src="Figure7-30.png"/>
      </figure>

    </example>

    <para id="para7">
      It is interesting to note that every time a BSC is used one bit
      is sent across the channel with probability of error of
      <m:math><m:ci>ε</m:ci></m:math>.  The question is how much
      information or how many bits can be sent per channel use,
      reliably.  Before we consider the above question a few
      definitions are essential.  These are discussed in <cnxn document="m10178" strength="7">mutual information</cnxn>.
    </para>

  </content>
</document>
