Page 34 - Integrated Wireless Propagation Models
P. 34

12    C h a p t e r   O n e


               and
                                                                                 (1.6.4.2)
               where B5 is the bandwidth of the transmitted signal, T 5  is the symbol interval (e.g., recip­
               rocal bandwidth), and L1, and B are the delay spread and the bandwidth of the fading
                                         H
               channel, respectively.
               1.6.5  Signal Fading Caused by Time-Delay Spread: Frequency-Selective Fading
               Basically, based on multipath time-delay spread, the fading could be divided into two
               groups: flat fading and frequency-selective fading. Frequency-selective fading is when
               the coherence bandwidth of the channel is larger than the bandwidth of the signal.
               Coherence bandwidth is defined as within such a bandwidth if the signal is frequency
               insensitive with a linear phase, as mentioned previously. If the coherence bandwidth of
               the channel is smaller than the bandwidth of the signal, it is called flat fading. In this
               case, the same magnitude fading would happen to all signals with different frequency
               components.
                  Frequency-selective fading is called such when two signals with different frequen­
               cies received, after having gone through a radio channel, have different fading character­
               istics. The condition of creating frequency-selective fading is different from that of flat
               fading. When the bandwidth of a received signal is greater than the bandwidth of a radio
               channel in which a constant gain and linear phase response are preserved, then the
               received signal is spread in time and is distorted; thus, its fading characteristics change
               and differ from flat fading. The distorted signal causes the time dispersion of its trans­
               mitted symbols and introduces the lSI. To avoid lSI, the symbol rate has to be calculated
                                                                               1
               so that the symbol interval is longer than the delay spread, as shown in Fig.  . 6.3.2. ( a).
                                                                                    1
               The characteristics of a frequency-selective fading channel is depicted in Fig.  . 6.5. .
                                                                                   1
                                                                               1
                                                                   o
                  For frequency-selective fading, when the symbol interval  f   the transmitted signal
               is greater than the time-delay spread, the spectrum S(f) of the signal has a bandwidth
                                                                           >
               B,, which is greater than the coherence bandwidth Be of the channel, B5  B e
                  Frequency-selective fading channels are wideband channels since B5 > Be In sum­
               mary, a transmitted signal is suffering in a frequency selective fading if
                                                                                 (1.6.5.1)
                              _ s __:__: (t) �

                                  _       h(t)      r(t)
                                      t_  _  _  _    _J
                     s(t)                 h(t)               r(t)
                 SL

                    0  T  s   t
                     S(f)





               FIGURE 1.6.5.1  Characteristics of frequency-selective fading channel.
   29   30   31   32   33   34   35   36   37   38   39