Pj. Cullen et al., WIDE-BAND MEASUREMENT AND ANALYSIS TECHNIQUES FOR THE MOBILE RADIO CHANNEL, IEEE transactions on vehicular technology, 42(4), 1993, pp. 589-603
The appearance of high bit rate digital radio systems would appear to
demand the existence of wide-band channel models, planning tools, and
simulation capabilities. The inclusion of these wide-band effects in p
lanning tools would, in many cases, appear to be the most obvious way
of improving upon the accuracy of existing propagation modeling techni
ques, especially in areas subject to multipath interference, for examp
le, in urban areas surrounded by mountains. Detailed channel informati
on is particularly important in the development of these more refined
propagation models. This means that the mobile radio planner needs acc
ess to the local scattering function rather than just the average sign
al strength; moreover, it is increasingly desirable that experimentali
sts have the capability to locate scatterers (measure the direction of
arrival azimuthal as well as elevation) to aid in multipath model dev
elopment. In this paper, we present an overview of various approaches
to wide-band channel data acquisition (channel sounding) and data anal
ysis for the locally QWSSUS channel. In particular, we outline how hig
h-speed digital signal processing techniques can be employed to constr
uct a nearly ideal real-time wide-band channel sounder. The digital so
under which we describe employs a TMS32050A to implement a matched fil
ter type receiver. The probe signal is a carrier BPSK modulated with a
binary maximal length sequence; the received signal is processed in r
eal time, yielding a complex time variant channel impulse response. A
dynamic range in excess of 40 dB is achieved. We also briefly discuss
techniques to use this data for stored channel simulation. Finally, by
way of pointing-out that much more information can be extracted from
channel sounding data than is generally appreciated, we briefly discus
s the use of probability theory (Bayesian techniques) in model selecti
on.