Page 116 - Integrated Wireless Propagation Models

94 C h a p t e r T w o

is acceptable to extrapolate to as little as 6 dB of loss. The other two curves shown rep­
resent extremes of clear terrain and very rough terrain, so they provide insight into the
variability that can be expected for any given value of normalized clearance.

2.18 On-Body ModeP3
The on-body model is a new concept. When the terminal handset is used, the body size,
shape, and posture, as well as the orientation and distance of the antenna from the
body of the user will become a part of the system. It has been called the body-area net­
work (BAN).33 This BAN will affect the reception of the mobile signal and need to be
included in the prediction of the overall path loss of the mobile radio propagation. We
include it in this chapter because it is an area network. Also, it affects the performance
of all cellular networks, from macrocell to in building (picocell). There are two models.

2 . 18.1 Model 1
The received signal may be represented as the combination of a constant LOS signal
and a Rayleigh-distributed time-varying component. The spatial subchannel linking
the ith receive element and the jth transmit element in an (N x M) MIMO channel matrix
H can be represented as

(2.18.1.1)

where K is the Rician factor, <Jl; is the phase of the jth transmit element to ithsubchannel
j
in the constant component, and z .. (t) is the correlated NLOS component. The phase <p
q
q
is randomly distributed over [0, 2n]. The received power term p, can be modeled for a
given transmitter-receiver separation as
0
p,(d) = p, (d ) - 10n log(:) + X sha d (d)
o
where X sha d (d) is the lognormally distributed shadowing term. The first term in
Eq. (2.1 8 . 1 . 1 ) corresponds to the LOS component, and the second term corresponds
to NLOS component.

2 . 18.2 Model 2
In model l, the receive antenna did not encounter identical statistics owing to the varia­
tion in their position, orientation, and the amount of shadowing. In model 2, the index
I is attached to the receive power (p) ; and to the Rician K-factor in order to indicate that
they depend on the receive antenna. It follows that the hi (t) of the MIMO channel
j
matrix in this model can be represented as
K ( p )i (p )i
h (t) = -- '- e l<i';; + -'- z (t) (2.18.2.1)
·
'
' I K + l K +l ' I
I I
The two models are used to predict the on-body antenna orientation and position. For
each on-body channel, the transmitting array was placed at the belt position at the left
side of the body, while the receiving array was placed at the right side. The receiving
array that was placed on the head is called belt-head channel, and that placed at the
chest is called the belt-chest channel. Model l has been shown to well represent the belt­
chest channel, while model 2 provides a more reliable representation of the belt-head