Transcript Document 7330956

CHANNEL MODEL for
INFOSTATIONS

Can this be the model for outdoors? 
Andrej Domazetovic,
WINLAB – February, 23
OBJECTIVE
Assuming that the channel is Ricean and using the
measurements by Feuerstein, Rappaport et. al. in San
Francisco (2-ray model) try to develop the channel model
proposal described as the behavior of Ricean K-factor with
respect to transmitter-receiver distance.
INITIAL ASSUMPTIONS

Low transmitter antenna heights (3, 4 and 5m)

Receiver antenna height 1.7m

Clear line of sight path - no shadowing

Carrier frequency 5.1 GHz

Channel bandwidth 100 MHz

Omnidirectional antennas

No mobility (yet)
OUTLINE

Brief overview of standard 2-ray propagation model

Brief overview of Propagation over the earth

Closer look into propagation issues

Modified model

Link to Ricean K-factor

Real antenna pattern

Conclusions/Questions
Standard 2-ray propagation model
1
  
Pr d   Pt 
 Gt Gr
L
 4d 
2
Friis free space equation:
EIRP E 2
Pd 

4d 2 R fs
Relation between power
and electric field:
2
E
Pr d   Pd Ae 
Ae
120
Where: EIRP - effective isotropic radiated power, E - magnitude of
radiating portion of electric field in the far field, Rfs - free space intrinsic
impedance and Ae - antenna effective aperture
Source:
[] Rappaport - Wireless Communications
Standard 2-ray propagation model
The electric field at receiver:
ETOT d , t  
ETOT  ELOS  EGR
E0 d 0
Ed
  d 
  d   
cos  c  t     (1) 0 0 cos  c  t   
d
c 
d 
c 
 
 
assuming: large distance from the
transmitter, Taylor series approximations,
perfect ground reflection...
Source:
[] Rappaport - Wireless Communications
ht2 hr2
Pr  Pt Gt Gr 4
d
Standard 2-ray propagation model
In measurements performed in San Francisco, it
was shown that 2-ray model is fairly good model
for microcellular urban environment
It was also shown that the path loss within first
Fresnel zone clearance is purely due to spherical
spreading of the wave front:
decreases as d-2 and not d-4
(10m being the minimum T-R distance)
Source:
[] Feuerstein, Rappaport et. al. - Path loss, Delay spread and Outage models as
Functions of Antenna Height for Microcellular System Design - TVEH, Aug, 1994
Standard 2-ray propagation model
Source:
[] Feuerstein, Rappaport et. al. - Path loss, Delay spread and Outage models as
Functions of Antenna Height for Microcellular System Design - TVEH, Aug, 1994
Standard 2-ray propagation model
Fresnel zone clearance
df 
1


2

2 2



   
2      
2 2
2
2
2
4
  ht  hr
  ht  hr
Source:
[] Feuerstein, Rappaport et. al. - Path loss, Delay spread and Outage models as
Functions of Antenna Height for Microcellular System Design - TVEH, Aug, 1994
Propagation over a plane earth
Propagation over smooth, conducting, flat earth
Bullington:
2
  
j
j
Pr  Pt 
G
G
1

R
e

(
1

R
)
A
e

....
 t r
 4d 
2
Where:
first term - direct wave
second term - reflected wave
third term - surface wave
rest - induction field and ground secondary effects
 - phase difference between reflected and direct paths
Source:
[] W.C. Jakes - Microwave Mobile Communications
ASSUMTIONS
1
  
Pr d   Pt 
 Gt Gr
L
 4d 
2
Friis free space equation:
• The formula is a valid predictor for Pr for d which are in the
far-field of the transmitting antenna - Fraunhofer region i.e.
when inductive and electrostatic fields become negligible
and only radiation field remains
df=2D2/ , df>>D and df>>
• For fc = 5.1GHz and the antenna size D = 10cm
df=33.9cm , df>>10cm and df>>5.9cm
• If D (largest linear dimension of antenna) and fc increase,
so does df - attention must be paid
Source:
[] Rappaport - Wireless Communications
ASSUMTIONS
First Fresnel zone distance:
Antenna height:
fd:
3m
70.47m
4m
118.29m
5m
179.6m
for fc=5.1GHz
Mobile height:1.7m
Since wavelength=5.9cm, the Bullington equation
also holds (surface wave can be neglected)
Source:
[] Feuerstein, Rappaport et. al. - Path loss, Delay spread and Outage models as
Functions of Antenna Height for Microcellular System Design - TVEH, Aug, 1994
[] W.C. Jakes - Microwave Mobile Communications
Ricean K-factor
ETOT d , t  
K
E0d0
Ed
cosct   R 0 0 cosct   
d
d 
Power of specular component
Power of scattered component
 d  
K 




R
d
d


2
d   d 2  ht  hr 2
d   d 2  ht  hr 2
Source:
[] Rappaport - Wireless Communications
[] Steele - Mobile Radio Communications
Propagation Mechanisms
Ei
Ei
Er
i

r
t

Et
E-field in plane of incidence
Vertical polarization
Source:
[] Rappaport - Wireless Communications
Er
i
r
t
Et
E-field normal to plane of incidence
Horizontal polarization
Propagation Mechanisms
Reflection coefficient (Fresnel) depends on material properties,
frequency, incident angle…
It is often related to relative permittivity value:






j
0 r
(for lossy dielectric) - some energy absorbed
2f
Type of surface
 (S/m)

If material is good
Poor ground
0.001
4
conductor (f</r0)
Average ground
0.005
15
- not sensitive to f
Good ground
0.02
25
Sea water
5
81
For lossy dielectrics:
Fresh water
0.01
81
- 0, r - const. with f
Brick
0.01
4.44
but  may be
Limestone
0.028
7.51
sensitive
Glass at 10 GHz
0.005
4
Source:
[] Rappaport - Wireless Communications
[] W.C. Jakes - Microwave Mobile Communications
Propagation Mechanisms
From Maxwell’s equations and Snell’s Law:
R|| 
R 
  r sin  i   r  cos2  i
 r sn i   r  cos2  i
sin  i   r  cos2  i
sin  i   r  cos2  i
Er  REi
When the first medium is free space and 1  2
Source:
[] Rappaport - Wireless Communications
Reflection coefficient
Ricean K-factor
Ricean K-factor
Reflection coefficient
Reflection coefficient
Ricean K-factor
Ricean K-factor
Ricean K-factor
Ricean K-factor
Real antenna issues
Ricean K-factor - antenna
Ricean K-factor - antenna
Close scatters – practical issue
Assuming 100MHz bandwidth  200Msamples/second 
1.5m path distance in order to detect another path wave
Some hints that look promising
Source:
[] Bolcskei et. al, “Fixed Broadband Wireless Access: State of the Art, Challenges, and
Future Directions“, IEEE Communication magazine, Jan 2001.
Conclusions/Questions
1.
2.
2.
What do you think IMW or JFAI?
What to pursuit?
- If this idea holds, how to prove it?
- If not, should COSTs/ITUs/etc. be investigated better
and picked one of those models?
If the channel is really that good  why OFDM?
- Simplicity for Downlink (no PAPR headache,
implementable on Winlab hardware)
- DS-CDMA (no near-far, fully orthogonal code set,
multiple access…)