Spatio-Temporal Channel Modelling

References:
W.R. Braun and U. Dersch, "A physical mobile radio channel model", IEEE Trans. Vehicular Technology, vol. 40, 1991, pp. 472-482. U. Dersch and R.J. Ruegg, "Simulations of the time and frequency selective outdoor mobile radio channel", IEEE Trans. Vehicular Technology, vol. 42, 1993, pp. 338-344.

Contents
Introduction Some channel parameters Codit modelling for uniform linear array Matlab demonstration

Introduction
Deterministic (geometric) modelling
- e.g. specific micro/pico cell environment - example: local area sample of the Codit model - can be used as basis for stochastic modelling

Stochastic modelling
- often based on large measurement campaign - e.g. the so-called Metra MIMO model in 3GPP - example: typical correlation matrices for certain environments

"Channel model should be based on physical propagation environment"

Antenna Array Principle

Geometrical modelling of 2D radio channel

Parameters:

number of Tx/Rx antennas & their spacing/ geometry

number of scatterers/multipath components and their

- azimuth directions - power - delay - fading characteristics

In case of large antenna separation/ low correlation parallel 1D channel models can be applied --> Rxx MIMO Terminal Rx
x

radio channel Ry
y

Base station

Some radio channel parameters

Coherence time (vs. time-selective fading): tc = 1/ Bd = 1/Doppler spread
e.g. 2GHz, 15m/s: f d,max = v/ =15/0.15 Hz = 100 Hz => Bd = 200 Hz => tc = 5 ms

Coherence bandwidth (vs. frequency selective fading): fc = 1/ Tm = 1/Delay spread

e.g. typical urban: f c = 1/(2 s) = 500 kHz

Delay spread Delay Tm

Some radio channel parameters

Angular spread,
scatterers

Antenna array

= azimuth angle
Measurements: Angular power spectrum often Laplacian distributed

Angular spread of received paths is defined as the square root of the sample second central moment

= S

N n =1

( n E{ })2 P( n )
P ,total
where

E[] is the mean or expected value of angles of arrival i , P(i) is the power of signal coming from scatterer i and P,total is the total power

Angular spread is the standard deviation of angles of arrival weighted with path energies.

Some radio channel parameters

"Coherence beamwidth" of an antenna array BWc = / D = /Antenna aperture
e.g. linear array: BW c = 48o /L (L = array length)

analogy: 1/delay spread = fc /space spread = BWc => angularly selective fading => angular diversity

Impairments of the radio channel

Signal variation caused by multiplicative and additive factors

+ White noise

Path loss Shadowing Fast fading

100s m - 10s km 10s m - 100s m /2: 1cm - 10s cm

In the following we consider only fast fading characteristics

Codit: Wide Band Channel Model

macro-cell channel model (BTS antenna well above the surroundings) short term variation -UE moving over small areas (tens of ) local area samples (WSSUS model) long term variation of mean power has not been considered

"Lee's model"
Base station antenna array
Uniform Linear Array (ULA)

Local reflector environment around UE

Angular spread r

"Urban"

 Complex field Ei from one scatterer i:

Ei (t ) = ai 0 e j ( i 0 + kvt cos i 0 ) +

N waves l =1

ail e j ( il + kvt cos il )

- ail is the partial wave amplitude for subpath l - il is a random phase term - k is the wave number (=2/) - v is the mobile terminal (UE) velocity - is the wavelength of the carrier -il denotes the incident angle between the scatterer i and the UE velocity vector - Nwaves is the number of partial waves (per scatterer)

 Time-variant channel impulse response function from multiple scatterers at different delays:
N i =1

h(t , ) =

Ei (t ) ( i )

- N is the number of scatterers - is the Dirac delta function - i time delay associated with a particular scatterer

 Spatio-temporal channel impulse response at antenna element m of a uniform linear array:

2 ( m 1) d sin

hm (t , ) = h(t , )e

- d denotes the inter-element spacing is the azimuth direction of the reflector

ULA

kd sin() 1 2 1

Delay difference => phase difference: l = d sin = 2 d sin / = k d sin k = 2 /