Pulse or digital communications – Spread spectrum – Direct sequence
Reexamination Certificate
1998-12-18
2001-08-21
Le, Amanda T. (Department: 2634)
Pulse or digital communications
Spread spectrum
Direct sequence
C375S344000
Reexamination Certificate
active
06278725
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention is directed to an automatic frequency control (AFC) loop multipath combiner for a RAKE receiver, and more particularly, to an AFC combiner which removes the Doppler frequency offsets from all the RAKE fingers.
2. Discussion of the Prior Art
Typically in terrestrial communication, a receiver receives a transmitted signal which has traveled through a direct path and indirect paths. Propagation through the indirect paths, referred to as multipath propagation, results from the transmitted signal being reflected and refracted by surrounding terrain. The multipath signals traveling through the indirect paths undergo frequency and time offsets.
To exploit the energy in the multiple components of multipath propagation of a transmitted signal, a RAKE receiver is used which has multiple parallel demodulators for receiving different multipath components of the transmitted signal. Each multipath component demodulator is called a “finger” of the RAKE receiver. The RAKE receiver identifies and acquires the multiple components of multipath propagation with the aid of a pilot signal. As well known to those skilled in the art, a RAKE receiver collects and combines the energy from the distinct paths.
Typically a RAKE receiver uses an automatic frequency control (AFC) loop for initial frequency acquisition and Doppler frequency adjustment of the received signal which has been disturbed by noise and multipath. The Doppler frequency offsets of the disturbed or faded signals are often unknown. A balanced discrete quadri-correlator or cross-product automatic frequency control (CP-AFC) loop structure is used to obtain the unknown frequency offsets.
In the CP-AFC loop, the unknown frequency offset is obtained through differentiation as will be described in connection with equations (5) and (7). To derive the CP-AFC, consider an optimal phase estimator structure
10
shown in FIG.
1
. As shown in
FIG. 1
, a received signal y(t) is provided to two mixers
12
,
14
, which respectively receive locally oscillating signals
16
,
18
having a 90° phase difference. These two signals are provided from a local oscillator, such as a voltage controlled oscillator (VCO)
20
, where one of the signals passes through a 90° phase shifter
22
. The outputs of the two mixers
12
,
14
are provided to two integrators (or lowpass filters)
32
,
34
, respectively.
The received signal y(t) is expressed by equation (1):
y
(
t
)={square root over (2)}
A
sin[{circumflex over (&ohgr;)}
t+
(&ohgr;−{circumflex over (&ohgr;)})
t+&thgr;]+n
(
t
) (1)
where:
{circumflex over (&ohgr;)} is the frequency of the local oscillator
20
;
&ohgr; is the frequency of the received signal y(t);
&thgr; is an unknown constant carrier phase; and
n(t) is noise.
Combining the frequency difference term (&ohgr;−{circumflex over (&ohgr;)}) with the unknown constant carrier phase e into one unknown time variant phase &thgr;(t), equation (1) is rewritten as equation (2):
y
(
t
)={square root over (2)}
A
sin[{circumflex over (&ohgr;)}
t+&thgr;
(
t
)]+
n
(
t
) (2)
In the noise free case, the outputs of the integrators or lowpass filters 32, 34 are given by equations (3) and (4):
y
c
(
t
)=
A
cos&thgr;(
t
) (3)
and
y
s
(
t
)=
A
sin&thgr;(
t
) (4)
As seen from equations (3) and (4), the purpose of the lowpass filters (LPFs)
32
,
34
is to suppress the double frequency term resulting from the product of y(t) with the locally oscillating signals
16
,
18
. The difference (&ohgr;−{circumflex over (&ohgr;)}) between the received signal frequency ({circumflex over (&ohgr;)}) and the local oscillator frequency (&ohgr;) is given by equation (5):
ω
-
ω
^
=
ⅆ
ⅆ
t
θ
(
t
)
=
ⅆ
ⅆ
t
[
a
tan
(
y
s
(
t
)
y
c
(
t
)
)
]
(
5
)
where &thgr;(t) is the output of the phase estimator
10
shown in FIG.
1
.
Using the identity shown in equation (6):
ⅆ
ⅆ
t
[
atan
f
(
t
)
]
=
f
′
(
t
)
1
+
f
2
(
t
)
(
6
)
equation (5) can be expressed as equation (7):
ω
-
ω
^
=
ⅆ
ⅆ
t
θ
(
t
)
=
ⅆ
ⅆ
t
[
atan
(
y
s
(
t
)
y
c
(
t
)
)
]
=
y
c
′
(
t
)
y
s
(
t
)
-
y
s
′
(
t
)
y
c
(
t
)
(
7
)
Because of the differentiator in equation (5), a CP-AFC structure realization of equation (5) is also known as the differentiator AFC. In the discrete domain, the differentiator AFC structure can be easily derived by replacing the derivative dy(t)/dt at time t=n&Dgr;T by the expression shown in equation (8):
(
ⅆ
ⅆ
t
y
(
t
)
&RightBracketingBar;
)
t
=
n
ΔT
=
y
(
n
Δ
T
)
-
y
(
n
Δ
T
-
Δ
T
)
Δ
T
=
y
(
n
)
-
y
(
n
-
1
)
Δ
T
(
8
)
where &Dgr;T represents the sampling period. The analog differentiator dy(t)/dt has a system transfer function H(s)=s, whereas the discrete system has the transfer function given by equation (9) which can be deduced from equation (8):
H
(
z
)
=
1
-
z
-
1
Δ
T
(
9
)
Consequently, the mapping between the analog and the discrete domains is governed by equation (10):
s
=
1
-
z
-
1
Δ
T
(
10
)
Note that the mapping in equation (10) is only suitable for lowpass and bandpass filters having relatively small resonant frequencies.
In order to derive the structure for the discrete-time differentiator AFC, equation (8) is substituted into equation (7) to yield equation (11):
y′
s
(
t
)
y
c
(
t
)−
y′
c
(
t
)
y
s
(
t
)≈[1
/&Dgr;T][y
s
(
n−
1)
y
c
(
n
)−
y
c
(
n−
1)
y
s
(
n
)] (11)
The realization of equation (11) is a discrete differentiator AFC (or CP-AFC) loop structure depicted in
FIG. 2
, which will be described later.
The relation in equation (11) can be further expressed as equation (12):
y′
s
(
t
)
y
c
(
t
)−
y′
c
(
t
)
y
s
(
t
)≈[1
/&Dgr;T]
sin(&Dgr;
T&Dgr;&ohgr;
) (12)
where &Dgr;&ohgr;=&ohgr;−{circumflex over (&ohgr;)}.
When operating in the linear region (i.e., theoretically &Dgr;&ohgr;&Dgr;T<<1)), the error signal D(&ohgr;−{circumflex over (&ohgr;)}) is directly proportional to the difference between the received signal and local oscillator frequencies. The relationship is no longer linear when &Dgr;&ohgr; becomes large.
FIG. 2
shows a typical cross-product (CP) AFC
100
having a frequency discriminator (FD)
110
which is a realization of the expression shown in equation (11). Similar to the phase estimator
10
of
FIG. 1
, the received signal y(t) is provided to the two mixers
12
,
14
for down conversion using locally oscillated signals
16
,
18
which are 90° apart and provided from the VCO
20
, where one signal is phase shifted by a 90° phase shifter. For clarity, the 90° phase shifter
22
, shown in
FIG. 1
, is omitted from FIG.
2
.
The down-converted signals pass through respective analog-to-digital A/D converters
120
,
125
, integrators
130
,
135
, and dump or low-pass filters
140
,
145
. The output signals of the dump filters
140
,
145
are indicated as y
s
and y
c
, respectively, which are the input signals to the frequency discriminator
110
. The first signal y
s
passes though a first delay element
150
and a first mixer or multiplier
155
. Similarly, the second signal y
c
passes though a second delay element
160
and a second mixer or multiplier
165
. The first signal y
s
is also provided to the second mixer
165
, while the second signal y
c
is also provided to the first mixer
155
.
The outputs of the two mixers
155
,
165
are provided to a combining circuit
Kamgar Farbod
Rouphael Antoine J.
Dicran Halajian
Le Amanda T.
Philips Electronics North America Corporation
LandOfFree
Automatic frequency control loop multipath combiner for a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Automatic frequency control loop multipath combiner for a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automatic frequency control loop multipath combiner for a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2466759