Maximum likelihood symbol timing estimator

Details Maximum likelihood symbol timing estimator Maximum likelihood symbol timing estimator

1999-10-08

2002-01-22

Ghebretinsae, Temesghen (Department: 2631)

Pulse or digital communications

Receivers

Particular pulse demodulator or detector

C375S355000, C714S794000

Reexamination Certificate

active

06341147

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a digital communication, and in particular to a maximum likelihood symbol timing recovering unit for a symbol timing recovering process of a received signal.
2. Description of the Background Art
FIG. 1
is a block diagram illustrating a digital communication system of a known base bandwidth. As shown therein, the conventional digital communication system includes a transmitter
110
having a coder
111
and a transmission filter
122
, a symbol decision unit
123
, a decoder
124
and a timing recovering unit
125
.
The operation of the conventional digital communication system will be explained.
When a binary bit BITS is inputted into the transmitter
110
, the binary bit BITS is mapped in a certain alphabet by a coder
111
and is changed to a symbol a
m
. The symbol a
m
is converted into an analog signal s(t) by the transmission filter
112
having an impact response g(t) and is transmitted to the communication channel
130
.
At this time, the communication channel
130
has a response characteristic b(t) in the case of a wired channel and becomes a receiving signal r(t) mixed with a summing noise n(t) by a thermal noise generated in the transmitter and receiver with respect to the signal s(t) transmitted to the receiver
120
by a wire or wireless method.
In addition, in the receiver
120
which receives the receiving signal r(t), the sampling of the symbol period is performed by the sampler
122
with respect to a signal which passes through the receiving filter
121
having the impact response f(t) like the impact response g(t) of the transmission filter
112
as a transmission characteristic.
Therefore, when the symbol decision unit
123
decides the At sampled data, the measured symbol value is obtained, and this symbol value is recovered into a binary bit by the decoder
124
.
At this time, the sampler
122
samples an output signal of the receiving filter
121
by a sampling clock and transmits the data to the symbol decision unit
123
. The sampling clock is generated by the timing recovering unit
125
which receives a receiving signal r(t).
The operation of the above-described circuit block will be explained with reference to a pulse amplitude modulation(PAM) of
FIG. 2 and a
frequency spectrum of the Nyquist pulse.
In the transmitter
110
, the pulse g(t) having an amplitude decided based on the value of the transmission symbol a
m
is generated at every symbol period T, and the thusly generated pulse g(t) is duplicated and becomes a transmission pulse s(t) which is a PAM signal as shown in FIG.
2
A. The above-described operation may be implemented by the following Equation 1.
s
(
t
)
=&Sgr;a
m
g
(
t−mT
)  Equation 1
The transmission pulse s(t) has a value “0” at all symbol points except for OT′.
The receiver
120
may recover the symbol value a
m
at the transmission side in the case that the value is read at every symbol period nT by the sampler(or A/D converter)
122
.
At this time, the operation for deciding an integrated impact response p(t) with respect to a response g(t) of the transmission filter
112
, a response b(t) of the channel
130
, and a response f(t) of the receiving filter
121
so that a zero cross point may occur at every symbol period nT except for OT′ is called as a Nyquist criterion. This response p(t) may be expressed in the following Equation 2.
p
(
t
)
=g
(
t
)
*b
(
t
)
*f
(
t
)  Equation 2
As shown in the spectrum of
FIG. 2B
, “&agr;” is a roll-off factor. When this value varies to 0~1, the over frequency bandwidth is changed to 0%~100%.
In addition, in the spectrum,
“
0
~
1
2
⁢
T
”
bandwidth is called as a signal bandwidth, and the operation for carrying the transmission symbol a
m
by multiple times is called as a Nyquist ratio transmission.
At this time, the frequency component at
1
2
⁢
T
,
namely, at the bandwidth edge point includes a timing information which is important for the symbol timing recovery. If the over frequency is decreased, the bandwidth occupying width of the channel is decreased, and it is difficult to obtain the timing recovery.
Therefore, in the digital communication, the output of the demodulator must be periodically sampled at the timing of t
m
=mT+&tgr; based on the symbol rate.
Here, T represents a symbol interval, and &tgr; represents a delay time which occurs during a transfer from the transmitter to the receiver.
In order to implement a periodical sampling operation, a clock signal is required for the receiver. The process for extracting the clock signal from the receiver is called as a symbol timing recovery.
Various methods for the symbol timing recovery are known.
First, in the spectrum recovering method(spectral line method), a band-pass filter is tuned at a bandwidth edge portion of the signal spectrum with respect to the receiving signal which passed through the linear or non-linear apparatus for thereby extracting a timing information.
For the binary signal, there are a method for checking a zero cross point and a method for using a point at which an inclination at the sampling time of the receiving signal becomes a timing information.
As an important factor for a selection of the timing recovering method, there are an area of the over frequency bandwidth and a level number of the signal. The case that random symbol value affects the timing information by a larger PAM signal which exceeds a certain signal level is called as a self-noise.
The circuit for the current timing recovery is directed to implementing a digital circuit.
Therefore, it is possible to enhance a reliability of the circuit operation by implementing a digital circuit for the sampling clock occurrence because that the signal process is digitally performed.
The digital implementation of the circuit for a timing recovery is obtained by a data interpolation method, a combination with a channel equalizer, etc. For example, a maximum likelihood symbol timing estimator is known.
The maximum likelihood (ML) symbol timing recovering unit uses a recovering technique for forming a likelihood function with respect to the receiving signal and estimating a timing phase for maximizing the likelihood function. The construction is different based on a DA(Data-Aided) ML, a DD(Decision-directed) ML, and a NDD (Non-Decision-directed) ML modes.
FIG. 3
is a block diagram illustrating a maximum likelihood symbol timing recovering unit using a DA(Data-Aided) ML mode as an example of the conventional art. As shown therein, the maximum likelihood symbol timing recovering unit includes a sampler
201
for sampling the matched and filtered signal at a certain period and outputting a digital signal q
k
{circumflex over ((&tgr;))} to a channel equalizer, a differential unit
202
for differentiating the matched and filtered signal, a sampler
203
for sampling the output signal of the differential unit
202
and outputting a digital signals
ⅆ
ⅆ
τ
⁢
q
k
⁢
(
τ
)
^
,
a multiplier
204
for multiplying an accurate symbol a
m
transmitted in a preamble format and an output signal
ⅆ
ⅆ
τ
⁢
q
k
⁢
(
τ
)
^
of the sampler
203
, a k-term accumulator
205
for accumulating the outputs of the multiplier
204
at the m-symbol interval of the observing period, and a voltage adjusting oscillator VCO
206
for outputting an oscillation frequency to the samplers
201
and
203
using an output value of the accumulator
205
as an adjusting voltage.
The operation of the first example of the conventional art will be explained with reference to FIG.
3
.
In the conventional art for the DA-ML mode, the matched and filtered signal is inputted into the sampler
201
and the differential unit
202
, and an output of the differential unit
202
is inputted into the sampler
203
. The-output of the sampler
203
and a symbol received in a preamble format are multiplied by the multiplier
204
.
The resultant values accumulated from the multip

Affiliated with

Also associated with

Rating

Maximum likelihood symbol timing estimator does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Maximum likelihood symbol timing estimator, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Maximum likelihood symbol timing estimator will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2841809