Title of Invention

A PROCESS FOR RE-TRANSMITTING SINGLE FREQUENCY SIGNALS AND A SINGLE FREQUENCY SIGNAL REPEATER

Abstract The invention discloses a process for re-transmitting single frequency signals and a single frequency signal repeater, where coupling occurs between the transmitting antenna and the receiving antenna, and where the process is of the type used in a single frequency signal repeater and comprises the steps of: [a] receiving a first radio frequency signal having a particular receiving power, [b] optionally converting said first radio frequency signal into a process signal, [c] filtering, amplification and automatically controlling the gain of said signal, [d] canceling said coupling between said transmitting antenna and said receiving antenna, [e] reconverting, as the case may be, said process signal into a second radio frequency signal, [f] amplifying the power of said second radio frequency signal, [g] output filtering, and [g] transmission.
Full Text A PROCESS FOR RE-TRANSMITTING SINGLE FREQUENCY SIGNALS
AND A SINGLE FREQUENCY SIGNAL REPEATER
DESCRIPTION
The invention relates to a process for re-transmitting single
frequency signals, of the type used in a single frequency signal repeater
and comprising the steps of: [a] receiving a first radio frequency signal
through a receiving antenna, the first radio frequency signal having a
receiving power, [b] optionally, converting the first radio frequency signal
into a process signal, [c] filtering1 the signal, [d] amplifying the signal, [e]
automatically controlling the signal gain, [f] reconversion, as the case may
be, of the process signal into a second radio frequency signal, [g]
amplifying the power of the second radio frequency signal, [h] output
filtering the second radio frequency signal, and [i] transmitting the second
radio frequency signal through a transmitting antenna, where coupling
takes place between the transmitting antenna and the receiving antenna.
The invention also relates to a single frequency signal repeater, of the type
comprising: [a] a receiving antenna, for receiving a first radio frequency
signal, having a receiving power, [b] a base unit, for converting, optionally,
the first radio frequency signal into a process signal, filtering the signal,
amplifying the signal, automatically controlling the signal gain, and
reconverting, as the case may be, the process signal into a second radio
frequency signal, [c] a power amplifier unit, [d] an output filter, and [e] a
transmitting antenna, where the transmitting antenna and the receiving
antenna may undergo coupling.
The present invention relates, therefore, to signal processing, and
the corresponding device therefor, the repeater, to be incorporated in DVB
(Digital Video Broadcasting), DAB (Digital Audio Broadcasting), GSM, etc.
repeaters that transmit in the same channel as they receive, whereby they
operate in single frequency. This invention allows these repeaters to have
high gains, so that for one same received signal level its transmitted power
is increased and, consequently, its coverage area is increased. The
ultimate aim of the system is to be able to cover the same area of service
with a smaller number of repeaters, thereby reducing costs.
The main limitation of the repeaters that operate in single frequency
resides in the fact that the reception and transmission frequency of the
repeater is the same, whereby a certain degree of coupling exists between
the transmitting and receiving antennas, that is to say, the receiving
antenna receives an echo of the transmitted signal. This can cause the
repeater to oscillate. Also, said coupling distorts the signal frequency. In
accordance with the state of the art, an effective way of avoiding this, or of
reducing it to insignificant values, is achieved by reducing the repeater's
gain. However, a consequence of this is that the area of coverage thereof is
also reduced.
An attempt has been made in the repeaters used so far to alleviate
this problem by means of, for example, the use of reduced coupling
transmitting and receiving antennas. However, this solution is expensive
and not very satisfactory, since is not possible to avoid the coupling
completely. By way of example, let us consider the case in which the
transmitted signal is reflected from an object close to the repeater (a tree, a
car, etc.). The echo caused by the reflection from said object will introduce
a coupling between antennas not contemplated at the time of designing
their radiation diagrams, whereby it will not be possible to avoid it.
It is an object of the invention to overcome these drawbacks. This
object is achieved by means of a process for re-transmitting single
frequency signals of the type first mentioned above, having a cancellation
step of said coupling between said transmitting antenna and said receiving
antenna. That is to say, conceptually, it is not a question of avoiding the
formation of the coupling or echo, which is a solution that has shown itself
to be expensive and of limited results, but rather the process is capable of
"knowing" the coupling that is occurring, and cancels it. Since the received
signal is really the sum of two components: one component is the "true"
signal it is desired to transmit, and the other component is due to the
coupling, the process according to the invention eliminates the component
due to the coupling of the received signal before transmitting it. This allows
a number of additional advantages to be attained. Thus, for example, the
process allows "a priori" unforeseen couplings to be cancelled in a simple
way, such as for example, those caused by the environment.
Preferably the first radio frequency signal is converted into a
process signal, such as, for example, into an intermediate frequency (F1)
signal, into a baseband signal, etc., but it is also possible to have the whole
process carried out on the first radio frequency signal, without any
conversion.
An additional improvement consists of carrying out an adaptive
process that permanently estimates the coupling between antennas, even
while the repeater is operative. This allows the cancellation of time-variable
coupling, such as may be generated by mobile elements in the
environment.
Preferably the cancellation step comprises a negative feedback.
Since the received signal is really the sum of two components, such as has
already been stated above, by means of the negative feedback, a signal
equal (or very similar) to the component due to the coupling is subtracted
from the received signal. Thus, the signal that is transmitted is free from the
component due to the coupling.
The proposed process is more agile and cheaper than the existing
ones, thanks to its implementation not requiring any alteration of the
radiation diagrams between antennas, but a simple filtering of the
transmitted signal that is easily reconfigurable and the cost of which is low.
Also the present invention proposes a single frequency signal
repeater of the type first mentioned above, wherein it includes a device
adapted to cancel said coupling between said transmitting antenna and
said receiving antenna.
Accordingly, the present invention provides
A method for re-transmitting single frequency signals, of the type used in a single frequency signal repeater and
comprsing the steps of.
[a] receiving a first radio frequency signal through a rece.ving antenna said first rad o frequency signal having
a receiving power, [b] converting said first rad o frequency signal into a process signal (RF/FI), [c] fiiterirg sad
process signal, [d] amplifying said process signal, [e] automatically controlling said process Signal gain (CAG),
[f] reconversion of sad process signa! into a second radio frequency signa! (F1/RF), [g] ampliying (AP) the
power of said second radio frequency signal, [h] output filtering sad second radio frequency signal H tans-
miting said second radio frequency signal through a traismiting antenra where a coupling (A(s); takes place
between said tansmiting antenna and said receivig antenna, and [i] a step of adaptive cancel ation with an
adapt ve a!gorthm of said coupling (A(s)) between said transmitting antenna and said receiv.ng anterra said
step of adapt ve cancellation comprising a negatve feedback of a 'eedback sigra! (F(s)) and said step of
adaptive cancel at on operating on digitzed signals characterized in that said step of adaptive cancelation
takes, as input sgnal, said process sigral digitized after said automatic gain control step (CAG).
The present invention also provides
A method for re-transmitting single frequency signals, of the type used in a single frequency signal repeater and
comprising the steps of:
[a] receiving a first radio frequency signal through a receiving antenna, said first radio frequency signal having
a receiving power, [b] filtering said f.rst radio frequency signal, [c] amplifying said first radio frequency signal,
[d] automatically controlling said first radio frequency signal gain (CAG), [e] amplifying (AP) the power of a
second radio frequency signal, [f] output filtering said second radio frequency signal, [g] transmitting said
second radio frequency signal through a transmitting antenna, where a coupling (A(s)) takes place between
said transmitting antenna and said receiving antenna, and [h] a step of adaptive cancellation with an adaptive
algorithm of said coupling (A(s)) between said transmitting antenna and said receiving antenna, said step of
adaptive cancellation comprising a negative feedback of a feedback signal (F(s)), said step of adaptive cancellation
operating on digitized signals, and said step of adaptive cancellation having as output signal said
second radio frequency signal, characterized In that said step of adaptive cancellation takes, as input signal,
said first radio frequency signal digitized after said automatic gain control step (CAG).
The present invention further provides
A singe frequency signal repeater, of the type comprising:
[a] a receiving antenna, for receiving a first radio frequency signal, with a receiving power, [b] a base unit (UB),
for converting said first radio frequency signal into a process signal, filtering said process signal, amplifying
said process Signal, automatically controlling said process signal gain (CAG), and reconverting said process
signal into a second radio frequency signal, [c] a power amplifier unit (AP), [d] an output filter (FS), [e] a
transmitting antenna, where said transmitting antenna and said receiving antenna may undergo coup'ing (A
(s)), and [f] an adaptive device with an adaptive algorithm for canceling said coupling (A(s)) between said
transmitting antenna and said receiving antenna, said adaptive device comprising a negative feedback of a
'eedback signal (F(s)) and said adaptive device operating on digitized signals, characterized in that said
adaptive device takes, as input signal, said process signal digitized after automatically controlling said process
signal gain (CAG).
The present invention further provides
A single frequency signal repeater, of the type comprising:
[a] a receiving antenna, for receiving a first radio frequency signal, with a receiving power, [b] a base unit (UB),
for filtering said first radio frequency signal, amplifying said first radio frequency signal, and automatically
controling said first radio frequency signal gain (CAG), [c] a power amplifier unit (AP) [d] an output filter (FS).
[e] a transmitting antenna, wrere said transmitting antenna and said receiving antenna may undergo couplirg
(A(s)), and [f] an adapt.ve device with an adaptive algorithm for canceling said coupling (A(s)) between said
transmiting antenna and said receivirg antenna, said adaptive device comprising a regative feedback of a
feedback signal (F(s)), said adapt've device operating on digitized signals and said adaptive device having as
output signal a second radio frequency signal, characterized In that said adaptive device taxes, as input
signal, said first radio frequency signal dgitized after automatically controlling said first radio frequency signal
gain (CAG), wnere said second radio frequency signal is input to said power amplifier unit (AP).
Other advantages and features of the invention will be appreciated
from the following description, wherein, without any limitative character,
preferred embodiments of the invention are related, with reference to the
accompanying drawings, in which:
Figure 1 is a generic block diagram of a conventional single
frequency signal repeater.
Figure 2 is a simplified model of a conventional single frequency
signal repeater.
Figure 3 is a block diagram of an example a conventional single
frequency signal repeater.
Figure 4 illustrates a simplified model of a conventional single
frequency signal repeater, with the transfer function expressed in complex
variable.
Figure 5 illustrates a simplified model of a single frequency signal
repeater according to the invention.
Figure 6 is a block diagram of a single frequency signal repeater
according to the invention.
Figure 7 is a block diagram of an analog delay line.
Figure 8 illustrates a time graph showing the variation of the
repeater's gain during the acquisition and follow-up phases.
Figure 9 illustrates a first alternative strategy for sampling the signal
used in the estimate of the filter response.
Figure 10 illustrates a second alternative strategy for sampling the
signal used in the estimate of the filter response.
Figure 11 illustrates a repeater with the whole intermediate
frequency step implemented digitally.
Figure 12 illustrates a third alternative strategy for sampling the
signal used in the estimate of the filter response.
Figure 13 illustrates a scheme of an analog/digital (A/D) conversion
step.
Figure 14 illustrates a scheme of a digital/analog (D/A) conversion
step.
Figure 15 illustrates an adaptive filter broken down into two blocks.
Figure 16 illustrates a block diagram of the analog implementation
of the calculation of coefficients according to the LMS algorithm.
Figure 17 illustrates an analog embodiment of the correlation loop
for obtaining the coefficients.
Figure 18 illustrates a two-cell portion equivalent to that of Figure 17
of the analog embodiment of the adaptive filter.
Figure 19 illustrates embodiment of an integrator by means of an
operational amplifier.
Figure 20 illustrates a scheme of a resistive distributor by means of
a star network.
Figure 21 illustrates a block diagram of a repeater according to the
invention with a digital embodiment of the adaptive filter.
Figure 22 illustrates the structure of the digital embodiment of the
adaptive filter.
Figure 23 illustrates the calculation of the coefficients of the
adaptive filter.
Figure 24 illustrates analog-digital conversions for the adaptive filter.
Figure 25 illustrates the detailed structure of the adaptive filter.
Figure 26 illustrates a diagram of the Data-Framer.
Figure 27 illustrates the structure of each of the filter coefficients
(tap).
Figure 28 illustrates the Table Generator.
Table 1 shows the contents of the Look-up Table (LUT).
Figure 1 shows a generic block diagram of a single frequency signal
repeater. In accordance with this diagram, the incoming radio frequency
(RF) signal SE (received by the receiving antenna), is converted into an
intermediate frequency (Fl) signal, and an intermediate frequency pass
band filter FFI filters other undesired signals received. Subsequently a
converter converts the intermediate frequency signal back into a radio
frequency signal again, in the same channel as the received radio
frequency signal. Finally the signal is amplified to the required output level
by means of an output power amplifier AP and is transmitted ST.
A single frequency signal repeater is, substantially, a filtered
amplifier, and can be modeled as shown in Figure 2. Within the
corresponding signal bandwidth, the repeater acts as an amplifier AMP,
followed by a delay cell, of value t. This delay is due to the intermediate
frequency pass band filter FFI. The receiving and transmitting antennas are
not perfectly insulated from one another, whereby coupling takes place
between them, generating a feedback of the output signal in the input
signal. Said coupling effect can be modeled as a feedback line of gain B, as
shown in Figure 2. Additionally, this coupling line also has a delay, but this
delay is much smaller than the delay t, whereby it can be ignored.
The amplitude response is not flat, but rather rippled, depending on
the product AB, the expression of which is:
Another form of expressing the product AB is as the gain margin,
which may be defined as the difference between the antenna insulation and
the repeater gain:
The key point in the operation of a single frequency signal repeater
is the compromise between a minimum gain margin, allowing a maximum
gain to be achieved, and, therefore, the maximum output power, and the
maximum authorized amplitude ripple.
Figure 3 is a block diagram of an example of a conventional single
frequency signal repeater. It comprises a base unit UB, a power amplifier
unit AP and an output filter unit FS.
The base unit UB carries out the following functions:
a)conversion of a first radio frequency signal into an intermediate
frequency signal;
b)filtering the intermediate frequency signal;
c) amplification;
d) automatic gain control;
e) conversion of the intermediate frequency signal into a second radio
frequency signal;
The power amplifier unit AP amplifies the power.
The output filter FS carries out a final filtering of the signal to
eliminate undesired out-of-band signals.
The base unit UB consists of the following components:
1. an input filter FE which eliminates the undesired signals received
out-of-band;
2. a first converter C0NV1 comprising, in turn, a means for converting
the radio frequency channel received into an intermediate
frequency, an intermediate frequency filter FFI, and the automatic
gain control CAG;
3. a linearity pre-corrector PRLIN, which is a circuit compensating for
the intrinsically non-linear behavior of the amplifier unit;
4. a second converter C0NV2 including a means for converting
intermediate frequency to the radio frequency output channel and
the output level control circuit;
5. a synthesizer SINT generating and supplying the signal from the
local oscillator to both frequency converters.
Figures 4 and 5 illustrate the basic differences between the already
existing single frequency repeaters and the proposed invention, where the
notation (s) refers to the representation of the linear systems by means of
their transfer function in terms of the complex variable s. Figure 4 illustrates
the basic block diagram of a conventional single frequency signal repeater,
where the repeater's main chain receives an input signal SE and transmits
an output signal ST. The main chain has been modeled with a response
H(s), and a gain G. The coupling between antennas has also been
indicated in dash line, modeled linearly by means of a response A(s). It is
thus observed that the signal SE entering the repeater REP through the
receiving antenna is the sum of the desired received signal S1 plus a
coupling signal. In a conventional repeater, the stability of the repeater
depends on the loop gain G-H(s)-A(s), and the only way to guarantee that it
is stable is to reduce the gain G or to reduce the coupling A(s). Figure 5
illustrates the block diagram of a repeater incorporating one embodiment of
the proposed invention. It consists of a negative feedback of the transmitted
signal ST that is processed by means of the adaptive filter FAD, expressed
as complex variable as F(s), and subsequently it cancels the echoes, that is
to say the coupling, at the repeater input. It is evident that now the
repeater's stability depends on the loop gain G-H(s)-(A(s)-F(s)), whereby to
guarantee it, it is sufficient to achieve F(s)=A(s), without necessity of
reducing G or A(s).
The proposed process is to some extent similar to the echo
cancellation systems used in communications over long distance telephone
lines. Nevertheless, the objectives are different. Indeed, the objective of the
proposed invention is to allow a repeater having a large gain at the same
time as it stays stable to be implemented and, for example in the case of a
DVB signal, the dispersion with time of the transmitted signal ST is sensibly
inferior to the duration of the cyclic prefix. On the other hand, in the echo
cancellation system used in communications over long distance telephone
lines, there is no problem of instability, and the desired objective is to
eliminate from the received signal the component of transmitted signal
introduced by the hybrid in the passage from 2 wires to 4 wires. Also, there
are notable differences in the problems originated in the implementation of
both applications.
The description of the invention requires the specification of the
following aspects of its implementation: A) architecture for inserting the
adaptive processing in the single frequency signal repeater, B) architecture
for implementing the filter for the processing of the signal, C) algorithm for
estimating the filter coefficients.
There is described hereinafter in detail one possible configuration of
the proposed invention, implemented totally with analog technology. This is
referred to hereinafter as analog configuration. Subsequently there are
briefly described some digital alternatives for the implementation thereof,
referred to as digital configurations that are distinguished from the analog
configuration in that they implement the adaptive filter digitally, but their
principles are common to the latter, whereby they should be considered to
be parts of the same invention.
Analog configuration
The block diagram of Figure 6 describes with more detail one
embodiment of a repeater according to the invention, showing in a
summarized way all the repeater steps that are relevant to the invention.
In the repeater of Figure 6, the coupling has been modeled linearly
and the repeater has an adaptive filter FAD which estimates the value of
the coupling in the frequency band occupied by the signal and cancels its
contribution to the transmitted signal ST, so that the repeater behaves as if
it did not exist. Preferably the proposed repeater works in intermediate
frequency (FI). However other alternatives may also be used operating in
baseband or in other frequencies such as is commented upon in the digital
configurations section. The repeater of Figure 6 is provided, additionally,
with a radio frequency filter FRF, and a radio frequency amplifier ARF at
the input end of the equipment.
The adaptive filter FAD takes the main chain signal after the CAG
(Automatic Gain Control). This detail is important with a view to establishing
the performance of the system, since in this case the variations with time of
the filter coefficients are due only to the variation of the coupling between
antennas, but not to variations in the repeater gain. Nevertheless, it is also
possible to arrange for the adaptive filter FAD to take the signal before the
CAG.
The adaptive filter FAD is implemented preferably by means of an
analog delay line and multipliers that weight and add the signal to the
output of each of the delay cells T (Figure 7). The number of cells and
multipliers depends on the compromise established between the complexity
of the system and the level of cancellation of the couplings that it is desired
to achieve. On the other hand, the delay introduced by each cell T should
be chosen in agreement with the value of the intermediate frequency and
the band width of the signal. Indeed, the implemented filter has a periodic
frequency response, whereby it is necessary to introduce certain
restrictions in the delay with the purpose of guaranteeing that there is
freedom to cancel the coupling between antennas in the whole frequency
band occupied by the signal.
The filter coefficients are preferably constantly adaptively estimated
while the repeater is in operation. This is because the coupling between
antennas is unknown a priori, since it depends on the configuration of the
main chain (antennas, filters and amplifiers used) and of the environment
where the repeater is located (nearby obstacles, reflectivity and distance of
the latter, etc.). Also, it is not very effective to estimate their value a priori
since the environment can change with time (movement of the foliage of
nearby trees, movement of cars or people, etc.).
One embodiment of the proposed invention estimates the filter
coefficients based on the optimization of a quadratic cost function:
minimization of the signal power at point A of Figure 6. This criterion is
based on the statistical property of non-correlation between the desired
received signal and the signal induced in the antenna by the coupling,
thanks to which it is shown that the power at point A is minimum when it
has been possible to cancel said coupling. In order to guarantee correct
operation of the criterion, it should be ensured that the main chain
introduces a delay equal to or longer than the minimum for which the
desired non-correlation is fulfilled. Optionally this non-correlation delay can
be introduced in the conversion step.
The minimization of the power in the point A is equivalent to a
criterion of minimum mean squared error that can be optimized with very
diverse adaptive algorithms. The fact that an analog implementation is used
has made it advisable to limit the algorithms applied to the simplest and
whose behavior is well documented in the literature. Preferably the Least
Mean Square {LMS) and normalized LMS (NLMS) algorithms can be used,
as well as the simplified versions thereof based on the sign function (see
Ref. [2], [6], [8], [11]). Nevertheless, the algorithm used is not the object of
this invention, and any algorithm the convergence and good behavior in
follow-up of which is guaranteed (see other alternative algorithms in the
digital configurations section) could be used.
The adaptive filter FAD is faced with two basic restrictions. The first
restriction is the need to guarantee a minimum delay in the main chain, as
previously mentioned. This delay can be introduced in said conversion
step. The second restriction resides in the fact that the coupling can cause
a signal of a level substantially lower than the desired signal in the
receiving antenna (otherwise the repeater would start to oscillate), whereby
the signal to noise ratio (SNR) for the purpose of identification of the
coupling is very low. This requires that a very slow evolution of the adaptive
system coefficients be forced with the objective of compensating the loss of
SNR with a time averaging of the signal.
Figure 5 illustrates the closed loop structure of the single frequency
repeater REP, the coupling between antennas and the adaptive filter FAD.
This structure causes the cancellation errors to be fed back to the system
and to the transmitted signal, with the possibility of causing the system to
start oscillating - it is recalled that the stability depends on the loop gain
GH(s)-(A(s)-F(s)) -. To avoid this, it is advisable to contemplate
mechanisms that assure that the system stays stable at all times. This
drawback, which does not appear in the echo cancellation systems used in
communications over long distance telephone lines, makes it advisable to
establish two phases in the operation of the proposed invention: the
acquisition phase FADQ and that of follow-up FSEG (Figure 8). The
acquisition phase FADQ is carried out only once, during the initialization or
start up of the repeater. During this phase, the repeater gain stays low
(G,n/), so that it is stable independently of the cancellation level reached by
the adaptive system. In it, the adaptive algorithm estimates the value of the
optimal coefficients of the adaptive filter and reduces the coupling to levels
below the desired one. The gain stays at low level for sufficient time to
allow the convergence of the algorithm and, subsequently, is increased
slowly until it reaches its regular value (Gfin). In the follow-up phase FSEG
the repeater operates normally, having attained the desired gain and
cancellation levels. Nevertheless, the adaptation algorithm of the adaptive
filter remains in operation to detect and to follow possible variations of the
frequency response of the coupling between antennas without having to
reinitiate the repeater. These changes in gain in the acquisition phase are
operated preferably in the intermediate frequency amplifiers AFI.
The analog configuration of the invention has two basic limitations.
In the first place, the limitations in complexity imposed by the very fact of
being analog, either in the number of coefficients of the adaptive filter, in
the type of adaptive filter or even in the adaptive algorithm used. In the
second place, the technological problems attached to the implementation of
the adaptive algorithm, such as for example the offset of the integrators
used in LMS, although alternatives can be found that alleviate the
seriousness of this problem (see [2]). Nevertheless, a preferred solution to
both problems is to use, at least in part, digital technology, whereby the
alternatives that the digital configuration of the proposed invention offers
are described below.
Digital configurations
Diverse alternative digital implementations to the basic analog
configuration described can be devised. They are all based on the same
principles that constitute the basis of the proposed invention. However, they
are distinguished from the basic configuration in that the adaptive filter
operates on the digitized signal, whereby it has the degrees of freedom
offered by the digital processing of the signal as well as the drawbacks that
the introduction of sampling and reconstruction steps of the analog signal
represent.
Architecture
The digital implementation of the adaptive filter requires the
digitization of the signals involved in the filtering and filter coefficient
estimation processes. Four alternative architectures are contemplated,
according to at what point of the main chain the signals are sampled and
reconstructed. These four options are shown in Figures 9-12.
Figures 9, 10 and 12 correspond to different strategies for sampling
the signal used in estimating the adaptive filter FAD response, while Figure
11 illustrates the case in which it is opted to implement the whole Fl
(intermediate frequency) step of the repeater digitally. In this case, the
signal can be regenerated (demodulation and modulation), substantially
improving the repeater's performance. In all cases, the signals can be
sampled in baseband (called I/Q sampling), so that the analytical signal is
recovered, or in the pass band signal (called Fl sampling), either in the Fl
signal or transferred to another lower frequency that is more convenient
from the sampling point of view. It should be borne in mind that the
sampling alternatives commented upon below (generation of the analytical
signal, heterodynation) are also applicable to the analog implementation
described as basic configuration, although the technological complexity of
the implementation thereof dissuades its use. The application of one or
another sampling type to each of the A/D converters gives place to four
possible configurations that differ from each other in that they work with real
or complex coefficients and/or error signal. Although the four combinations
lead to similar solutions and cancellation levels, all of them require a
different design of the sampling frequency and of the number of coefficients
of the adaptive filter, although said design is always based on the principles
already expressed for the analog configuration. Generally speaking, the
adaptive algorithms afford a quicker convergence in those configurations
that work with complex coefficients, but this improvement is offset by a
greater technological complexity of the implementation of the A/D and D/A
conversion. In all cases, the election of the sampling frequency will depend
on the same parameters as in the analog configuration already commented
upon (value of the intermediate frequency adopted in the repeater and
signal band width) as well as of the selectivity of the filters in the main
chain.
The A/D and D/A conversion steps are implemented in a different
way according to whether I/Q or Fl sampling is elected. Figure 13 illustrates
a scheme for the A/D conversion step. The antialiasing filter FAL and the
local oscillator (f1) are necessary or unnecessary according to which
sampling method is elected and how selective the filters of the main chain
are. Figure 14 illustrates a scheme for the D/A conversion step. The
reconstruction filter FRE compensates for the part of the distortion
introduced by the D/A converter that has not been corrected digitally. The
application or otherwise of the local oscillator (f2) depends on the sampling
type and frequency used, while the pass band filter FPB is only necessary if
heterodynation is applied with the local oscillator and the subsequent filters
of the main chain are not sufficiently selective.
Finally, in all the previous architectures it might be desirable to aid
the adaptive filter with a system compensating for possible distortions or
phase jitters or frequency offsets ([4]). Although the adaptive filter can in
theory compensate for said distortions, in practice, above all in multicarrier
systems such as is DVB, it cannot adapt itself sufficiently quickly to follow
said perturbations of the carrier.
Filter implementation architecture for signal processing.
The filter can be implemented by means of an in-line delay
architecture or by means of a lattice network. On the other hand, a finite
(FIR) or infinite (IIR) impulse response filter can be used. The IIR-filters
provide a better performance for same number of coefficients, but have
limitations in their combination with adaptive algorithms.
If the environment where the repeater is located causes very late
echoes of the transmitted signal to appear in the receiving antenna (with a
delay greatly in excess of the filter sampling period), is advisable to break
the adaptive filter down into two blocks FAD I and FAD II (see Figure 15),
one of them, FAD I, operating with the transmitted signal and the other,
FAD II, with the same signal delayed by a time interval similar to the
appearance time of the late echoes.
The filtering process can be implemented in the time or frequency-
domain. In the latter case, the use of the FFT or other filter banks with
multirate structures allows the computational cost to be reduced and the
convergence of the adaptive algorithm ([10], [5]) to be improved.
Nevertheless, the frequency implementation of adaptive filters operating in
environments where coupling between antennas is subject to variations
with time limits the capacity of following said changes and, furthermore,
always causes the introduction of a delay in the adaptive filter that can limit
its coupling cancellation capacity.
Adaptive algorithms for estimating filter coefficients.
The adaptive filter coefficients can be estimated with any algorithm
that guarantees convergence to the correct solution, that is to say, to the
one in which the adaptive filter correctly estimates the coupling between
antennas, and which, furthermore, is capable of following the variations
with time thereof. From among said algorithms, the most common and
studied are those which follow criteria: I) of minimum mean squared error or
II) of least squares. Both criteria can be used both in delay line architecture
and in lattice architecture.
There is given below a non-exhaustive list of the applicable
algorithms, in attention to whether they are implemented with FIR or IIR
response.
In the case of the FIR-filter ([3]), and taking criterion I), one
possibility is to use the steepest descent family of algorithms, which,
starting out from any initial value of the coefficients correct it with an
increment in the opposite direction to that of the gradient which has the
power surface depending on the coefficients. The gradient is calculated in
an exact statistical way; for this reason, these algorithms need to have prior
knowledge of the characteristics of the echoes or coupling to be cancelled
and cannot properly be described as adaptive.
Where the echoes or coupling are unknown, one solution is to have
recourse to the family of differential steepest descent algorithms which
calculate the gradient starting from differences in the error function
originated by perturbations which are caused in the coefficients.
Nevertheless, these algorithms converge slowly and a quicker solution is to
use the stochastic gradient, that is to say, to use an instantaneous
calculation of the exact expression of the gradient. Where the error function
to be minimized depends linearly on the coefficients, said algorithms are
included in the so-called LMS (Least Mean Square) algorithm and variants
such as the NLMS {Normalized LMS), the P vector algorithm or other
variants that have scenarios in which that the interfering noise is not
distributed in an uniform or white way in frequency, but rather do it in a
colored way. Where the error function to be minimized depends non-
linearly on the filter coefficients or depends recursively thereon, the
stochastic gradient algorithms calculate the gradient with a small number of
iterative applications of the rule of the differential calculus chain. Where the
error function does not depend linearly on the coefficients, said algorithm is
a non-linear extension of the LMS called the back propagation algorithm
([1]). In the case to which this patent relates, the gradient of the error
depending on the filter coefficients must be derived from the rule of the
chain. In Figure 6 the error signal at the output of the adaptive filter is
shown to be the one that, properly amplified, is transmitted by the repeater.
Therefore, it is the error signal that will be coupled with the desired received
signal at the input to the receiving antenna and it will become part of the
error signal again. Therefore, the LMS algorithm is an approximation of the
gradient that allows the calculation to be reduced regarding the rule of the
chain, which is the one that should really be applied. Only if the
intermediate frequency filter FFI, shown in Figure 6, introduces an
appropriate delay, will the transmitted signal be incorrelated with the error
signal and the LMS will make an increasingly valid approximation of the
gradient.
In the case of the FIR-filter, and considering criterion II), the most
feasible and effective algorithm is the Kalman filter or variants thereof: RLS
or Recursive Least Squares and the rapid Kalman filter.
In the case of IIR-filters, implementation by means of lattice network
is preferable to the delay line, since it allows the stability of the adaptive
filter to be easily monitored during the convergence phase. The filter
coefficients can be estimated by means of steepest descent algorithms,
algorithms based on the methodology of Steiglitz-McBride or on
hyperstable algorithms ([12], [7]). The first group is based on the minimum
mean squared error criterion, based either on the output error or on the
equation error. It is preferred to minimize the output error (e.g. using the
RPE or Recursive Prediction Error algorithm [9]), since the cost function
based on the equation error does not guarantee convergence to the optimal
solution under conditions of a low signal to noise ratio, as is the case of the
proposed invention. Among the algorithms of the third group there is the
very well known SHARF or Simplified Hyperstable Adaptive Recursive
Filter, having the qualities desired in an adaptive algorithm.
Finally, it must be borne in mind that if it is opted to implement the
whole Fl step of the repeater digitally (Figure 11), complete digital
processing of the received signal can be carried out, that regenerates it by
means of demodulation and subsequent modulation. In this case,
advantage can be taken of the training sequences proper to the signals
retransmitted, if there were any, such as for example the pilot tones of the
DVB signal, which can be taken as a reference when calculating the error
function that will govern any of the described algorithms. The advantage of
taking the training sequences as a reference is that they allow the adaptive
algorithm to work under better SNR conditions, as well as facilitating the
use of different architectures for inserting the adaptive system in the single
frequency signal repeater.
EXAMPLE OF ANALOG EMBODIMENT
Introduction
There is described one example of an analog embodiment of a DVB
signal repeater, which is based on a finite impulse response adaptive FIR-
filter of a structure according to the one shown in Figure 7, composed of a
delay line. As illustrated in Figure 7, the input signal E of the adaptive filter
is injected into the delay line. Subsequently, the signals present at the
output of each delay are multiplied by their respective coefficients, and
finally the results of such multiplications are added to obtain the output
signal S of the adaptive filter. In this embodiment, the obtaining of the value
of the coefficients is based on the LMS or Least Mean Squares algorithm.
Figure 16 illustrates the implementation by means of a correlation loop of
the calculation of coefficients according to the LMS algorithm. Said
algorithm calculates each of the coefficients according to the following
expression:
where the signal x(t) is the delay output signal, the signal r(t) is the signal
present in the receiving antenna, ea(t) is the error signal after the
cancellation, and a(t) is the ensemble of coefficients. ma is the adaptation
constant, which fixes the speed of convergence of the algorithm, as well as
the magnitude of the oscillation of the coefficients relative to the final
solution. The elements M1 and M2 are multipliers, the element INT is the
integrator and the element SUM is the adder of the signals multiplied by the
respective coefficients.
Approach to the embodiment
The single frequency or isofrequency repeaters usually have
intermediate frequency processing, which allows high selectivity filters to be
used.
The adaptive filter can be embodied at the same intermediate
frequency as that of the repeater, or at a second lower intermediate
frequency, should the possible advantages of operating at a lower
frequency compensate for the increment in complexity on having to add a
converter from the first to the second intermediate frequency before the
adaptive filter, and another converter from the second to the first
intermediate frequency, at the output from said filter. The block diagram of
the implementation example in question is the one illustrated in Figure 6.
Blocks used
The group of basic blocks that compose the adaptive filter in its
analog embodiment is as follows:
Delay blocks
Multipliers
Integrators
Distributors and combiners
Signal amplifiers
Adder
Figure 17 illustrates the block diagram of one of the cells where a
signal x(t-iT) is seen to be input, a distributor DIS distributes it, on one
hand, to a delay T to obtain an output x(t-(i+1)T), on the other hand, to a
multiplier M2, and, again, to a multiplier M1. The multiplier M1 also receives
the error signal e(t).
Figure 18 illustrates a stretch of two cells, CEL1 and CEL2, of the
analog embodiment of the adaptive filter.
Embodiment of each block
Delay blocks
The delay blocks can be embodied by means of the following
techniques:
Surface acoustic wave (SAW) delay lines
Resonator circuits, for example LC, ceramic or dielectric.
Pass band filters with sufficiently flat group delay
Transmission lines, for example coax lines, microstrip or stripline.
Multipliers
The multipliers can be embodied by means of the following
techniques:
- Multiplier circuits, for example four quadrants.
- Variable gain amplifiers.
- Attenuators with electronic control, for example based on PIN
diodes.
Integrators
The integrators can be embodied by means of the following
techniques:
- By means of operational amplifiers
- By means of first order low pass filters.
The scheme of Figure 19 illustrates an integrator embodied by
means of an operational amplifier, where E indicates the input to and S the
output from the integrator.
Distributors and combiners
The distributors and combiners can be embodied by means of the
following techniques:
- With coils and capacitors, for example by means of Wilkinson type
networks.
- Resistive, for example by means of star or delta networks.
- By means of directional couplers.
An example of a resistive distributor embodied by means of a star
resistor network is shown in Figure 20, where E also indicates the input to
and S the output from the distributor.
Signal amplifiers
The signal amplifiers can be embodied by means of the following
techniques:
Operational amplifiers with sufficient bandwidth.
Integrated monolithic amplifiers.
Directly by means of discret circuits based on transistors, following
the traditional radio frequency amplifier design techniques.
Adder
The adder adds the signals obtained at the output of each of the
multipliers M2 of the cells.
EXAMPLE OF DIGITAL EMBODIMENT
Introduction
Figure 21 shows the block diagram of the transmitter, also for DVB
signals, in which the adaptive filter is implemented digitally. In a similar way
to the analog case, the RF input signal SE is converted into intermediate
frequency (Fl), where a FI filter FFI rejects the possible out-of-band signals,
to finally convert it again to RF. In the digital embodiment, the
corresponding A/D and D/A conversions are required, as illustrated in
Figure 21.
The transmitter operates as follows: from the signal r[n] received by
the antenna (which in fact is formed by the desired signal plus the echoes
caused by the transmitting antenna) there is subtracted an estimate s[n] of
the undesired signals obtained by the adaptive filter FAD, which gives an
error signal e[n] that is fed through the Fl filter FFI and is reconverted to
RF.
The structure of the digitally embodied adaptive filter is shown in
Figure 22. Essentially it is a variable coefficient FIR-filter, where these are
updated periodically by the LMS algorithm.
For the instant n, the LMS algorithm calculates the new coefficients
as follows:
h[n] = h[n -1] + m . e[n]¦ x[n]
This operation, schematically illustrated in Figure 23, is carried out
for each of the filter coefficients.
The constant m has the function of adjusting the adaptation step.
The greater its value, the quicker will be the convergence of the algorithm;
in exchange, the coefficients will have a greater oscillation around the
optimum solution. There is, therefore, a compromise between convergence
speed and stability of the coefficients around the optimum solution.
Digital embodiment
In Fl we have the DVB pass band signal, with a bandwidth of
7.61MHz, which is digitized by means of a 12-bit A/D conversion. The
subsequent processing is based on programmable logic digital integrated
circuits of the FPGA (Field Programmable Gate Array) type.
Figure 24 shows more particularly the blocks required to be able to
embody the analog/digital interface.
A more detailed structure of the adaptive filter is as shown in Figure
25. It is seen to be formed by the following parts:
Data-Framer DF: Its function is to fragment the input data x[n] (12
bits) into four parts, giving 3 bits to which a control bit is added. Because
four cycles will be needed to process each sample, the clock frequency of
the system should be four times the sampling frequency.
Table Generator TG: calculates the partial products of each
coefficient and loads them in the LUT (Look-Up Table) of the corresponding
tap.
LMS Algorithm: This calculates the new coefficients depending on
the error signal e[n] and on x[n], and supplies them to the Table Generator
TG, through cdata.
Tap Filter coefficients: Each Tap has assigned its own coefficient,
and multiplies it by the input data.
Adder tree: this is a registered adder tree that obtains the sum of all
the partial results supplied by the coefficients.
Each of the parts comprising the system will be shown in greater
detail below. Thus, Figure 26 is the diagram of the Data-Framer DF. The 12
bit data (x[n]) are fed to a register REG12 at the sampling frequency FMUE,
and are output, through a control register REGC fragmented into 4 parts of
4 bits each (each of these parts is known as an octet), at a frequency
4xFMUE 4 times higher than the sampling frequency FMUE. The three
least significant bits of each octet correspond to three bits of the input data;
the most significant bit is the control bit, and is set to 1 only when the fourth
and last octet of the data being processed is output. This is achieved by
means of a 2-bit counter C0N2 and the three multiplexors Mux1, Mux2 and
Mux3. Initially the counter C0N2 is set to 0, therefore the multiplexors
Mux1, Mux2 and Mux3 select the signal at their 0 input, that is, Mux1
selects bit 2, Mux2 bit 1 and Mux3 bit 0, that is to say, the three least
significant bits of the sample. The control signal SCO is the most significant
bit in the octet, and in this case is 0. Subsequently the counter is set to 1,
and bits 5, 4 and 3 (the following three least significant bits) are selected
and the control signal continues to be 0. Subsequently, when the counter is
set to 2, bits 8, 7 and 6 are selected, with the control signal still being 0.
Finally the counter is set to 3, bits 11,10 and 9 (the most significant in the
sample) are selected, and the control signal SCO is activated. In the
following iteration the counter will be reset to 0 and there will be a new
sample stored in the register REG12, the cycle beginning again.
Figure 28 is an illustration of the diagram corresponding to the Table
Generator TG. This calculates the partial products of each coefficient
according to Table 1.
It could have been implemented simply with a 12*3 bit multiplier plus
some minor components, but the configuration shown has been opted for,
since is better as far as space occupation is concerned. The operation
thereof is rather complex, although it is known to one skilled in the art,
whereby only the idea will be described here: if Table 1 is observed, it will
be seen that an accumulator is ideal for calculating the value of the memory
positions 0h to 7h, it is only necessary to add the coefficient to the last result
obtained to get the current value. The same may be said of the positions 8h
to Bh and of Ch to Fh, only in the last named it is necessary to complement
the coefficient. The counter is used to access the consecutive memory
positions of the LUT. The function of Mux5 is to supply either 0 or the adder
output to the output register. The control signal of Mux5 is the OR of the 3
least significant bits of the counter. Therewith 0 is selected for the memory
positions 0 and 8 (as shown in Table 1). On the other hand, the adder is fed
back through its own output, thereby obtaining the results for the addresses
1 to 7 and 9 to 11. The addresses 12 to 15 require negative results, to this
end the two's complement of the coefficient is selected by means of Mux4.
Finally, 0 is selected for the addresses 12 to 15 by means of Mux6, since
from now on the results become negative.
Figure 27 shows the structure for each tap comprising the filter. It
consists of three clearly differentiated parts:
Time Skew Buffer TSB: this is a 4x4 bit shift register, thus the TSB
is able to house a complete sample (divided into 4 parts of 4 bits each). At
each clock cycle it delivers an octet to the Partial Product Multiplier PPM
and also to the following tap. From the TSB there is also output the
First_oct signal that is only activated when the least significant octet of the
four forming the sample is being output. The Time Skew Buffer TSB
receives the information data1 coming from the control register REGC.
Partial Product Multiplier PPM: this consists of two RAM (called
LUT: Look-Up-Table) that store the partial products of the coefficient
according to Table 1. At the same time as one LUT is accessed to read the
partial result of the multiplication, the Table Generator TG is writing the
partial products corresponding to the new coefficient that the LMS algorithm
will have calculated to the other LUT, by means of the data buses and addr.
The multiplexors Mux8 and Mux9 are controlled by means of the bank_sel
signal and are complementary, that is to say, when one selects its 0 input,
the other selects its 1 input, and vice versa. This allows the addr and data
signals to be addressed to the LUT corresponding thereto. The multiplexor
Mux10, also controlled by the bank_sel signal, selects the data outputted
by the LUT containing the partial products of the current coefficient (it
should remember that meanwhile the other LUT is being updated with the
partial products of the new coefficients). By means of the two AND gates
and the bank_sel and tap_sel signals both write_en signals that enable the
writing to the corresponding LUT are being generated.
Scaling Accumulator SCA: Its mission is to accumulate the partial
results of each octet properly to obtain the complete solution of the
multiplication (24 bits in all) at its output. It will be seen that it is an adder
fed back by its own output, suitably escalated (the 13 most significant bits
are fed back and the most significant bit is replicated three times). Mux11,
which is controlled by means of the first_oct signal, allows the first octet to
pass directly to the output; the other three octets forming the sample pass
through the adder.
Obviously all the steps and elements described above are
schematic, to facilitate the understanding of the invention. Details have not
been included that are evident to one skilled in the art (additional filtering
steps, etc.), and which, moreover, do not affect the concept of the
invention. Thus, for example, the steps of filtering the process signal, of
amplification of the process signal, of automatic gain control, etc. are
frequently not done at one same time, but rather they are done in diverse
steps. It should be understood, therefore, that where the existence of a
step, for example, of filtering, is mentioned, it should not be understood that
only one filtering step exists, but rather the signal in question is filtered,
independently of the number of steps in which said filtering is carried out.
Also, the order indicated in the steps of the process is purely a
descriptive order and does not have to coincide with the true order of the
process. The only intention is to say that the process comprises said steps,
that is to say, that it includes them, but it is not indicating that the sequence
of performance of the steps is the one indicated. In fact, for example, the
filtering, amplification and automatic gain control steps of the process signal
that, as has already been indicated previously, are usually carried out in
several stages, do not always follow the order indicated in the text, it is
even frequent that the different stages of a step are interlaced with the
stages of the other steps. Therefore, it is desirable to insist on the fact that
the steps mentioned in the text and the claims only indicate the existence of
said steps, without restricting the number of stages in which they are
carried or the order in which they take place.
References
[1] J.A.Anderson and E.Rosenfeld (Eds.), Neurocomputing:
Foundations of Research, M.I.T. Press 1988
[2] R.T.Compton Jr., Adaptive Antennas. Concepts and Performance,
PRENTICE-HALL 1988, ISBN: 0-13-004151-3.
[3] R.D.Gitlin, J.F.Hayes, S.B.Weinstein, Data Communication
Principles, Plenum Press 1992.
[4] D. Harman, J.D. Wang, E J.J. Werner, Frequency Offset
Compensation Techniques for Echo-Cancellation Based Modems,
Conference Record Globecom'87, Tokyo, Japan.
[5] Youhong Lu and Joel M.Morris, "Gabor Expansion for Adaptive
Echo Cancellation", IEEE Signal Processing Magazine, Vol.16, N°2 pag.68-
80, March 1999, ISSN: 1053-5888.
[6] Odile Macchi, Adaptive Processing. The Least Mean Squares
Criterion with Applications in Transmission, John Wiley & Sons 1995, ISBN:
0-471-93403-8
[7] Phillip A.Regalia. Adaptive IIR Filtering in Signal Processing and
Control Marcel Dekker 1995. ISBN: 0-8247-9289-0.
[8] J.R.Rosenberger and J.Thomas, "Performance of an Adaptive Echo
Canceller Operating in a Noisy, Linear, Time-Invariant Environment", The
Bell System Technical Journal, Vol.50 N°3 pag.785-813, March 1971
[9] John J.Shynk, "Adaptive IIR Filtering", IEEE ASSP Magazine, Vol.6
N°2 pag.4-21, April 1989, ISSN: 0740-7467
[10] John J.Shynk, "Frequency-Domain and Multirate Adaptive Filtering",
IEEE Signal Processing Magazine, Vol.9, N°1 pag. 14-37, January 1992,
ISSN: 1053-5888
[11] B.M.Sondhi, "An Adaptive Echo Canceller", The Bell System
Technical Journal, Vol.46 N°3 pag.497-511, March 1967
[12] John R.Treichler, C.Richard Johnson and Michael G.Larimore,
Theory and Design of Adaptive Filters, Wiley & Sons 1987, ISBN: 0-471-
83220-0
Claims
1. A method for re-transmitting single frequency signals, of the type used in a single frequency signal repeater and
comprising the steps of.
[a] receiving a first radio frequency signal through a receiving antenna, said first rad o frequency signal having
a receiving power, [b] converting said first radio frequency signal into a process signal (RF/FI), [c] filtering sad
process signal, [d] amplifying said process signal, [e] automatically controlling said process signal gain (CAG),
[f] reconversion of said process signal into a second radio frequency signal (FI/RF), [g] ampl.fying (AP) the
power of said second radio frequency signal, [h] output filtering said second radio frequency signal, [:] trans-
mitting said second radio frequency signal through a transmitting antenna where a coupling (A(s)) takes place
between said transmitting antenna and said receiving antenna, and [j] a step of adaptive cancellation wth an
adaptive algorithm of said coupling (A(s)) between said transmitting antenna and said receiving antenna said
step of adaptive cancellation comprising a negative feedback of a feedback signal (F(s)) and said step of
adaptive cancellation operating on digitized signals, characterized in that said step of adaptive cancelation
takes, as input signal, said process signal digitized after said automatic gain control step (CAG).
2. A method for re-transmitting single frequency signals, of the type used in a single frequency signal repeater and
comprising the steps of:
[a] receiving a first radio frequency signal through a receiving antenna, said first radio frequency signal having
a receiving power, [b] filtering said first radio frequency signal, [c] amplifying said first radio frequency signal,
[d] automatically controlling said first radio frequency signal gain (CAG), [e] amplifying (AP) the power of a
second radio frequency signal, [f] output filtering said second radio frequency signal, [g] transmitting said
second radio frequency signal through a transmitting antenna, where a coupling (A(s)) takes place between
said transmitting antenna and said receiving antenna, and [h] a step of adaptive cancellation with an adaptive
algorithm of said coupling (A(s)) between said transmitting antenna and said receiving antenna, said step of
adaptive cancellation comprising a negative feedback of a feedback signal (F(s)), said step of adaptive can-
cellation operating on digitized signals, and said step of adaptive cancellation having as output signal said
second radio frequency signal, characterized in that said step of adaptive cancellation takes, as input signal,
said first radio frequency signal digitized after said automatic gain control step (CAG).
3. The method as claimed inclaim 1, wherein it introduces an non-correlation delay in said process signal that estab-
lishes an non-correlation between said process signal and said feedback signal (F(s)) and wherein said non-
correlation delay is introduced in said conversion step (RF/FI).
4. The method as claimed in claim 1 or 3, wherein said feedback signal (F(s)) is negatively fedback to a signal of the
group formed by said process signal before said automatic gain control step (CAG) and said process signal after
said automatic gain control step (CAG).
5. The method as claimed inclaim 2, wherein said feedback signal (F(s)) is negatively fedback to a signal of the group
formed by said first radio frequency signal before said automatic gain control step (CAG) and said first radio fre-
quency signal after said automatic gain control step (CAG).
6. The method as claimed in at least one of claims 1, 3 or 4, wherein said cancellation step time averages said process
signal or said first radio frequency signal.
7. The method as claimed in at least one of claims 1 to 6, wherein said cancellation step is implemented in a domain
of the group formed by a time domain and a frequency-domain.
8. The method as claimed in at least one of claims 1 to 7, wherein said adaptive algorithm follows a criterion of the
group formed by the criterion of minimum mean squared error and the criterion of least squares.
9. The method as claimed in claim 8, where said adaptive algorithm follows a criterion of minimum mean squared
error, wherein said adaptive algonithm is an adaptive algonithm of the group formed by a steepest descent adaptive
algonithm, a differential steepest descent adaptive algonithm, a stochastic gradient adaptive algonithm, an LMS
(Least Mean Squares) algonithm, an NLMS (Normalized Least Mean Squares) algonithm, an adaptive algonithm
based on the sign function, a P vector adaptive algonithm, and a back propagation adaptive algorithm.
10. The method as claimed inclaim 8, where said adaptive algonithm follows a criterion of least squares, wherein said
adaptive algonithm is a Kalman filter.
11. The method as claimed in at least one of claims 1 to 7, wherein said adaptive algonithm is based on the methodology
of Steiglitz-McBride or on the simplified hyperstable recursive adaptive filter (SHARF).
12. The method as claimed in at least one of claims 1, 3, 4 or 6, wherein said cancellation step is broken down into at
least one first sub-step (FAD I) and one second sub-step (FAD II), where said first sub-step (FAD I) operates with
said process signal and said second sub-step (FAD II) operates with said process signal delayed in time.
13. The methodas claimed in at least one of claims 2 to 11, wherein said cancellation step is broken down into at least
one first sub-step (FAD I) and one second sub-step (FAD II), where said first sub-step (FAD I) operates with said
first radio frequency signal and said second sub-step (FAD II) operates with said first radio frequency signal delayed
in time.
14. The methodas claimed in at least one of claims 8 to 13, where said single frequency signal comprises training
sequences, wherein said training sequences are taken as a reference signal at the time of calculating an error
function that will govern said adaptive algonithm.
15. A single frequency signal repeater, of the type comprising:
[a] a receiving antenna, for receiving a first radio frequency signal, with a receiving power, [b] a base unit (UB),
for converting said first radio frequency signal into a process signal, filtering said process signal, amplifying
said process signal, automatically controlling said process signal gain (CAG), and reconverting said process
signal into a second radio frequency signal, [c] a power amplifier unit (AP), [d] an output filter (FS), [e] a
transmitting antenna, where said transmitting antenna and said receiving antenna may undergo coupling (A
(s)), and [f] an adaptive device with an adaptive algonithm for canceling said coupling (A(s)) between said
transmitting antenna and said receiving antenna, said adaptive device comprising a negative feedback of a
feedback signal (F(s)) and said adaptive device operating on digitized signals, characterized in that said
adaptive device takes, as input signal, said process signal digitized after automatically controlling said process
signal gain (CAG).
16. A single frequency signal repeater, of the type comprising:
[a] a receiving antenna, for receiving a first radio frequency signal, with a receiving power, [b] a base unit (UB),
for filtering said first radio frequency signal, amplifying said first radio frequency signal, and automatically
controlling said first radio frequency signal gam (CAG), [c] a power amplifier unit (AP), [d] an output filter (FS),
[e] a transmitting antenna, where said transmitting antenna and said receiving antenna may undergo coupling
(A(s)), and [f] an adaptive device with an adaptive algonithm for canceling said coupling (A(s)) between said
transmitting antenna and said receiving antenna, said adaptive device comprising a negative feedback of a
feedback signal (F(s)), said adaptive device operating on digitized signals and said adaptive device having as
output signal a second radio frequency signal, characterized in that said adaptive device takes, as input
signal, said first radio frequency signal digitized after automatically controlling said first radio frequency signal
gain (CAG), where said second radio frequency signal is input to said power amplifier unit (AP).
17. The repeater as claimed in claim 15, wherein said feedback signal (F(s)) is negatively fedback to a signal of the
group formed by said process signal before said automatic gain control (CAG) and said process signal after said
automatic gain control (CAG).
18. The repeater as claimed in claim 16, wherein said feedback signal (F(s)) is negatively fed back to a signal of the
group formed by said first radio frequency signal before said automatic gain control (CAG) and said first radio
frequency signal after said automatic gain control (CAG).
19. The repeater as claimed in claim 15 or 17, wherein said device time averages said process signal or said first radio
frequency signal.
20. The repeater as claimed in at least one of claims 15 to 19, wherein said device is implemented in a domain of the
group formed by the time domain and the frequency-domain.
21. The repeater as claimed in at least one of claims 15 to 20, wherein said device comprises an adaptive filter (FAD)
that follows said adaptive algonithm and wherein said adaptive filter (FAD) is implemented by means of an in-line
delay architecture or by means of a lattice network.
22. The repeater as claimed in at least one of claims 15 to 20, wherein said device comprises an adaptive filter (FAD)
that follows said adaptive algonithm and wherein said adaptive filter (FAD) has a finite impulse response (FIR filter)
or an infinite impulse response duration (IIR-filter).
23. The repeater as claimed in at least one of claims 15 to 22, wherein said adaptive algonithm follows a criterion of
the group formed by the criterion of minimum mean squared error, and the criterion of least squares.
24. The repeateras claimed in claim 23, where said adaptive algonithm follows the criterion of minimum mean squared
error, wherein said adaptive algonithm is an adaptive algonithm of the group formed by a steepest descent adaptive
algonithm, a differential steepest descent adaptive algonithm, a stochastic gradient adaptive algonithm, an LMS
(Least Mean Squares) algonithm, an NLMS (Normalized Least Mean Squares) algonithm, an adaptive algonithm
based on the sign function, a P vector adaptive algonithm, and a back propagation adaptive algonithm.
25. The repeater as claimed in claim 23, where said adaptive algonithm follows a criterion of least squares, wherein
said adaptive algonithm is a Kalman filter.
26. The repeater as claimed in at least one of claims 15 to 20, wherein said adaptive algonithm is based on the meth-
dology of Steiglitz-McBride or on the simplified hyperstable recursive adaptive filter (SHARF).
27. The repeater as claimed in at least one of claims 15 to 26, where said single frequency signal comprises training
sequences, wherein said training sequences are taken as a reference signal at the time of calculating an error
function that will govern said adaptive algonithm.
28. The repeater as claimed in at least one of claims 15 and 17 to 19, wherein said device comprises an adaptive filter
(FAD) that follows said adaptive algonithm and wherein said adaptive filter (FAD) is broken down into at least a
first block (FAD I) and a second block (FAD II), where said first block (FAD I) operates with said process signal
and said second block (FAD II) operates with said process signal delayed in time.
29. The repeater as claimed in at least one of claims 16 to 20 and 23 to 27, wherein said device comprises an adaptive
filter (FAD) that follows said adaptive algonithm and wherein said adaptive filter (FAD) is broken down into at least
a first block (FAD I) and a second block (FAD II), where said first block (FAD I) operates with said first radio
frequency signal and said second block (FAD II) operates with said first radio frequency signal delayed in time.
30. A method for re-transmitting single frequency signals, substantially as herein
described, particularly with reference to and as illustrated in the accompanying
drawings.
31. A single frequency signal repeater, substantially as herein described, particularly with
reference to and as illustrated in the accompanying drawings.
The invention discloses a process for re-transmitting single
frequency signals and a single frequency signal repeater, where coupling
occurs between the transmitting antenna and the receiving antenna, and
where the process is of the type used in a single frequency signal repeater
and comprises the steps of: [a] receiving a first radio frequency signal
having a particular receiving power, [b] optionally converting said first radio
frequency signal into a process signal, [c] filtering, amplification and
automatically controlling the gain of said signal, [d] canceling said coupling
between said transmitting antenna and said receiving antenna, [e]
reconverting, as the case may be, said process signal into a second radio
frequency signal, [f] amplifying the power of said second radio frequency
signal, [g] output filtering, and [g] transmission.

Documents:

in-pct-2002-1041-kol-granted-abstract.pdf

in-pct-2002-1041-kol-granted-assignment.pdf

in-pct-2002-1041-kol-granted-claims.pdf

in-pct-2002-1041-kol-granted-correspondence.pdf

in-pct-2002-1041-kol-granted-description (complete).pdf

in-pct-2002-1041-kol-granted-drawings.pdf

in-pct-2002-1041-kol-granted-examination report.pdf

in-pct-2002-1041-kol-granted-form 1.pdf

in-pct-2002-1041-kol-granted-form 18.pdf

in-pct-2002-1041-kol-granted-form 2.pdf

in-pct-2002-1041-kol-granted-form 3.pdf

in-pct-2002-1041-kol-granted-form 5.pdf

in-pct-2002-1041-kol-granted-gpa.pdf

in-pct-2002-1041-kol-granted-reply to examination report.pdf

in-pct-2002-1041-kol-granted-specification.pdf

in-pct-2002-1041-kol-granted-translated copy of priority document.pdf


Patent Number 225471
Indian Patent Application Number IN/PCT/2002/1041/KOL
PG Journal Number 46/2008
Publication Date 14-Nov-2008
Grant Date 12-Nov-2008
Date of Filing 12-Aug-2002
Name of Patentee MIER COMMUNICACIONES, S.A.
Applicant Address POLIGON INDUSTRIAL CONGOST, PARCEL.LA 4-S, 08530-LA GARRIGA (BARCELONA)
Inventors:
# Inventor's Name Inventor's Address
1 IBORRA ARCHS DOMENEC POLIGON INDUSTRIAL CONGOST, PARCEL.LA 4-S 08530-LA GARRIGA (BARCELONA)
2 MIER ALBERT PEDRO POLIGON INDUSTRIAL CONGOST, PARCEL.LA 4-S 08530-LA GARRIGA (BARCELONA
3 MERITXELL LAMARCA OROZCO MARIA RAMBLA DE CATALUNYA 106 08008-BARCELONA
4 NAJAR MARTON MONTSERRAT MARQUES DE CORNELLA 73-75 08940-CORNELLA DE LLOBREGAT (BARCELONA)
5 PEREZ NEIRA ANA ISABEL CAN MALAGRIDA 13,08338-PREMIA DE DALT(BARCELONA)
6 BARBA MIQUEL XAVIER POLIGON INDUSTRIAL CONGOST, PARCEL.LA 4-S 08530-LA GARRIGA (BARCELONA)
7 VAZQUEZ GRAU GREGORI MARINA 228-232, 08013-BARCELONA
PCT International Classification Number H04B 7/15
PCT International Application Number PCT/ES01/00035
PCT International Filing date 2001-02-08
PCT Conventions:
# PCT Application Number Date of Convention Priority Country
1 P 200000379 2000-02-18 Spain