Title of Invention	A METHOD FOR PROCESSING DIGITAL DATA VALUES
Abstract	The invention relates to a method for processing digital data valued according; to which a diffirential value is formed from a respective current digital data value and a respective predicted value with the aid of a predictor method, compressed 12 ves being generated from thedifferential valuse in a Rice process that succeeds the predictor method. The aim of the invention is to 12 Ty out said method relatively rapidly. To achieve this, the respective predicted value is obtained using a signal model, which discover the expected temporal progression of the digital data values.

Title of Invention

A METHOD FOR PROCESSING DIGITAL DATA VALUES

Abstract

The invention relates to a method for processing digital data valued according; to which a diffirential value is formed from a respective current digital data value and a respective predicted value with the aid of a predictor method, compressed 12 ves being generated from thedifferential valuse in a Rice process that succeeds the predictor method. The aim of the invention is to 12 Ty out said method relatively rapidly. To achieve this, the respective predicted value is obtained using a signal model, which discover the expected temporal progression of the digital data values.

Full Text	Description Method for processing digital data values The invention relates to a method for processing digital data values, in which a predictor method is used to form a difference value from the respective current digital data value and from a respective predicted value, and compressed values are produced from the difference values using a RICE method which follows the predictor method. A method such as this is known to those skilled in the act from, for example, the Internet site which can be called up at http;//www.monkeysaudio,com/theory.html (called up on November 11, 2002). This site explains how digital audio data can be compressed for the purpose of compression with the aid of a predictor method and a subsequent RICE method. Those audio data values which immediately precede the current audio data values are in each case used to calculate predicted values which are required for the predictor method. Measurement values from processes that are being carried out, for example from production or from conversion processes, are also usually recorded in digital form for closed-loop and open- loop control of industrial plant, and are, made availab1e as output information for subsequent processing. The subsequent processing may, for example, be readjustment of a processed variable. In this case, it is often necessary Co transfer the recorded measurement values. in the form of digital data values or digital measurement data between different technical plant, or to Store such measurement values for subsequent processing. In order to reduce the amount of data PCT/DE 2003/003862 -2- 2002F11723WOIN required for transmission of the digital measurement data, provision is made in conjunction with a closed-loop filling level control system that is known from the document WO 01/91031 for the digital measurement data to be compressed by means of digital data compression, before being transmitted. This makes it possible to reduce the bandwidth required for data transmission. One object of the invention is to specify a comparatively fast method for compression of digital data values. For a method of the type mentioned initially, this object is achieved according to the invention by obtaining the respective predicted value by means of a signal model which describes the expected time profile of the digital data values. The major, advantage which is achieved by the invention over the prior art is that a significant reduction is. achieved in the memory space requirement for the compressed values determined from the digital data values, and at high speed, with the aid of the proposed method. This is because, the digital data, values can be compressed without any losses in real time, that is to say immediately after they have been, recorded. The predicted values for the predictor method are determined as quickly end reliably as possible in that, when the pzedictor method is being carried out, the predicted values are determined on the basis of a signal model which is suitable for describing the time profile of the digital data values. The use of a signsl model such as this means that the predicted values are available to form the difference values without the need. PCT/DE 2003/003862 - 3 - 2002F11723WOIN for a calculation in each case on the basis of previous data values, and thus without any time delay. The space required in the memory is leas than that for uncompressed data values. The proposed method results in the memory space requirement being reduced with minimal computation power for processing of the digital data values, since only simple computation operations are required, such as addition, subtraction and bit manipulations. This makes it possible to use the stated method for processing digital data values even in technical equipment which has the computation power. by means of suitable processors only to a restricted extent. The requirement tor computation power for compression, which is not available for other functions, is minimized. The compression process carried out using the method according to the invention means that, in, this case, only data which contains no. information is removed. The digital measurement data can thus be reproduced completely, that is to say without any losses, during the decompression process,. One embodiment of the method according to the invention comprises the signal model being determined on the basis of a sine function or a cosine function with a constant period and its respective harmonics, a decaying exponential function or a sine function with a decaying exponential function as an envelope curve. This advantageously allows the respective predicted value to be determined particularly quickly, since no other calculations need be carried out. PCT/DE 2003/003862 4 - 2002F11723WOIN A further solution for the object mentioned above comprises on the basis of the method of the type mentioned initially the digital data values being obtained, according to the invention from periodic signals and at least that digital data value which was recorded one period befor the current digital data value being used as the respective predicted value associated with the respective current digital data value. The required computation power can advantageously be kept low by the use for the predicted value of the respective date value recorded one period before a current data value. No additional computations tieed be carried, out, so that the method can be carried out at high speed. In order to improve the accuracy of the respective predicted value, it is also possible. to use, in addition, to the digital data value recorded precisely one period before the respective current digital data value, further digital data values, which. were recorded an integer number of periods before the respective carrent digital data value, in order to form the respective predicted value. These can be linked to one another by simple mathematical functions, such as averaging and/or weighting, so that the computation power required for this purpose can be kept low. One advantageous further development of the method according to the invention also provides that, in the RICE method, an overflow, which is determined by subtraction of a data length (which has been predicted by means of a RICE prediction value; of the respective current difference value and a current data length of this difference value, is. compared with a limit value which can be predetermined, and the difference value is"emitted with a predetermined maximum data length as a compressed value PCT/DE 2003/003862 - 5 - 2002F11723WOIN if the overflow exceeds the limit value. The major advantage of this development is that, in the case of successive difference values whose data length fluctuates to a major extent, the RICE method is used only when it can be used effectively, that is to say with Little overflow. If there is a. large overflow, the respective compressed value is emitted with a maximum predetermined data length. Furthermore, one advantageous further development of the method according to the invention provides that the compressed values are transmitted via a data transmission path, and the difference values are then recovered from the compressed values using a RICE decpding method, and the digital data values are determined from the difference values and the respective predicted values by addition using a reverse predictor method. Transmission of the compressed measurement values via a data transmission path in this case covers both wire-based transmission methods and wireless transmission methods, such as radio transmissions, According to this development, decompression. of the compressed values can advantageously be carried out at a point which is remote from the point at which they were compressed. For example, the compressed value can be transmitted from an in-field appliance in an industrial plant via a data bus to a central computer, which evaluates the digital data values once the compressed values have been decompressed. In order to simplify subsequent processing of the compressed values or to transmit additional information, one preferred embodiment of the invention allows the compressed Values to be provided with header data. PCT/DE 2003/003862 - 6 - 2002F11723WOIN According to a further advantageous embodiment of the method according to the invention, the data values are formed from input measurement variables from in-field appliances. In this context, by way of example, in-field appliances are protective or control appliances such as those which are normally used in industrial, for example energy, chemical or petrochemical plant. In comparison to audio technology, in which a virtually indefinite time is available for compression of the digital data values, the comparison prscess in in-field appliances must be carried out particularly quickly, owing to the restricted computation and storage capacities. The method according to the invention can thus be used particularly advantageously in this case. Protective and/or control appliances in energy plant can also advantageously be used as in-field appliances. The invention will be explained, in more detail in the following text using exemplary embodiments. In this case, in the figures: Figure 1 shows a schematic illustration in order to explain a method for compression of the digital data values; and Figure 2 shows a schematic illustration in order to explain a method for decompression of compressed- digital data values. Figure 1 shows a schematic illustration in order to explain a method for compression, using digital data values which have been recorded using an in-field appliance that is not illustrated. PCT/DE 2003/003862 -7- 2002F11723WOIN Since the digital data values in the present exemplary embodiment are intended to be formed from (analog) input measurement variables of the in-field appliance, they are referred to in the following text as digital measurement data. A measurement device 1 is used to record measurement values, which are converted to digital measurement data in the normal way. First of all, a predictor method is carried out in order to process the digital measurement data. The predictor method is part of a. processing process for compression of the digital measurement data. in this case, predicted values for the digital measurement data are determined by means of a predictor device 2. The predictor method first of all reduces the data length (the number of bits) which is required to represent the digital measurement data, by forming difference values by subtraction from the digital measurement data and suitable predicted values. The predicted values that are used should in this case be as close as possible to the real digital measurement data, so that the difference values which result from the subtraction process are as small as possible. The further processing is then carried out using difference values whose required data length is considerably less than that of the digital measurement data. The predicted values for the predictor method can be formed using a signal modal which, for example, is based, on a sine function or a cosine function, as well as their respective harmonics, a decaying exponential function or a sine function with a decaying exponential function as an envelope curve. In this case, these are in each case functions which are represented and processed computationally with little effort. PCT/DE 2003/003862 -8- 2002F11723WOIN and are in some cases periodic. Previous digital measurement data which has been recorded before the digital measurement data to be compressed at that time is taken into account in the creation of the signal model for determination of the predicted models. Characteristic properties, such as the amplitude, period duration and decay behavior, of the previous digital measurement data ace used to form the signal model. Properties such as these can be calculated by simple computation operations and nevertheless allow the predicted values to be determined relatively reliably. In the simplest case, the previous digital measurement data. from precisely one previous recording period can be adopted unchanged as predicted values. The most recently described procedure, in which reference is made to a previous recording period, is particularly feasible when the recorded measurement values and thus the digital measurement data have or has a periodic behavior, as is frequently the case, for example, in energy plant. When the digital measurement data is processed for the first time using the predictor method, a start value must be provided for the first predicted value. As is shown in Figure 1, the digital measurement data is transmitted (10) from the measurement device 1 and the respective determined predicted values are transmitted (20) from the predictor apparatus 2 to a subtraction device 3, respectively, in which the predicted values are subtracted from the respectively associated digital measurement data. The subtraction process results in difference values, which are then passed (30) to on apparatus A in order to carry out a RICE method. PCT/DE 2003/003862 - 9 - 2002F11723WOIN In the RICE method, which ia known per se and whose details can be found from the references by a person skilled in the art (see, for example, http://WWW.monkeysaudio,com/theory.html [called up on November 13, 2002], the data length of the difference values which are produced by the predictor method are reduced, so that, in the end, compressed values are produced and emitted (40) . The same data length could, in each case, be used for transmission and storage of the difference values. These would have to correspond at Least to the maximum, possible data length of a difference value. However, since the required data length is generally not constant in the case of successive difference values, but can fluctuate from one difference value to the next, a not inconsiderable amount of memory space would the wasted for the transmission and storage of the difference values with the maximum data length. This will be explained in more detail using the following example: let us assumed that three digitally represented difference values are to be transmitted, to be precise 11010110, 1101 and 10110. If a constant date length were used for transmission, the maximum occurring data length of the difference values would have to be used, that is to say in this case 8 (data length of first difference value}. In consequence, the three difference values would have to be transmitted in the form 110101100000110100010110. The zeroes which are inserted in the maximum data length for transmission of the smaller difference values (1101, 10110) waste unnecessary memory space, since no additional in format ion is transmitted. The RICE method now provides an algorithmic method for scrambled difference values which result from the predictor method in a suitable manner, that is to say in a manner which PCT/DE 2003/003862 - 10 - 2002F11723WOIN is optimized with regard to memory space, to form compressed values. The basic idea for the RICE method is the compression of the difference values using a data Length which is matched to the respective difference value. A so-called RICE code is introduced in order to separate successive compressed values, which are obtained from the difference values, and in order to scramble information which it may possibly not be possible to represent with an excessively small data length. This will be explained for the following text. The BICE method requires a RICE predicted value for the expected data length of the respective subsequent difference value. The compressed values which are produced by means of the RICE method are then in general stored using the data length predicted by means of the RICE predicted value. If the data length of the difference values which are produced using the previous predictor method is greater than the data length predicted using the RICE predicted value, the overflow (the most significant bits which can no longer be represented in the predicted data length} are scrambled in the RICE code. The RICE code comprises a number of binary values 0, with this number resulting directly from the overflow, and a final binary value 1. The RICE code and a resultant value which is obtained, taking into account the predicted data length from the respective difference value, are attached directly to one another, in order to obtain a compressed value. The RICE, predicted value for the data length is obtained from the values for the data length of a specific number of previous digital measurement data items, in which case these may be weighted differently as a function of the time interval from PCT/DE 2003/003862 - 11 - 2002F11723WOIN the RICE predicted value to be estimated at that time. If the discrepancies between the actual data length of the respective data value and the data length predicted using the RICE predicted value are too great, the RICE method becomes ineffective. For this reason; the RICE method is used only up to a specific difference between the predicted data length and the actual data length. If this limit value is exceeded in the method, a maximum data length is used rather than the RICE predicted value. The fact that the limit value has been exceeded is identified by a particular value in the RICE codes, which cannot normally occur (limit value not exceeded). In order to explain the compression method further, the following text refers to three successive difference values which, by way of example, can be represented as a numerical sequence of a first binary value, (ll00lll0110) , of a second binary value (10110) and of a third binary value (1101111). The RICE predicts valve for the data length 15 obtained from the data length of the respective previous difference value of the corresponding binary value. The limit value or the maximum permissible difference between the actual data length of the respective binary value and the predicted RICE predicted value is assumed to be 4, with the maximum data length that can be transmitted being assumed to be 16. These assumptions result in the situation illustrated in Table 1 when the first binary value is transmitted with its actual data length. PCT/DE 2003/003862 - 13 - 2002F11723WOIN The bit sequence for the second and the third binary value is thus 100000G10110 00101111, For comparison purposes, it should be mentioned that the bit sequence for the second and the third binary value without the use of the RICE method but using the maximum data length that occurs (11, the data length of the first binary value) would have been 000000l0ll0 0000ll0llll, that is to say 22 bits, instead of the 20 bits produced by means of the RICE method. Figure 2 shows a schematic illustration, in order to explain a decompression method. In this case, the method explained in conjunction with Figure 1 is carried out in the reverse sequence. The compressed values are supplied (100) to a RICE apparatus 20. The difference values are recovered as the result of the RICE decoding method which can be carried out using the RICE apparatus 20, and the difference values are joined together (200) and (300), respectively, in an addition apparatus 21 with the predicted values for the predictor method, which are produced by means of a predictor apparatus 22, so that, in the end, the digital measurement data is emitted once again (400) . The start values for the pradicted values for the decompression process (see Figure 2) must either be agreed as fixed with the predicted values for the compression process (see Figure 1) , or must be transmitted together with the compressed values. In order to optimize the processing of the compressed values for practical applications in industrial plant, in particular energy plant, it is possible to provide for the compressed values to be provided with header values. The header data may, for example, comprise information about a data length of the difference values, the number of the difference values, a type and parameter of the respective predicted value for the predictor method, as well as a type and parameter of the PCT/DE 2003/003862 - 14 - 2002F11723WOIN respective RICE predicted value for the data length (RICE coding method). Furthermore, information may be included about the start values of the predicted values for the predictor method, and/or of the RICE, predicted value. The compression process with the aid of the method described in conjunction with Figure 1 can expediently be carried out immediately after the recording of the digital measurement values, in real time. Decompression of the compressed values can then advantageously be carried out for further use shortly before being used, in which case the (decompressed) digital measurement data can be used, for example, for display or simulation purposes. PCT/DE 2003/003862 - 15 - 2002F11723WOIN Patent Claims 1. A method for processing digital data values, in which a predictor method is used to form a difference value from the respective current digital data value and from a respective predicted value, and compressed values are produced from the difference values using a RICE method which follows the predictor method, characterized in that the respective predicted value is obtained by means of a signal model which describes the expected time profile of the digital data values. 2. The method as claimed in claim 1, characterized in that the signal model is determined on the basis of a sine, function or a cosine function with a constant period and ins respective harmonics, a decaying exponential function or a sine function with a decaying exponential function as an envelope curve. 3. A method for processing digital data values, in which a predictor method is used to form a difference value from the respective current digital data value and from a respective predicted value, and compressed values are produced from the difference values using a RICE method which follows the predictor method, characterised in that the digital data values are obtained from periodic signals and at least that digital data value which was recorded one period before the current digital data value is used as the respective predicted value associated with the respective current digital data value. PCT/DE 2003/003862 - 16 - 2002F11723WOIN 4. The method as claimed in one of the preceding claims, characterired in that, in the RICE method, an overflow, which is determined by subtraction of a data length (which has been predicted by means of a RICE prediction value) of the respective current difference value and a current data length of this difference value, is compared with a limit value which can be predetermined, and the difference value is emitted with a predetermined maximum data length as a compressed value if the overflow exceeds the limit value. 5. The method as claimed in one of the preceding claims, characterized in that the compressed values are transmitted via a data transmission pathy and the difference values are then recovered from the compressed values using a RlCE decoding method, and the digital data values are determined from the difference values and the respective predicted values by addition using a reverse predictor method. 6. The method as claimed in one of the preceding claims,. characterised in that the compressed digital measurement data is provided with header data, 7. The method as claimed in one of the preceding claims, characterized in that the digital data values are formed from input measurementt variables from in-field equipment. PCT/DE 2003/003862 - 17 - 2002F11723WOIN 8. The method as claimed in claim 7, characterized in that protective and/or control appliances in energy, installations are used as in-field appliances. The invention relates to a method for processing digital data valued according; to which a diffirential value is formed from a respective current digital data value and a respective predicted value with the aid of a predictor method, compressed 12 ves being generated from thedifferential valuse in a Rice process that succeeds the predictor method. The aim of the invention is to 12 Ty out said method relatively rapidly. To achieve this, the respective predicted value is obtained using a signal model, which discover the expected temporal progression of the digital data values.

Full Text

Description
Method for processing digital data values
The invention relates to a method for processing digital data
values, in which a predictor method is used to form a
difference value from the respective current digital data value
and from a respective predicted value, and compressed values
are produced from the difference values using a RICE method
which follows the predictor method.
A method such as this is known to those skilled in the act
from, for example, the Internet site which can be called up at
http;//www.monkeysaudio,com/theory.html (called up on
November 11, 2002). This site explains how digital audio data
can be compressed for the purpose of compression with the aid
of a predictor method and a subsequent RICE method. Those audio
data values which immediately precede the current audio data
values are in each case used to calculate predicted values
which are required for the predictor method.
Measurement values from processes that are being carried out,
for example from production or from conversion processes, are
also usually recorded in digital form for closed-loop and open-
loop control of industrial plant, and are, made availab1e as
output information for subsequent processing. The subsequent
processing may, for example, be readjustment of a processed
variable. In this case, it is often necessary Co transfer the
recorded measurement values. in the form of digital data values
or digital measurement data between different technical plant,
or to Store such measurement values for subsequent processing.
In order to reduce the amount of data

PCT/DE 2003/003862 -2-
2002F11723WOIN
required for transmission of the digital measurement data,
provision is made in conjunction with a closed-loop filling
level control system that is known from the document
WO 01/91031 for the digital measurement data to be compressed
by means of digital data compression, before being transmitted.
This makes it possible to reduce the bandwidth required for
data transmission.
One object of the invention is to specify a comparatively fast
method for compression of digital data values.
For a method of the type mentioned initially, this object is
achieved according to the invention by obtaining the respective
predicted value by means of a signal model which describes the
expected time profile of the digital data values.
The major, advantage which is achieved by the invention over the
prior art is that a significant reduction is. achieved in the
memory space requirement for the compressed values determined
from the digital data values, and at high speed, with the aid
of the proposed method. This is because, the digital data, values
can be compressed without any losses in real time, that is to
say immediately after they have been, recorded. The predicted
values for the predictor method are determined as quickly end
reliably as possible in that, when the pzedictor method is
being carried out, the predicted values are determined on the
basis of a signal model which is suitable for describing the
time profile of the digital data values. The use of a signsl
model such as this means that the predicted values are
available to form the difference values without the need.

PCT/DE 2003/003862 - 3 -
2002F11723WOIN
for a calculation in each case on the basis of previous data
values, and thus without any time delay.
The space required in the memory is leas than that for
uncompressed data values. The proposed method results in the
memory space requirement being reduced with minimal computation
power for processing of the digital data values, since only
simple computation operations are required, such as addition,
subtraction and bit manipulations. This makes it possible to
use the stated method for processing digital data values even
in technical equipment which has the computation power. by means
of suitable processors only to a restricted extent. The
requirement tor computation power for compression, which is not
available for other functions, is minimized.
The compression process carried out using the method according
to the invention means that, in, this case, only data which
contains no. information is removed. The digital measurement
data can thus be reproduced completely, that is to say without
any losses, during the decompression process,.
One embodiment of the method according to the invention
comprises the signal model being determined on the basis of a
sine function or a cosine function with a constant period and
its respective harmonics, a decaying exponential function or a
sine function with a decaying exponential function as an
envelope curve. This advantageously allows the respective
predicted value to be determined particularly quickly, since no
other calculations need be carried out.

PCT/DE 2003/003862 4 -
2002F11723WOIN
A further solution for the object mentioned above comprises
on the basis of the method of the type mentioned initially
the digital data values being obtained, according to the
invention from periodic signals and at least that digital data
value which was recorded one period befor the current digital
data value being used as the respective predicted value
associated with the respective current digital data value. The
required computation power can advantageously be kept low by
the use for the predicted value of the respective date value
recorded one period before a current data value. No additional
computations tieed be carried, out, so that the method can be
carried out at high speed.
In order to improve the accuracy of the respective predicted
value, it is also possible. to use, in addition, to the digital
data value recorded precisely one period before the respective
current digital data value, further digital data values, which.
were recorded an integer number of periods before the
respective carrent digital data value, in order to form the
respective predicted value. These can be linked to one another
by simple mathematical functions, such as averaging and/or
weighting, so that the computation power required for this
purpose can be kept low.
One advantageous further development of the method according to
the invention also provides that, in the RICE method, an
overflow, which is determined by subtraction of a data length
(which has been predicted by means of a RICE prediction value;
of the respective current difference value and a current data
length of this difference value, is. compared with a limit value
which can be predetermined, and the difference value is"emitted
with a predetermined maximum data length as a compressed value

PCT/DE 2003/003862 - 5 -
2002F11723WOIN
if the overflow exceeds the limit value. The major advantage of
this development is that, in the case of successive difference
values whose data length fluctuates to a major extent, the RICE
method is used only when it can be used effectively, that is to
say with Little overflow. If there is a. large overflow, the
respective compressed value is emitted with a maximum
predetermined data length.
Furthermore, one advantageous further development of the method
according to the invention provides that the compressed values
are transmitted via a data transmission path, and the
difference values are then recovered from the compressed values
using a RICE decpding method, and the digital data values are
determined from the difference values and the respective
predicted values by addition using a reverse predictor method.
Transmission of the compressed measurement values via a data
transmission path in this case covers both wire-based
transmission methods and wireless transmission methods, such as
radio transmissions, According to this development,
decompression. of the compressed values can advantageously be
carried out at a point which is remote from the point at which
they were compressed. For example, the compressed value can be
transmitted from an in-field appliance in an industrial plant
via a data bus to a central computer, which evaluates the
digital data values once the compressed values have been
decompressed.
In order to simplify subsequent processing of the compressed
values or to transmit additional information, one preferred
embodiment of the invention allows the compressed Values to be
provided with header data.

PCT/DE 2003/003862 - 6 -
2002F11723WOIN
According to a further advantageous embodiment of the method
according to the invention, the data values are formed from
input measurement variables from in-field appliances. In this
context, by way of example, in-field appliances are protective
or control appliances such as those which are normally used in
industrial, for example energy, chemical or petrochemical
plant. In comparison to audio technology, in which a virtually
indefinite time is available for compression of the digital
data values, the comparison prscess in in-field appliances
must be carried out particularly quickly, owing to the
restricted computation and storage capacities. The method
according to the invention can thus be used particularly
advantageously in this case.
Protective and/or control appliances in energy plant can also
advantageously be used as in-field appliances.
The invention will be explained, in more detail in the following
text using exemplary embodiments. In this case, in the figures:
Figure 1 shows a schematic illustration in order to explain a
method for compression of the digital data values;
and
Figure 2 shows a schematic illustration in order to explain a
method for decompression of compressed- digital data
values.
Figure 1 shows a schematic illustration in order to explain a
method for compression, using digital data values which have
been recorded using an in-field appliance that is not
illustrated.

PCT/DE 2003/003862 -7-
2002F11723WOIN
Since the digital data values in the present exemplary
embodiment are intended to be formed from (analog) input
measurement variables of the in-field appliance, they are
referred to in the following text as digital measurement data.
A measurement device 1 is used to record measurement values,
which are converted to digital measurement data in the normal
way. First of all, a predictor method is carried out in order
to process the digital measurement data. The predictor method
is part of a. processing process for compression of the digital
measurement data. in this case, predicted values for the
digital measurement data are determined by means of a predictor
device 2.
The predictor method first of all reduces the data length (the
number of bits) which is required to represent the digital
measurement data, by forming difference values by subtraction
from the digital measurement data and suitable predicted
values. The predicted values that are used should in this case
be as close as possible to the real digital measurement data,
so that the difference values which result from the subtraction
process are as small as possible. The further processing is
then carried out using difference values whose required data
length is considerably less than that of the digital
measurement data.
The predicted values for the predictor method can be formed
using a signal modal which, for example, is based, on a sine
function or a cosine function, as well as their respective
harmonics, a decaying exponential function or a sine function
with a decaying exponential function as an envelope curve. In
this case, these are in each case functions which are
represented and processed computationally with little effort.

PCT/DE 2003/003862 -8-
2002F11723WOIN
and are in some cases periodic. Previous digital measurement
data which has been recorded before the digital measurement
data to be compressed at that time is taken into account in the
creation of the signal model for determination of the predicted
models. Characteristic properties, such as the amplitude,
period duration and decay behavior, of the previous digital
measurement data ace used to form the signal model. Properties
such as these can be calculated by simple computation
operations and nevertheless allow the predicted values to be
determined relatively reliably.
In the simplest case, the previous digital measurement data.
from precisely one previous recording period can be adopted
unchanged as predicted values. The most recently described
procedure, in which reference is made to a previous recording
period, is particularly feasible when the recorded measurement
values and thus the digital measurement data have or has a
periodic behavior, as is frequently the case, for example, in
energy plant. When the digital measurement data is processed
for the first time using the predictor method, a start value
must be provided for the first predicted value.
As is shown in Figure 1, the digital measurement data is
transmitted (10) from the measurement device 1 and the
respective determined predicted values are transmitted (20)
from the predictor apparatus 2 to a subtraction device 3,
respectively, in which the predicted values are subtracted from
the respectively associated digital measurement data. The
subtraction process results in difference values, which are
then passed (30) to on apparatus A in order to carry out a RICE
method.

PCT/DE 2003/003862 - 9 -
2002F11723WOIN
In the RICE method, which ia known per se and whose details can
be found from the references by a person skilled in the art
(see, for example, http://WWW.monkeysaudio,com/theory.html
[called up on November 13, 2002], the data length of the
difference values which are produced by the predictor method
are reduced, so that, in the end, compressed values are
produced and emitted (40) .
The same data length could, in each case, be used for
transmission and storage of the difference values. These would
have to correspond at Least to the maximum, possible data length
of a difference value. However, since the required data length
is generally not constant in the case of successive difference
values, but can fluctuate from one difference value to the
next, a not inconsiderable amount of memory space would the
wasted for the transmission and storage of the difference
values with the maximum data length. This will be explained in
more detail using the following example: let us assumed that
three digitally represented difference values are to be
transmitted, to be precise 11010110, 1101 and 10110. If a
constant date length were used for transmission, the maximum
occurring data length of the difference values would have to be
used, that is to say in this case 8 (data length of first
difference value}. In consequence, the three difference values
would have to be transmitted in the form
110101100000110100010110. The zeroes which are inserted in the
maximum data length for transmission of the smaller difference
values (1101, 10110) waste unnecessary memory space, since no
additional in format ion is transmitted.
The RICE method now provides an algorithmic method for
scrambled difference values which result from the predictor
method in a suitable manner, that is to say in a manner which

PCT/DE 2003/003862 - 10 -
2002F11723WOIN
is optimized with regard to memory space, to form compressed
values. The basic idea for the RICE method is the compression
of the difference values using a data Length which is matched
to the respective difference value. A so-called RICE code is
introduced in order to separate successive compressed values,
which are obtained from the difference values, and in order to
scramble information which it may possibly not be possible to
represent with an excessively small data length. This will be
explained for the following text.
The BICE method requires a RICE predicted value for the
expected data length of the respective subsequent difference
value. The compressed values which are produced by means of the
RICE method are then in general stored using the data length
predicted by means of the RICE predicted value. If the data
length of the difference values which are produced using the
previous predictor method is greater than the data length
predicted using the RICE predicted value, the overflow (the
most significant bits which can no longer be represented in the
predicted data length} are scrambled in the RICE code. The RICE
code comprises a number of binary values 0, with this number
resulting directly from the overflow, and a final binary value
1. The RICE code and a resultant value which is obtained, taking
into account the predicted data length from the respective
difference value, are attached directly to one another, in
order to obtain a compressed value.
The RICE, predicted value for the data length is obtained from
the values for the data length of a specific number of previous
digital measurement data items, in which case these may be
weighted differently as a function of the time interval from

PCT/DE 2003/003862 - 11 -
2002F11723WOIN
the RICE predicted value to be estimated at that time.
If the discrepancies between the actual data length of the
respective data value and the data length predicted using the
RICE predicted value are too great, the RICE method becomes
ineffective. For this reason; the RICE method is used only up
to a specific difference between the predicted data length and
the actual data length. If this limit value is exceeded in the
method, a maximum data length is used rather than the RICE
predicted value. The fact that the limit value has been
exceeded is identified by a particular value in the RICE codes,
which cannot normally occur (limit value not exceeded).
In order to explain the compression method further, the
following text refers to three successive difference values
which, by way of example, can be represented as a numerical
sequence of a first binary value, (ll00lll0110) , of a second
binary value (10110) and of a third binary value (1101111). The
RICE predicts valve for the data length 15 obtained from the
data length of the respective previous difference value of the
corresponding binary value. The limit value or the maximum
permissible difference between the actual data length of the
respective binary value and the predicted RICE predicted value
is assumed to be 4, with the maximum data length that can be
transmitted being assumed to be 16. These assumptions result in
the situation illustrated in Table 1 when the first binary
value is transmitted with its actual data length.

PCT/DE 2003/003862 - 13 -
2002F11723WOIN
The bit sequence for the second and the third binary value is
thus 100000G10110 00101111, For comparison purposes, it should
be mentioned that the bit sequence for the second and the third
binary value without the use of the RICE method but using the
maximum data length that occurs (11, the data length of the
first binary value) would have been 000000l0ll0 0000ll0llll,
that is to say 22 bits, instead of the 20 bits produced by
means of the RICE method.
Figure 2 shows a schematic illustration, in order to explain a
decompression method. In this case, the method explained in
conjunction with Figure 1 is carried out in the reverse
sequence. The compressed values are supplied (100) to a RICE
apparatus 20. The difference values are recovered as the result
of the RICE decoding method which can be carried out using the
RICE apparatus 20, and the difference values are joined
together (200) and (300), respectively, in an addition
apparatus 21 with the predicted values for the predictor
method, which are produced by means of a predictor apparatus
22, so that, in the end, the digital measurement data is
emitted once again (400) . The start values for the pradicted
values for the decompression process (see Figure 2) must either
be agreed as fixed with the predicted values for the
compression process (see Figure 1) , or must be transmitted
together with the compressed values.
In order to optimize the processing of the compressed values
for practical applications in industrial plant, in particular
energy plant, it is possible to provide for the compressed
values to be provided with header values. The header data may,
for example, comprise information about a data length of the
difference values, the number of the difference values, a type
and parameter of the respective predicted value for the
predictor method, as well as a type and parameter of the

PCT/DE 2003/003862 - 14 -
2002F11723WOIN
respective RICE predicted value for the data length (RICE
coding method). Furthermore, information may be included about
the start values of the predicted values for the predictor
method, and/or of the RICE, predicted value.
The compression process with the aid of the method described in
conjunction with Figure 1 can expediently be carried out
immediately after the recording of the digital measurement
values, in real time. Decompression of the compressed values
can then advantageously be carried out for further use shortly
before being used, in which case the (decompressed) digital
measurement data can be used, for example, for display or
simulation purposes.

PCT/DE 2003/003862 - 15 -
2002F11723WOIN
Patent Claims
1. A method for processing digital data values, in which a
predictor method is used to form a difference value from
the respective current digital data value and from a
respective predicted value, and compressed values are
produced from the difference values using a RICE method
which follows the predictor method,
characterized in that
the respective predicted value is obtained by means of a
signal model which describes the expected time profile of
the digital data values.
2. The method as claimed in claim 1,
characterized in that
the signal model is determined on the basis of a sine,
function or a cosine function with a constant period and
ins respective harmonics, a decaying exponential function
or a sine function with a decaying exponential function as
an envelope curve.
3. A method for processing digital data values, in which a
predictor method is used to form a difference value from
the respective current digital data value and from a
respective predicted value, and compressed values are
produced from the difference values using a RICE method
which follows the predictor method,
characterised in that
the digital data values are obtained from periodic signals
and at least that digital data value which was recorded
one period before the current digital data value is used
as the respective predicted value associated with the
respective current digital data value.

PCT/DE 2003/003862 - 16 -
2002F11723WOIN
4. The method as claimed in one of the preceding claims,
characterired in that,
in the RICE method, an overflow, which is determined by
subtraction of a data length (which has been predicted by
means of a RICE prediction value) of the respective
current difference value and a current data length of this
difference value, is compared with a limit value which can
be predetermined, and the difference value is emitted with
a predetermined maximum data length as a compressed value
if the overflow exceeds the limit value.
5. The method as claimed in one of the preceding claims,
characterized in that
the compressed values are transmitted via a data
transmission pathy and the difference values are then
recovered from the compressed values using a RlCE decoding
method, and the digital data values are determined from
the difference values and the respective predicted values
by addition using a reverse predictor method.
6. The method as claimed in one of the preceding claims,.
characterised in that
the compressed digital measurement data is provided with
header data,
7. The method as claimed in one of the preceding claims,
characterized in that
the digital data values are formed from input measurementt
variables from in-field equipment.

PCT/DE 2003/003862 - 17 -
2002F11723WOIN
8. The method as claimed in claim 7,
characterized in that
protective and/or control appliances in energy,
installations are used as in-field appliances.

The invention relates to a method for processing digital data valued according; to which a diffirential value is formed
from a respective current digital data value and a respective predicted value with the aid of a predictor method, compressed 12 ves
being generated from thedifferential valuse in a Rice process that succeeds the predictor method. The aim of the invention is to 12 Ty
out said method relatively rapidly. To achieve this, the respective predicted value is obtained using a signal model, which discover
the expected temporal progression of the digital data values.

Documents:

« Previous Patent

Next Patent »

Patent Number

218533

Indian Patent Application Number

01304/KOLNP/2005

PG Journal Number

14/2008

Publication Date

04-Apr-2008

Grant Date

02-Apr-2008

Date of Filing

06-Jul-2005

Name of Patentee

SIEMENS AKTIENGESELLSCHAFT

Applicant Address

WITTELSBACHERPLATZ 2, 80333 MUNCHEN, GERMANY

Inventors:

#	Inventor's Name	Inventor's Address
1	TORSTEN KERGER	LENTHER STEIG 12, 13629 BERLIN, GERMANY
2	ROLAND KIND	PASTEURSTR. 19, 14482 POTSDAM, GERMANY

PCT International Classification Number

H03M7/40; G10L19/00; H03M7/30

PCT International Application Number

PCT/DE2003/003862

PCT International Filing date

2003-11-19

PCT Conventions:

#	PCT Application Number	Date of Convention	Priority Country
1	102 58 472.9	2002-12-09	Germany