DE102022120401A1 - Method and device for encoding a data stream, data storage device with such a device, method for decoding codes and computer-readable memory - Google Patents

Abstract

The application relates to a method for coding a data stream (DS), the data stream (DS) having bits in a continuous sequence, comprising: • receiving a first number of bits of the data stream (DS), • segmenting the first number of bits, wherein There are segments which correspond to packets (P) which have a length which corresponds to a second number of bits, and where there is at least one short segment which has less than the second number of bits, • padding the at least one short segment a packet (P) by adding padding bits (X), so that the combined length of short segment and padding bits (X) results in the second number of bits, • Encoding the packets (P), each packet (P) being one of the second Number of bits is mapped to a respective code (C) by the coding, the second number of bits being chosen so that the respective code (C) can be efficiently calculated and / or stored. The application further relates to a device (10 ) for encoding a data stream (DS) and a method for decoding codes (C).

Description

Technical area

The application relates to a method for encoding a data stream or a device for encoding such a data stream or a data storage device with such a device or a method for decoding codes and a computer-readable memory.

background

When encoding data, a specific sequence of bits represents a symbol of an alphabet, which is mapped to a code through the encoding.

The object of the present application is to more efficiently store the data transmitted in the data stream from a data stream.

This task is solved by the combinations of features of the independent patent claims according to this application.

Overview

• • Empfangen einer ersten Anzahl von Bits des Datenstromes,
• • Segmentieren der ersten Anzahl der Bits, wobei es danach Segmente gibt, die Paketen entsprechen, die eine Länge aufweisen, die einer zweiten Anzahl von Bits entspricht, und wobei es zumindest ein Kurz-Segment gibt, das weniger als die zweite Anzahl Bits aufweist,
• • Auffüllen des zumindest einen Kurz-Segments zu einem Paket durch Hinzufügen von Füllbits, sodass die kombinierte Länge des Kurz-Segments und der Füllbits die zweite Anzahl Bits ergibt,
• • Codieren der Pakete,

wobei jeweils ein Paket der zweiten Anzahl von Bits durch die Codierung auf einen Code abgebildet wird, wobei die zweite Anzahl von Bits so gewählt ist, dass sich der Code effizient berechnen und/oder speichern lässt.According to this application, a method for encoding a data stream is provided, the data stream having bits in a continuous sequence. This procedure has the following procedural steps:

• • Receiving a first number of bits of the data stream,
• • Segmenting the first number of bits, thereafter having segments corresponding to packets having a length corresponding to a second number of bits, and wherein there being at least one short segment having less than the second number of bits,
• • Padding the at least one short segment into a packet by adding padding bits so that the combined length of the short segment and the padding bits results in the second number of bits,
• • Encoding the packets,

wherein in each case a packet of the second number of bits is mapped onto a code by the coding, the second number of bits being chosen such that the code can be efficiently calculated and/or stored.

• eine Empfangseinrichtung zum Empfang einer ersten Anzahl von Bits des Datenstromes,
• einen Codierer, welcher eingerichtet ist,
  • eine erste Anzahl von Bits zu Segmentieren, wobei es danach Segmente gibt, die Paketen entsprechen, die eine Länge aufweisen, die einer zweiten Anzahl von Bits entspricht, und wobei es zumindest ein Kurz-Segment gibt, das weniger als die zweite Anzahl Bits aufweist,
  • das zumindest eine Kurz-Segment durch Hinzufügen von Füllbits zu einem Paket aufzufüllen, sodass die kombinierte Länge des zumindest einen Kurz-Segments und der Füllbits die zweite Anzahl von Bits ergibt, und
  • die Pakete zu codieren, wobei jeweils ein Paket mit der zweiten Anzahl von Bits durch die Codierung auf einen jeweiligen Code abgebildet wird, wobei die zweite Anzahl von Bits so gewählt ist, dass sich der jeweilige Code effizient berechnen und/oder speichern lässt.

Furthermore, according to this application, a device for encoding a data stream is provided, the data stream having bits in a continuous sequence, comprising:

• a receiving device for receiving a first number of bits of the data stream,
• an encoder which is set up
  • to segment a first number of bits, thereafter being segments corresponding to packets having a length corresponding to a second number of bits, and wherein there being at least one short segment having less than the second number of bits,
  • padding the at least one short segment by adding padding bits to a packet such that the combined length of the at least one short segment and the padding bits results in the second number of bits, and
  • to encode the packets, whereby one packet with the second number of bits is mapped to a respective code through the coding, the second number of bits being chosen so that the respective code can be efficiently calculated and/or stored.

Furthermore, according to this application, a data storage device is provided. The data storage device has the device for coding and a memory for storing the generated codes.

Furthermore, a method for decoding codes according to this application is provided, the codes being generated by the method for encoding the data stream.

Furthermore, a computer-readable memory is provided which contains program instructions which, when executed on a data processing system, cause the method for encoding the data stream to be carried out.

According to this application, a method for encoding a data stream or a device for encoding the data stream is proposed, wherein the data stream is segmented, i.e. divided piecewise. These segments are then filled to a predetermined number of bits if they do not already have this predetermined number of bits. This usually applies, for example, to the last segment in a data stream. However, it is possible for segments other than the last segment to be populated. Furthermore, it is possible for segments to be filled up accordingly by analyzing their bits in order to be optimally coded.

As the predetermined, therefore second number of bits for the segments, a number can be selected to which the code generated by the coding can be easily calculated and/or to which the code generated by the coding can be easily saved, for example as an integer value. Can calculate and/or store well e.g. B. mean efficient calculation and/or storage, in which less computing power is required for calculation or less memory is required for storage.

Now each segment can be coded and thus mapped to a code so that the code can be calculated and/or saved efficiently. For example, a so-called arithmetic coding can be used for this coding. This arithmetic coding has the advantage that its use can be used to approximate the so-called entropy of a message (here defined by a segment). This cannot be achieved by other codes, for example Huffman coding.

Entropy is a term in information theory and describes the minimum number of bits that are necessary per symbol or segment to transmit information. This entropy therefore represents a lower limit for the number of bits for a message. The goal, for example, is to get as close as possible to this entropy with the arithmetic code and thus ultimately be able to save a message in as space-saving a manner as possible.

For example, when simulating technical devices in a so-called hardware-in-the-loop arrangement (HIL), in which, for example, a control device is tested using simulated connected components, a large number of data, which are available in a continuous data stream, are generated . Another use case is the event-related or continuous storage of data in the vehicle for later analysis or playback of the data to test control devices.

Due to increasing complexity, for example in automated or autonomous driving functions or also in battery management systems, the number of data that is generated during such a simulation or that has to be recorded or stored in the vehicle continues to increase. Storage is necessary for subsequent analysis of this data. It is possible to store this data in a space-saving and at the same time low-loss or even loss-free manner through coding and compression. If the provided memory is not sufficient, additional memory can be provided.

A goal of this application is to avoid such additional storage as much as possible by using optimal coding. This is achieved by the method for encoding a data stream or the corresponding device.

The coding of the data stream is a mapping of the respective existing segment to a respective code. If possible, this code should have fewer bits than the segment. This means that the data stream can then be stored in a space-saving manner using the method according to the application or the device according to the application.

The data stream is a continuous sequence of bits. So there are originally no segments or packets available in the data stream. Therefore, the method or device according to this application will, after receiving the first number of bits of the data stream, segment this first number of bits in order to thereby provide the packets for coding. The first number of bits can be a few hundred or thousand bits.

A second number of bits is specified for each segment; the number is usually the same for each segment. The segments then have the same second number. However, there may be a so-called short segment, particularly at the end of the first number of bits, which does not have this second number of bits. This segment has a smaller number than this second number.

Therefore, according to this application, this or several of these short segments within the first number of bits are now filled up to the second number of bits by adding so-called filling bits. I.e. the combined length of the short segment(s) and the added padding bits equals the second number of bits. The filler bits are unlikely to add any meaningful information, but this error caused by the filler bits can be taken into account or checked for plausibility in the evaluation. Error correction can also be used for this.

If the short segments, preferably all short segments, are filled to the second number of bits by the filler bits, the packets are encoded, for example by arithmetic coding. Segments with a length equal to the second number of bits are now referred to as packets. Through the coding, the second number of bits is mapped to a code that is saved, for example. This second number of bits is determined so that coding runs optimally with this second number. This then emerges from the dependent claims.

The device according to this application is designed, for example, as software and/or hardware. If at least part of the device is designed as hardware, for example, then this hardware can be, for example, an FPGA, i.e. a field programmable gate array. Processors or controllers of various types are also possible.

The receiving device can be, for example, a memory or register that receives the data stream up to the first number and then the segmentation is carried out via the encoder as discussed above. The receiving device continually records the first number of bits from the data stream and then passes them on for further steps.

The encoder also performs the padding of the short segments by adding padding bits to such a packet with the second number of bits. Ultimately, the encoder encodes the packets. And the respective packets are mapped to a respective code with their second number of bits.

This code can be efficiently calculated and/or saved as described above.

The data storage device mentioned above is, for example, a computer which has the coding device described above and a memory for storing the generated codes. This memory can usually be an electronic and/or optical and/or magnetic memory.

Furthermore, a method for decoding these codes is also provided in order to be able to process the decoded codes, i.e. the packets, for further evaluation. A possible error that arises from filling can then be taken into account accordingly. However, no further action may be necessary if this error only has a very minor impact on the analysis of the data stream.

Furthermore, a computer-readable memory is provided which contains program instructions which are executed on the data processing system and which represent the method for encoding the data stream described above.

The measures and developments listed in the dependent claims make advantageous configurations of the method for coding the data stream possible.

In one embodiment, the short segment is at the end of the data stream. This allows the method to accordingly focus on the last segment after segmentation for padding into a packet by adding padding bits. So no further steps are necessary to evaluate the individual segments in terms of their length. On the other hand, it may be possible to also exchange bits in the segments to enable optimal coding.

In addition, it is proposed that the at least one short segment be arranged within the first number of received bits in such a way that the code for the resulting packet (filled with corresponding filler bits) and/or at least one further packet can be efficiently calculated or stored. A corresponding optimized segmentation is preferably determined and carried out by the encoder.

In one embodiment, the second number of bits is selected so that the code for the packets can be stored efficiently in binary format, with the coding in particular generating an integer number as the code for the respective packets. This is particularly advantageous when using so-called arithmetic coding.

The coding can in particular be an arithmetic coding and the second number of bits can correspond to 2 ⁿ -1, where n is a natural number. This determination of the second number of bits makes arithmetic coding particularly advantageous.

In addition, it is suggested that the padding bits be chosen so that the code for the associated packet can be efficiently calculated or stored. This can then mean lossless compression because it then enables storage to be as lossless as possible.

In one embodiment, the filler bits are selected so that the code for the associated packet can be efficiently stored in binary format, with an integer code being generated in particular by encoding the associated packet.

In addition, it is suggested that the code generated by the coding is output and/or stored.

In one embodiment, steps of the method are carried out in parallel, with the segmentation of already received bits and the encoding of already formed packets taking place in particular during the reception of the bits. Through such a parallelization of procedural tasks According to this application, the method can be carried out more quickly.

Fiaurenlist

Exemplary embodiments are shown in the drawing and are explained in more detail in the following description.

• 1 ein Flussdiagramm des Verfahrens,
• 2 einen kontinuierlichen Datenstrom, der gemäß dieser Anmeldung segmentiert wird und
• 3 ein Blockschaltbild der Einrichtung.

Show it

• 1 a flowchart of the procedure,
• 2 a continuous data stream that is segmented in accordance with this application and
• 3 a block diagram of the device.

The same reference numbers are used in the figures for the same or similar elements.

Fiauren description

1 shows the process in a flowchart. A data stream DS is processed in the first method step S1. In this first method step S1, the first number of bits of the data stream DS is received. It should be noted that the data stream DS has a continuous sequence of bits. This means that in method step S1 the device continuously receives the first number of bits and then passes them on.

This first number of bits of the data stream DS is segmented in the second subsequent method step S2. There are segments that already correspond to the predetermined packets P, which have a length that corresponds to a second number of bits. However, it can happen that at least one short segment is formed that has fewer than the second number of bits assigned to a packet P. It is possible that the short segment or segments are identified by the method or device and marked accordingly, or it is clear from the outset that the short segment is usually the last segment in the first number of bits acts.

The packets P, i.e. the segmented data stream DS, is then transferred to method step S3. Here, the at least one short segment is filled up to form a packet P by adding filler bits X, so that the combined length of the short segment and the respective filler bits X results in the second number of bits. At the end of method step S3, the received first number of bits of the data stream DS is now available in the form of packets P. I.e. all segments are now packets P. Packets P are defined by a length corresponding to the second number of bits. In the subsequent method step S4, these packets P can be encoded. A packet P of the second number of bits is mapped to a respective code C through this coding.

In method step S5, the code C is then stored in a memory M. This code C would therefore be output in S5.

2 now shows an exemplary continuous data stream DS which is segmented according to the method described in this application and where the last segment has been correspondingly filled with filler bits X. However, this is only to be understood as an example; the filler bits X could also be elsewhere in the data stream DS, i.e. not necessarily always in the last segment. In the present case, the second number of bits is given as 31 bits, each identified as packets P. However, since the last segment, i.e. the short segment in this case, had fewer than 31 bits, it was filled to 31 bits by seven filler bits X. The number 31 corresponds to 2 ⁵ - 1.

3 now shows the device with the connected components in a block diagram. The data storage device 50 has a device for encoding 10 and a device for decoding 20. Furthermore, a memory M is provided in which the codes C are stored or read out again.

The coding device 10 stores the codes C in the memory M, while the decoding device 20 reads these codes C from the memory M. The coding device 10 has a receiving device 12 for receiving the data stream DS. This means that the first number of bits of the data stream DS is received. This can be memory or registers or similar.

The encoder 14 now segments this first number of bits into segments. The segments correspond to the packets P if they have the length specified by the second number of bits, for example 31 bits. In this case, the packets P can then be encoded. In addition, there is at least one short segment that does not have this second number of bits, but fewer. The encoder 14 fills this at least one short segment by adding padding bits X to a packet P, so that the combined length of the short segment and the padding bits X results in the second number of bits. The segments are then encoded, as they are now only packets P. This coding can be, for example, arithmetic coding. I.e. the second The number of bits is mapped to a code C through the coding. This code C is efficiently stored in the memory M by the writing device 16.

The decoding device 20 can then read the code C out of the memory M again using the reading device 22 and assign it to the decoder 24. The decoder 24 decodes the codes C back into the packets P and sequences the packets P into a data stream DS, which is output again by the output device 26. The data stream DS can then be analyzed. The decoded data can optionally also be examined statically.

For example, the simulations mentioned above can be used to examine how the device that was being tested behaved.

Reference symbol list

D.S

data stream

C

code

P

package

X

Fill bit

10

Coding facility

12

Reception facility

14

Coder

16

Writing facility

20

Decoding facility

22

Reading facility

24

decoder

26

Output facility

50

Data storage device

M

Storage

S1-S5

Procedural steps

Claims (13)

Method for encoding a data stream (DS), the data stream (DS) having bits in a continuous sequence, comprising: • Receiving a first number of bits of the data stream (DS), • Segmenting the first number of bits, where there are segments corresponding to packets having a length equal to a second number of bits, and where there is at least one short segment having less than the second number (P) of bits , • Filling the at least one short segment into a packet (P) by adding padding bits (X), so that the combined length of the short segment and padding bits (X) results in the second number of bits, • Encoding the packets, whereby one packet (P) of the second number of bits is mapped to a respective code (C) through the coding, whereby the second number of bits is chosen such that the respective code (C) is calculated efficiently and/or can be saved. Procedure according to Claim 1 , where there is a short segment at the end of the data stream. Procedure according to Claim 1 or 2 , wherein at least one short segment is arranged within the first number of received bits in such a way that the code (C) can be efficiently calculated and/or stored for the resulting packet (P) and/or at least one further packet (P). Method according to one of the preceding claims, wherein the second number of bits is selected such that the code (C) for the packets (P) can be stored efficiently in binary format, in particular through the coding an integer number as the code (C) for the respective packages (P) are generated. Procedure according to Claim 4 , where the encoding is an arithmetic encoding and the second number of bits corresponds to 2 ⁿ -1, where n is a natural number. Method according to one of the preceding claims, wherein the fill bits (X) are selected so that the code (C) for the associated packet (P) can be efficiently calculated and/or stored. Procedure according to Claim 6 , whereby the filler bits (X) are chosen so that the code (C) for the associated packet (P) can be stored efficiently in binary format, with an integer code (C) being generated in particular by encoding the associated packet (P). . Method according to one of the preceding claims, wherein the code (C) generated by the coding is output and/or stored. Method according to one of the preceding claims, wherein steps of the method are carried out in parallel, with the segmentation of already received bits and the encoding of already formed packets taking place in particular during the reception of bits. Device (10) for coding a data stream, the data stream (DS) having bits in a continuous sequence, comprising: a receiving device (12) for receiving a first number of bits of the data stream (DS), an encoder (14) which is set up to segment the first number of bits, where there are segments which correspond to packets which have a length which corresponds to a second number of bits, and where there is at least one short segment which less than the second number (P) of bits, to fill up the at least one short segment by adding padding bits (X) to a packet (P), so that the combined length of short segment and padding bits (X) is the second number of bits results, and to encode the packets, each packet (P) of the second number of bits being mapped to a respective code (C) by the coding, the second number of bits being chosen so that the respective code (C ) can be calculated and/or stored efficiently. Data storage device (50) with a device (10) for coding Claim 10 , and a memory (M) for storing the generated codes (C). Method for decoding codes, which is carried out by a method according to one of Claims 1 until 9 were generated. Computer-readable memory which contains program instructions which, when executed on a data processing system, enable the execution of a method according to one of the Claims 1 until 9 cause.

Images