Title of Invention	"A METHOD AND DEVICE FOR CHANNEL ENCODING OR DECODING OF INFORMATION STRUCTURED IN FRAMES IN A MOBILE RADIO SYSTEM"
Abstract	A method for channel coding of information structured in frames in a mobile radio system, in which first information items (db) and second information items (mb ) for describing the coding of first information items are contained within a frame, characterized in that the second information items (mb) and a first portion (dbl) of the first information items (db) are uniformly channel-coded irrespective of the nature of the cooling for different types of coding.

Title of Invention

"A METHOD AND DEVICE FOR CHANNEL ENCODING OR DECODING OF INFORMATION STRUCTURED IN FRAMES IN A MOBILE RADIO SYSTEM"

Abstract

A method for channel coding of information structured in frames in a mobile radio system, in which first information items (db) and second information items (mb ) for describing the coding of first information items are contained within a frame, characterized in that the second information items (mb) and a first portion (dbl) of the first information items (db) are uniformly channel-coded irrespective of the nature of the cooling for different types of coding.

Full Text	Description Method and arrangement for channel coding and decoding of information structured in frames The invention relates to a method and an arrangement, for channel coding and decoding of information structured in frames, in particular for the purposes of adaptive multirate coding. Source signals and source information such as voice, audio, picture and video signals virtually always contain statistical redundancy, that is to say redundant information. This redundancy can be greatly reduced by source coding, thus allowing efficient, transmission and storace of the source signal. This reduction in redundancy gets rid of signal contents which were redundant before transmission and are based on prior knowledge of, for example, statistical parameters in the signal profile. The bit rate of the source-coded information (source bits or data bits) is also referred to as the source bit rate. During the source decoding after transmission, these components are added to the signal once again, so that virtually no loss of quality can be verified objectively. On the other hand, it is normal for signal transmission to add redundancy deliberately by means of channel coding once again, in order largely to correct for the influence of channel interference on the transmission. Additional redundant bits thus make it possible for the receiver or decoder to identify errors, and possibly also to correct them. The bit rate of the channel-coded information is also referred to as the gross bit rate. AMENDED SHEET -lain order to allow information, in particular voice data, picture data or other user data, to be AMENDED SHEET - 2 - transmitted, as efficiently as possible by means of the limited transmission capacities of a transmission medium, in particular a radio interface, this information to be transmitted is thus compressed by means of source coding before transmission, and is protected against channel errors by means of channel coding. Various methods are in each case known for doing this. For example, in the GSM {Global System for Mobile Communication) System, voice can be coded by means of a full rate voice codec a half rate voice codec, or an enhanced full rate voice codec. For the purposes of this application, the terms voice codec or coding also refer to a method for encoding and/or for corresponding decoding, which method may also cover sources and/or channel coding, and can also be applied to data other than voice data. In the course of the further development of the European mobile radio standard GSM, a new standard for coded voice transmission is being developed, which will allow the overall data rate and the splitting of the data rate between the source coding and channel coding to be set adaptively depending on the channel state and the network conditions (system load). Instead of the voice codecs described above, which have a fixed source bit rate, new voice codecs are intended to be used for this purpose, whose source bit rate is variable and is matched to changing frame conditions for information transmission. The main aims of such AMR (Adaptive Multirate) voice codecs are to achieve landline network quality for voice transmission in various channel conditions, and to ensure optimum distribution of the channel capacity, taking account of specific network - 2a - parameters. After carrying out a conventional source coding method, the compressed information is in structured form, in frames, in which case the source bit rate may differ from frame to frame depending on the code mode being used. AMENDED SHEET - 3 - In order to achieve standard gross bit rates, the information contained within a frame is channel-coded in a different manner depending on the source bit rate or code mode, in particular at a different rate, in such a manner that the gross bit rate after channel coding corresponds to the selected channel mode (half rate or full rate). For example, such an AMR voice codec can operate using the half rate (HR) channel in good channel conditions and/or in highly loaded radio cells. When the channel conditions are poor, a dynamic change should be made to the full rate (FR) channel, and vice versa. Within such a channel mode (half rate or full rate), various code modes are available for different voice and channel coding rates, and these are likewise selected to match the channel quality (rate adaptation) . In the process, the gross bit rate after channel coding remains constant within one channel mode (22.8 kbps for the full rate channel FR and 11.4 kbps for the half-fate channel HR) . This is intended to result in the best voice quality, taking account of the changing channel conditions. Thus, with such adaptive coding, different rates are used for voice coding (variable source bit rate) depending on the channel conditions in a transmission path, on the requirements for specific network parameters, and depending oh the voice. Since the gross bit rate after channel coding is intended to remain constant, an appropriately adapted variable number of error protection bits are added during channel coding. In order to decode such variably coded information after transmission, it is helpful for information about the coding method used at the transmission end, in particular the source bit rate and/or the type of AMENDED SHEET - 3a - channel coding used at the transmission end, to be known at the receiving end. For this purpose, it is possible for certain bits, so-called mode bits, AMENDED SHEET - 4 - to be generated at the transmission end, which, for example, indicate the rate used for source or channel coding. It is known for the mode bits to be protected and to be transmitted using a block code, independently of the source bits (data bits). In consequence, these so-called mode bits can be decoded first of all, with the source bits being determined subsequently, depending or. this first decoding result. A disadvantage cf this method is that the error frequency is relatively high in the mode bits since, particularly in mobile radio channels which are subject to fading, the correction capability in the decoder is low, owing to the short block length. Alternatively, it is possible to carry out the decoding in a number of steps. To do this, decoding is initially carried out using a first mode, and a CRC (Cyclic Redundancy Check) is used to determine whether this mode was worthwhile. If this is not the case, decoding is carried out using a further mode, and the result is checked once again. This method is repeated with all the modes until a sensible result is obtained. The disadvantage of this method is the high computation complexity, which leads to an increased power consumption, and to a decoding delay. It is known from US 5537410 for information to be transmitted in a frame in order to indicate a rate. The invention is thus based on the problem of specifying a method and an arrangement for channel coding and for decoding which allows information about AMENDED SHEET - 4a - the type of coding to be transmitted in a simple manner and reliably. This problem is solved by the features of the independent patent claims. Developments of the invention can be found in the dependent claims. AMENDED SHEET - 5 - In order to achieve the object, a method is specified in which a first portion of first information items, for example user information items, is channel coded, in a standard manner independently of the nature of the coding, for different types of coding. This ensures that a first portion of the first information items can also be used for decoding second information items for describing the coding of first information items, and better quality error correction for the second information items can thus be carried out. by the increase in the block length, associated with this, of the convolution codes used for coding the second information items. This allows multiple decoding, using the try and error principle described above, to be avoided. Information for describing the coding of first information items may in this case contain information for describing the source coding and/or the channel coding and/or other first information items for decoding, such as the type of coding (source and/or channel coding of the first information items) or the coding rate {source and/or channel coding of the first information items} . The invention can be used advantageously, particularly if the coding of the information can be carried out such that it is adapted to different types. In one refinement of the invention, the rate of channel coding of at least a second portion of the first information items for example the user information, is - 5a - matched to the quality of the transmission channel and/or to the network load. It is thus possible to match the channel coding to changing frame conditions in a communications system, - 6 - and to transmit this adaptation at the transmission end to a receiver in a simple and reliable manner. In further refinements, the second information items contain signaling information and/or information for describing the reception quality, in order to influence a transmitter as a function of the reception result at that time. It is thus possible to control the transmission of information based on the principle of a control loop. For channel coding, it is advantageous to use convolution codes, and to match the length of the first portion of the first information items, which is channel-coded in a standard manner, at least approximately to the length of influence of the convolution code being used. Furthermore, the problem is solved by a method for decoding of information structured in frames, in which a first portion of the first information items is also used for decoding second information items. This allows the coded transmission of second information ' items with a sufficiently long block length, and avoids complex multiple decoding based on the try and error principle described above. Decoding carried out in a manner such as this is advantageous in particular if one of the methods described above has been used to code the information at the transmission end as part of the information transmission process. The problem is also solved by arrangements for channel AMENDED SHEET coding and decoding of information structured in frames, in which a digital signal processor is in each case set up in such a manner that a first portion of the first information items can be channel-coded in a standard manner, independently of the nature of the coding, for different types of coding, AMENDED SHEET - 7 - and/or in such a manner that a first portion of the first information items can also be used for decoding second information items. These arrangements are particularly suitable for carrying out the methods according to the invention, or one of their developments explained above. Exemplary embodiments will be described and explained in the following text with reference to the following drawings. In this case, digital transmission of information is described, in particular. Nevertheless, the invention can also be applied to the storage of information, since the writing of information to a storage medium and the reading of information from a storage medium corresponds, in terms of the present invention to the transmission of information and the reception of information. Figure 1 shows an outline circuit diagram of a mobile radio system; Figure 2 shows a schematic illustration of major elements of a telecommunications transmission chain; Figure 3 shows a schematic illustration of an adaptive coding scheme; Figure A ' shows a schematic illustration of an adaptive coding scheme in the full-rate channel; Figure 5 shows a schematic illustration of an adaptive coding scheme in the half-rate channel; Figure 6 shows an outline circuit diagram of a processor unit. The structure of the mobile radio system shown in Figure 1 corresponds to that of a known GSM mobile AMENDED SHEET - 7a - radio system which comprises a large number of mobile switching centers MSC which are networked with one another and allow access to a landline network PSTN. Furthermore, these mobile switching centers MSC are each connected to at least one base station controller BSC, which may also be formed by a data processing system. Each base station controller BSC is in turn AMENDED SHEET - 8 - connected to at least one base station BS. A base station BS such as this is a radio appliance which can set up a radio link via a radio interface to radio appliances, which are referred to as mobile stations MS. The range of the signals from a base station, essentially defines a radio cell FZ. The allocation cf resources such as frequency bands to radio cells, and thus to the data packets to be transmitted, can be controlled by control devices such as the base station controllers BSC. Base stations 35 and a base station controller BSC car be combined to form a base station system BSS. The base station system BSS is in this case also responsible for the radio channel administration, the data rate matching, the monitoring of the radio transmission path, handover procedures, connection control and, possibly, for the allocation and signaling of the voice codecs to be used and, if required, transmits appropriate signaling information to the mobile stations MS. Such signaling information can also be transmitted via signaling channels. On the basis of the present description, the invention can also be used for signaling other information items, such as the type of information (data, speech, pictures, etc.) and/ox its coding, switching /information, using any desired transmission method, such as DECT, WB-CDMA or multimode transmission methods (GSM/WB-CDMA/TD-CDMA) within a UMTS (Universal Mobile Telephony System). Figure 2 shows a source Q which produces source signals qs which are compressed by a source coder QE, such as the GSM full-rate voice coding, to form symbol AMENDED SHEET - 9 - sequences composed of symbols. For parametric source coding methods, the source signals qs (for example, voice) produced by the source Q are subdivided into blocks (for example time frames), and these are ¦processed separately. The source coder QE produces quantized parameters (for example voice coefficients), which are also referred to in the following text as symbols in a symbol sequence, and which reflect the characteristics of the source in that particular block in a specific manner (for example the voice spectrum, filter parameters). After quantization, these symbols have a specific symbol value. The symbols in the symbol sequence and the corresponding symbol values are mapped onto a sequence of binary code words, each of which have a number of bit positions, by means of a binary mapping process (allocation rule), which is frequently written as part of the source coding QE. If, for example, these binary code words are processed further successively as a sequence of binary code words, then this results in a sequence of source-coded bit positions, which can be embedded in a frame structure. Methods which will not be explained here are used to considerably reduce, for example, the original rate of a telephone voice signal (65 kbps law, 104 kbps linear PCM) (approximately 5 kbps - 13 kbps, depending on the coding method) . Errors in this bit stream have different effects on the voice quality' after decoding. Errors in some bits lead to incomprehension or loud noises, while errors in other bits are scarcely perceptible. This leads to the bits being subdivided after the source coder QE into classes, which generally AMENDED SHEET - 9a - also have different protection against errors (for example: GSM full-rate codec: Class la, 1b and 2) After source coding has been carried out in such a manner, source bits or data bits db are produced/ - 10 - structured in frames, at a source bit rate which depends on the type of source coding. In mobile radio systems, convolution codes have been found to be efficient codes for subsequent channel coding. If the block length is long, these convolution codes have a high error correction capability and can be decoded with reasonable complexity. For the purposes of example, only rate 1/n convolution codes are dealt with in the following text. A convolution coder with a memory m produces n code bits via a register from the last m+1 data bits. As already explained above, bits are subdivided into classes during voice coding, and these classes are protected against errors in different ways. This is done by means of different rates during convolution coding. Rates greater than ½ are achieved by puncturing. A new voice and channel coding standard for existing GSM is currently being produced by the Standardization Group for. Mobile Radio Systems in Europe (ETSI). In this standard, the voice is intended to be source-coded and different rates using different coding modes, and the channel coding will be adapted appropriately so that the source-coded bit sequences will be coded with respect to channel interference in a channel coder CE, such as a convolution coder, in such a manner that the gross bit rate will still be 22.8 kbps (full-rate mode} or 11.4 kbps {half-rate mode). The particular source bit rate in this case varies depending on the voice (pause, hiss volume, the voice volume itself, strong or weak voice, etc.}, as a function of the channel conditions (good, poor channel} and as a function of AMENDED SHEET - 10a - network conditions (overloading, compatibility, etc.). The channel coding is adapted in a corresponding manner. The particular rate (for example by virtue of the particular code mode) being used and/or further information are/is transmitted as mode bits mb within the same frame. AMENDED SHEET - 11 - As is illustrated in Figure 3, for hierarchical coding purposes, the first portion of the data bits dbl is coded in the same way for all the source bit rates, voice coding rates and code modes being used. This first portion dbl may be the approximately 5-m first source bits. The trellis for this first portion is then set up in the channel decoder QD, and the mode bits are decided on first of all. The particular voice rate or the particular code mode is determined from these mode bits mb, and the second portion of the data bits db2 is also decoded in accordance with the decoding method used for this rate or this code mode. The first portion or another portion of the data bits dbl can also be channel-coded in a standard manner together with the mode bits mb, independently of the nature of the source coding, that is to say a first portion of the data bits dbl is channel-coded together with the mode bits mb independently of the particular code mode, and to be precise likewise using a standard method for all the code modes. This will be explained using a simple example, with reference to Figure 3: A source coding method produces frames or blocks using two different code modes with a length of 140 data bits db (case 1) or 100 data bits db (case 2), respectively. A mode bit mb which can also be transmitted in the same frame is intended to indicate which of the two block lengths has just been generated by the source coder QD. After channel coding, a frame with the length of 303 bits is intended to be produced in both cases-, which necessarily leads to different channel coding methods, at least with regard to the rate, for the two cases. It is now proposed, in both cases, for a first portion AMENDED SHEET - lla - of the data bits dbl, for example the. first 20 bits, to be channel-coded in a standard manner, for example with regard to the rate (rate 1/3}, AMENDED SHEET - 12 - the convolution codes used, the generator polynomials used or the memory used, and for the matching to the standard frame length of 303 bits to be carried out by using different rates (½ rate for case 1; 1/3 rate for case 2) in the channel coding for the second portion of the 120 (case 1) or 80 (case 2) data bits db2. In one embodiment variant of the invention, the mode bit or bits mb is or are to be channel-coded, in particular convolution-coded, in a standard manner together with the first portion of the data bits dbl ir. both cases, for example with regard to the rate (rate 1/3) of the convolution codes used, of the generator polynomials used, or of the memory used. During decoding, the trellis for a convolution decoder can be set up for the first 21 bits (one mode bit mb + 20 first data bits dbl) of a convolution decoder, without knowing what data block length has been used for the coding. If the trellis is set up over this length, the first bit (the mode bit mb) can be determined. The length of influence of the code is taken into account in the process, and the error rate is thus considerably lower than if the trellis were to be set up only for this first mode bit. Once this mode bit has been determined, the block length being used or the code mode used is also known, and the second portion of the data bits db2 is decoded at the 1/2 rate or the 1/3 rate as a function of this. The decoding complexity is thus only insignificantly greater than when decoding only one mode. A systematic convolution code can be used for poor channels, in order to keep the error rate below the channel error rate. A recursive code can be used in order to achieve - 13 - very good correction characteristics in good channels, despite this. For good channels, the error rate is higher than with a non-systematic non-recursive convolution code (previous GSM). However, this is evident only for an error rate of 10 or less. In this region, any errors which occur can be identified and disguised; the voice quality is not adversely affected. A scheme for both the half-rate channel and the full-rate channel will be described in the following text. Figure 4 shows the scheme for the full-rate channel (FR) : the voice coding generates four different rates at 13.3 kbps (code mode 1), 9.5 kbp/s (code mode 2), 3.1 kbp/s (code mode 3) and 6.3 kbp/s (code mode 4). The coding is carried out in frames or blocks with a duration of 20 ms. In addition, a CRC with 4 bits is added before the convolution coding in code mode 2, and 2 CRCs each of 3 bits, are added in both code modes 3 and 4. This leads to block lengths of 266 bits db (code mode 1), 199 bits db (code mode 2), 168 bits db (code mode 3) and 132 bits db (code mode 4) . Three code mode bits mb are provided at the start of each block or frame in order to signal the particular code mode and in order to transmit further signaling information. A recursively systematic convolution code at rates of 1/2 and 1/3 is used for the coding. 1/4 and 1/5 rates are produced by bit repetition, and higher rates by puncturing. Once again, the mode bits mb and a first portion of the data bits dbl are channel-coded in the same way for all four code modes, with the mode bits mb always being channel-coded at a rate 1/5, and the first portion of the data bits dbl always being channel-coded at a rate 1/3 or 1/4. Figure 5 shows the scheme for the half-rate channel (HR): the already explained principle of the same decoding of the mode bits - 14 - and of the first data bits is also implemented for the half-rate channel. Only the code modes 3 (8.1 kbps) and 4 (6.3 kbps) are used there, with the rate being increased to 11.4 kbps by channel coding. Since fewer code modes are used, 2 mode bits are sufficient in the half-rate channel. The convolution coder used is the same as that in the full-rate codec, but it is not terminated. As illustrated in Figure 2, bit sequences x or code bits which are channel-coded such as they are matched to the source coding are processed further in a modulator, which is not illustrated, and are then transmitted via a transmission path CH. Interference, such as fading or noise, occurs during transmission. The transmission path CH is located between a transmitter and a receiver. The receiver may contain an antenna, which is not illustrated, for receiving the signals transmitted via the transmission path CH, a sampling device, a demodulator for demodulating the signals, and an equalizer for elimination of intersymbol interference. Once again, for simplicity reasons, these devices have not been shown in Figure 1. The figure also does not show any possible interleaving and deinterleaving. The equalizer emits received values from a received' sequence y. Owing to interference during transmission on the transmission path CH, the received values 'have values other than "+1" and "-1". The channel coding is reversed in a channel decoder CD. The Viterbi ' algorithm is advantageously used for decoding convolution codes. - 15 - Depending on the convolution code memory to the length of influence during decoding is approximately 5m. The aim of this is to express the fact that, in general, up to this length of influence, errors in the code can still be corrected. The particular information bit is not corrected by any further removed code bits in the block to be decoded. The trellis of the decoder is set up approximetely as far as the 5.m removed data bit in order to achieve as low an error rate as possible for the first bit in a decoder. A decision on the first bit is then made. In a system with different source coder rates, the source coding of the first 5.m bits is also generally different. This means that the source decoding for these bits is also different, ana the decoding process must therefore be carried out differently depending on the source coding rate being used. Once channel decoding CD has been carried out, this results in the received mode bits mb and data bits db and source decoding QD is carried out to produce received source signals qs, which are emitted at the information sink S. In embodiment variants of the invention, other information, in particular control or signaling information, can also be transmitted by means of the mode bits, such as channel status information or responses to the signaling information (back-channel), information relating to the description of the coding used or the decoding to be used or other information which can be used for decoding the first information items. - 15a - Figure 6 shows a processor unit PE which may, in particular, be included in a communications device such as a base station BS or mobile station MS. It includes a AMENDED SHEET - 16 - control device STE, which essentially comprises a programmable microcontroller, and a processing device VE, which comprises a processor, in particular a digital signal processor, both of which may have write and read access to memory modules SPE. The microcontroller controls and monitors all the major elements and functions in a functional unit which includes the processor unit PE. The digital signal processor, a part of the digital signal processor, or a specific processor is responsible for carrying out the voice coding and voice decoding. The choice of voice codec can also be made by the microcontroller, or by the digital signal processor itself. An input/output interface I/O is used for inputting/outputting user data or control data, for example, to a control unit MMI, which may include a keyboard and/or a display. AMENDED SHEET 17 We Claim. 1. A method for channel coding of information structured in frames in a mobile radio system, in which first information items (db) and second information items (mb) for describing the coding of first information items are contained within a frame, characterized in that the second information items (mb) and a first portion (db1) of the first information items (db) are uniformly channel-coded irrespective of the nature of the coding for different types of coding. 2. The method as claimed in claim 1, wherein the coding of the information is carried out such that it is adapted to different types of codes. 3. The method as claimed in one of the preceding claims, wherein the rate of channel coding of at least a second portion of the first information items (db) is matched to the quality of the transmission channel and/or to the network load. 4. The method as claimed in one of the preceding claims, wherein second information items (mb) contain signalling information and /or information for describing the reception quality. 18 5. The method as claimed in one of the preceding claims, wherein convolution codes are used for channel coding, and wherein the first portion of the first information items, which is channel coded in a standard manner, corresponds at least to the length of influence of the convolution code. 6. A method for decoding of information structured in frames in a mobile radio system, in which first information items ( db) and second information items (mb) for describing the coding of first information items are contained within a frame, characterized in that a first portion of the channel-coded first information items (db) is used for channel decoding of the second information items (mb). 7. The method as claimed in claim 6, wherein a first portion of the first information items (db) uniformly channel-coded for different types of coding is used. 8. The method as claimed in one of claims 6 to 7, wherein the second information items (mb) are channel-decoded only once. 9. The method as claimed in one of claims 6 to 8, wherein the information to be decoded is coded using a method as claimed in one of claims 1 to 5. 119 10. A device for channel coding of information structured in frames, in a mobile radio system, in which first information items (db) and second information items (mb) for describing the coding of first information items are contained within a frame, the device comprising: a processor unit which is set up in such a manner that a first portion of the first information items can be uniformly channel coded irrespective of the nature of coding for different types of coding. 11. A device for decoding of information structured in frames in a mobile radio system, in which first information items (db) and second information items (mb) for describing the coding of first information items are contained within a frame, the device comprising a processor unit which is set up in such a manner that a first portion of the first information items (db) can also be used for channel decoding of the second information items (mb). 12. The device as claimed in claim 11, wherein a first portion of the first information items (db) uniformly channel coded for different types of coding can be used. A method for channel coding of information structured in frames in a mobile radio system, in which first information items (db) and second information items (mb ) for describing the coding of first information items are contained within a frame, characterized in that the second information items (mb) and a first portion (dbl) of the first information items (db) are uniformly channel-coded irrespective of the nature of the cooling for different types of coding.

Full Text

Description
Method and arrangement for channel coding and decoding of information structured in frames
The invention relates to a method and an arrangement,
for channel coding and decoding of information
structured in frames, in particular for the purposes of
adaptive multirate coding.
Source signals and source information such as voice, audio, picture and video signals virtually always contain statistical redundancy, that is to say redundant information. This redundancy can be greatly reduced by source coding, thus allowing efficient, transmission and storace of the source signal. This reduction in redundancy gets rid of signal contents which were redundant before transmission and are based on prior knowledge of, for example, statistical parameters in the signal profile. The bit rate of the source-coded information (source bits or data bits) is also referred to as the source bit rate. During the source decoding after transmission, these components are added to the signal once again, so that virtually no loss of quality can be verified objectively.
On the other hand, it is normal for signal transmission to add redundancy deliberately by means of channel coding once again, in order largely to correct for the influence of channel interference on the transmission. Additional redundant bits thus make it possible for the receiver or decoder to identify errors, and possibly also to correct them. The bit rate of the channel-coded information is also referred to as the gross bit rate.
AMENDED SHEET

-lain order to allow information, in particular voice data, picture data or other user data, to be
AMENDED SHEET

- 2 -
transmitted, as efficiently as possible by means of the limited transmission capacities of a transmission medium, in particular a radio interface, this information to be transmitted is thus compressed by means of source coding before transmission, and is protected against channel errors by means of channel coding. Various methods are in each case known for doing this. For example, in the GSM {Global System for Mobile Communication) System, voice can be coded by means of a full rate voice codec a half rate voice codec, or an enhanced full rate voice codec.
For the purposes of this application, the terms voice codec or coding also refer to a method for encoding and/or for corresponding decoding, which method may also cover sources and/or channel coding, and can also be applied to data other than voice data.
In the course of the further development of the European mobile radio standard GSM, a new standard for coded voice transmission is being developed, which will allow the overall data rate and the splitting of the data rate between the source coding and channel coding to be set adaptively depending on the channel state and the network conditions (system load). Instead of the voice codecs described above, which have a fixed source bit rate, new voice codecs are intended to be used for this purpose, whose source bit rate is variable and is matched to changing frame conditions for information transmission. The main aims of such AMR (Adaptive Multirate) voice codecs are to achieve landline network quality for voice transmission in various channel conditions, and to ensure optimum distribution of the channel capacity, taking account of specific network

- 2a -
parameters. After carrying out a conventional source coding method, the compressed information is in structured form, in frames, in which case the source bit rate may differ from frame to frame depending on the code mode being used.
AMENDED SHEET

- 3 -
In order to achieve standard gross bit rates, the information contained within a frame is channel-coded in a different manner depending on the source bit rate or code mode, in particular at a different rate, in such a manner that the gross bit rate after channel coding corresponds to the selected channel mode (half rate or full rate). For example, such an AMR voice codec can operate using the half rate (HR) channel in good channel conditions and/or in highly loaded radio cells. When the channel conditions are poor, a dynamic change should be made to the full rate (FR) channel, and vice versa. Within such a channel mode (half rate or full rate), various code modes are available for different voice and channel coding rates, and these are likewise selected to match the channel quality (rate adaptation) . In the process, the gross bit rate after channel coding remains constant within one channel mode (22.8 kbps for the full rate channel FR and 11.4 kbps for the half-fate channel HR) . This is intended to result in the best voice quality, taking account of the changing channel conditions. Thus, with such adaptive coding, different rates are used for voice coding (variable source bit rate) depending on the channel conditions in a transmission path, on the requirements for specific network parameters, and depending oh the voice. Since the gross bit rate after channel coding is intended to remain constant, an appropriately adapted variable number of error protection bits are added during channel coding.
In order to decode such variably coded information after transmission, it is helpful for information about the coding method used at the transmission end, in particular the source bit rate and/or the type of
AMENDED SHEET

- 3a -
channel coding used at the transmission end, to be known at the receiving end. For this purpose, it is possible for certain bits, so-called mode bits,
AMENDED SHEET

- 4 -
to be generated at the transmission end, which, for example, indicate the rate used for source or channel coding.
It is known for the mode bits to be protected and to be transmitted using a block code, independently of the source bits (data bits). In consequence, these so-called mode bits can be decoded first of all, with the source bits being determined subsequently, depending or. this first decoding result. A disadvantage cf this method is that the error frequency is relatively high in the mode bits since, particularly in mobile radio channels which are subject to fading, the correction capability in the decoder is low, owing to the short block length.
Alternatively, it is possible to carry out the decoding in a number of steps. To do this, decoding is initially carried out using a first mode, and a CRC (Cyclic Redundancy Check) is used to determine whether this mode was worthwhile. If this is not the case, decoding is carried out using a further mode, and the result is checked once again. This method is repeated with all the modes until a sensible result is obtained. The disadvantage of this method is the high computation complexity, which leads to an increased power consumption, and to a decoding delay.
It is known from US 5537410 for information to be transmitted in a frame in order to indicate a rate.
The invention is thus based on the problem of specifying a method and an arrangement for channel coding and for decoding which allows information about
AMENDED SHEET

- 4a -
the type of coding to be transmitted in a simple manner and reliably.
This problem is solved by the features of the independent patent claims. Developments of the invention can be found in the dependent claims.
AMENDED SHEET

- 5 -
In order to achieve the object, a method is specified in which a first portion of first information items, for example user information items, is channel coded, in a standard manner independently of the nature of the coding, for different types of coding.
This ensures that a first portion of the first information items can also be used for decoding second information items for describing the coding of first information items, and better quality error correction for the second information items can thus be carried out. by the increase in the block length, associated with this, of the convolution codes used for coding the second information items. This allows multiple decoding, using the try and error principle described above, to be avoided.
Information for describing the coding of first
information items may in this case contain information
for describing the source coding and/or the channel
coding and/or other first information items for
decoding, such as the type of coding (source and/or
channel coding of the first information items) or the
coding rate {source and/or channel coding of the first
information items} .
The invention can be used advantageously, particularly if the coding of the information can be carried out such that it is adapted to different types.
In one refinement of the invention, the rate of channel coding of at least a second portion of the first information items for example the user information, is

- 5a -
matched to the quality of the transmission channel and/or to the network load. It is thus possible to match the channel coding to changing frame conditions in a communications system,

- 6 -
and to transmit this adaptation at the transmission end to a receiver in a simple and reliable manner.
In further refinements, the second information items contain signaling information and/or information for describing the reception quality, in order to influence a transmitter as a function of the reception result at that time. It is thus possible to control the transmission of information based on the principle of a control loop.
For channel coding, it is advantageous to use convolution codes, and to match the length of the first portion of the first information items, which is channel-coded in a standard manner, at least approximately to the length of influence of the convolution code being used.
Furthermore, the problem is solved by a method for decoding of information structured in frames, in which a first portion of the first information items is also used for decoding second information items. This allows the coded transmission of second information ' items with a sufficiently long block length, and avoids complex multiple decoding based on the try and error principle described above.
Decoding carried out in a manner such as this is advantageous in particular if one of the methods described above has been used to code the information at the transmission end as part of the information transmission process.
The problem is also solved by arrangements for channel
AMENDED SHEET

coding and decoding of information structured in frames, in which a digital signal processor is in each case set up in such a manner that a first portion of the first information items can be channel-coded in a standard manner, independently of the nature of the coding, for different types of coding,
AMENDED SHEET

- 7 -
and/or in such a manner that a first portion of the first information items can also be used for decoding second information items. These arrangements are particularly suitable for carrying out the methods according to the invention, or one of their developments explained above.
Exemplary embodiments will be described and explained in the following text with reference to the following drawings. In this case, digital transmission of information is described, in particular. Nevertheless, the invention can also be applied to the storage of information, since the writing of information to a storage medium and the reading of information from a storage medium corresponds, in terms of the present invention to the transmission of information and the reception of information.
Figure 1 shows an outline circuit diagram of a mobile
radio system; Figure 2 shows a schematic illustration of major
elements of a telecommunications transmission
chain; Figure 3 shows a schematic illustration of an adaptive
coding scheme; Figure A ' shows a schematic illustration of an adaptive
coding scheme in the full-rate channel; Figure 5 shows a schematic illustration of an adaptive
coding scheme in the half-rate channel; Figure 6 shows an outline circuit diagram of a
processor unit.
The structure of the mobile radio system shown in Figure 1 corresponds to that of a known GSM mobile
AMENDED SHEET

- 7a -
radio system which comprises a large number of mobile switching centers MSC which are networked with one another and allow access to a landline network PSTN. Furthermore, these mobile switching centers MSC are each connected to at least one base station controller BSC, which may also be formed by a data processing system. Each base station controller BSC is in turn
AMENDED SHEET

- 8 -
connected to at least one base station BS. A base station BS such as this is a radio appliance which can set up a radio link via a radio interface to radio appliances, which are referred to as mobile stations MS.
The range of the signals from a base station, essentially defines a radio cell FZ. The allocation cf resources such as frequency bands to radio cells, and thus to the data packets to be transmitted, can be controlled by control devices such as the base station controllers BSC. Base stations 35 and a base station controller BSC car be combined to form a base station system BSS.
The base station system BSS is in this case also responsible for the radio channel administration, the data rate matching, the monitoring of the radio transmission path, handover procedures, connection control and, possibly, for the allocation and signaling of the voice codecs to be used and, if required, transmits appropriate signaling information to the mobile stations MS. Such signaling information can also be transmitted via signaling channels.
On the basis of the present description, the invention can also be used for signaling other information items, such as the type of information (data, speech, pictures, etc.) and/ox its coding, switching /information, using any desired transmission method, such as DECT, WB-CDMA or multimode transmission methods (GSM/WB-CDMA/TD-CDMA) within a UMTS (Universal Mobile Telephony System).

Figure 2 shows a source Q which produces source signals qs which are compressed by a source coder QE, such as the GSM full-rate voice coding, to form symbol
AMENDED SHEET

- 9 -
sequences composed of symbols. For parametric source coding methods, the source signals qs (for example, voice) produced by the source Q are subdivided into blocks (for example time frames), and these are ¦processed separately. The source coder QE produces quantized parameters (for example voice coefficients), which are also referred to in the following text as symbols in a symbol sequence, and which reflect the characteristics of the source in that particular block in a specific manner (for example the voice spectrum, filter parameters). After quantization, these symbols have a specific symbol value.
The symbols in the symbol sequence and the corresponding symbol values are mapped onto a sequence of binary code words, each of which have a number of bit positions, by means of a binary mapping process (allocation rule), which is frequently written as part of the source coding QE. If, for example, these binary code words are processed further successively as a sequence of binary code words, then this results in a sequence of source-coded bit positions, which can be embedded in a frame structure.
Methods which will not be explained here are used to considerably reduce, for example, the original rate of a telephone voice signal (65 kbps law, 104 kbps linear PCM) (approximately 5 kbps - 13 kbps, depending on the coding method) . Errors in this bit stream have different effects on the voice quality' after decoding. Errors in some bits lead to incomprehension or loud noises, while errors in other bits are scarcely perceptible. This leads to the bits being subdivided after the source coder QE into classes, which generally
AMENDED SHEET

- 9a -
also have different protection against errors (for example: GSM full-rate codec: Class la, 1b and
2) After source coding has been carried out in such a manner, source bits or data bits db are produced/

- 10 -
structured in frames, at a source bit rate which depends on the type of source coding.
In mobile radio systems, convolution codes have been found to be efficient codes for subsequent channel coding. If the block length is long, these convolution codes have a high error correction capability and can be decoded with reasonable complexity. For the purposes of example, only rate 1/n convolution codes are dealt with in the following text. A convolution coder with a memory m produces n code bits via a register from the last m+1 data bits.
As already explained above, bits are subdivided into classes during voice coding, and these classes are protected against errors in different ways. This is done by means of different rates during convolution coding. Rates greater than ½ are achieved by puncturing.
A new voice and channel coding standard for existing GSM is currently being produced by the Standardization Group for. Mobile Radio Systems in Europe (ETSI). In this standard, the voice is intended to be source-coded and different rates using different coding modes, and the channel coding will be adapted appropriately so that the source-coded bit sequences will be coded with respect to channel interference in a channel coder CE, such as a convolution coder, in such a manner that the gross bit rate will still be 22.8 kbps (full-rate mode} or 11.4 kbps {half-rate mode). The particular source bit rate in this case varies depending on the voice (pause, hiss volume, the voice volume itself, strong or weak voice, etc.}, as a function of the channel conditions (good, poor channel} and as a function of
AMENDED SHEET

- 10a -
network conditions (overloading, compatibility, etc.). The channel coding is adapted in a corresponding manner. The particular rate (for example by virtue of the particular code mode) being used and/or further information are/is transmitted as mode bits mb within the same frame.
AMENDED SHEET

- 11 -
As is illustrated in Figure 3, for hierarchical coding purposes, the first portion of the data bits dbl is coded in the same way for all the source bit rates, voice coding rates and code modes being used. This first portion dbl may be the approximately 5-m first source bits. The trellis for this first portion is then set up in the channel decoder QD, and the mode bits are decided on first of all. The particular voice rate or the particular code mode is determined from these mode bits mb, and the second portion of the data bits db2 is also decoded in accordance with the decoding method used for this rate or this code mode.
The first portion or another portion of the data bits dbl can also be channel-coded in a standard manner together with the mode bits mb, independently of the nature of the source coding, that is to say a first portion of the data bits dbl is channel-coded together with the mode bits mb independently of the particular code mode, and to be precise likewise using a standard method for all the code modes.
This will be explained using a simple example, with reference to Figure 3:
A source coding method produces frames or blocks using two different code modes with a length of 140 data bits db (case 1) or 100 data bits db (case 2), respectively. A mode bit mb which can also be transmitted in the same frame is intended to indicate which of the two block lengths has just been generated by the source coder QD. After channel coding, a frame with the length of 303 bits is intended to be produced in both cases-, which necessarily leads to different channel coding methods, at least with regard to the rate, for the two cases. It is now proposed, in both cases, for a first portion
AMENDED SHEET

- lla -
of the data bits dbl, for example the. first 20 bits, to be channel-coded in a standard manner, for example with regard to the rate (rate 1/3},
AMENDED SHEET

- 12 -
the convolution codes used, the generator polynomials used or the memory used, and for the matching to the standard frame length of 303 bits to be carried out by using different rates (½ rate for case 1; 1/3 rate for case 2) in the channel coding for the second portion of the 120 (case 1) or 80 (case 2) data bits db2.
In one embodiment variant of the invention, the mode bit or bits mb is or are to be channel-coded, in particular convolution-coded, in a standard manner together with the first portion of the data bits dbl ir. both cases, for example with regard to the rate (rate 1/3) of the convolution codes used, of the generator polynomials used, or of the memory used.
During decoding, the trellis for a convolution decoder can be set up for the first 21 bits (one mode bit mb + 20 first data bits dbl) of a convolution decoder, without knowing what data block length has been used for the coding. If the trellis is set up over this length, the first bit (the mode bit mb) can be determined. The length of influence of the code is taken into account in the process, and the error rate is thus considerably lower than if the trellis were to be set up only for this first mode bit. Once this mode bit has been determined, the block length being used or the code mode used is also known, and the second portion of the data bits db2 is decoded at the 1/2 rate or the 1/3 rate as a function of this.
The decoding complexity is thus only insignificantly greater than when decoding only one mode. A systematic convolution code can be used for poor channels, in order to keep the error rate below the channel error rate. A recursive code can be used in order to achieve

- 13 -
very good correction characteristics in good channels, despite this. For good channels, the error rate is higher than with a non-systematic non-recursive convolution code (previous GSM). However, this is evident only for an error rate of 10 or less. In this region, any errors which occur can be identified and disguised; the voice quality is not adversely affected.
A scheme for both the half-rate channel and the full-rate channel will be described in the following text.
Figure 4 shows the scheme for the full-rate channel (FR) : the voice coding generates four different rates at 13.3 kbps (code mode 1), 9.5 kbp/s (code mode 2), 3.1 kbp/s (code mode 3) and 6.3 kbp/s (code mode 4). The coding is carried out in frames or blocks with a duration of 20 ms. In addition, a CRC with 4 bits is added before the convolution coding in code mode 2, and 2 CRCs each of 3 bits, are added in both code modes 3 and 4. This leads to block lengths of 266 bits db (code mode 1), 199 bits db (code mode 2), 168 bits db (code mode 3) and 132 bits db (code mode 4) . Three code mode bits mb are provided at the start of each block or frame in order to signal the particular code mode and in order to transmit further signaling information. A recursively systematic convolution code at rates of 1/2 and 1/3 is used for the coding. 1/4 and 1/5 rates are produced by bit repetition, and higher rates by puncturing. Once again, the mode bits mb and a first portion of the data bits dbl are channel-coded in the same way for all four code modes, with the mode bits mb always being channel-coded at a rate 1/5, and the first portion of the data bits dbl always being channel-coded at a rate 1/3 or 1/4.

Figure 5 shows the scheme for the half-rate channel (HR): the already explained principle of the same decoding of the mode bits

- 14 -
and of the first data bits is also implemented for the half-rate channel. Only the code modes 3 (8.1 kbps) and 4 (6.3 kbps) are used there, with the rate being increased to 11.4 kbps by channel coding. Since fewer code modes are used, 2 mode bits are sufficient in the half-rate channel. The convolution coder used is the same as that in the full-rate codec, but it is not terminated.
As illustrated in Figure 2, bit sequences x or code bits which are channel-coded such as they are matched to the source coding are processed further in a modulator, which is not illustrated, and are then transmitted via a transmission path CH. Interference, such as fading or noise, occurs during transmission.
The transmission path CH is located between a transmitter and a receiver. The receiver may contain an antenna, which is not illustrated, for receiving the signals transmitted via the transmission path CH, a sampling device, a demodulator for demodulating the signals, and an equalizer for elimination of intersymbol interference. Once again, for simplicity reasons, these devices have not been shown in Figure 1. The figure also does not show any possible interleaving and deinterleaving.
The equalizer emits received values from a received' sequence y. Owing to interference during transmission on the transmission path CH, the received values 'have values other than "+1" and "-1".
The channel coding is reversed in a channel decoder CD. The Viterbi ' algorithm is advantageously used for decoding convolution codes.

- 15 -
Depending on the convolution code memory to the length of influence during decoding is approximately 5m. The aim of this is to express the fact that, in general, up to this length of influence, errors in the code can still be corrected. The particular information bit is not corrected by any further removed code bits in the block to be decoded.
The trellis of the decoder is set up approximetely as far as the 5.m removed data bit in order to achieve as low an error rate as possible for the first bit in a decoder. A decision on the first bit is then made. In a system with different source coder rates, the source coding of the first 5.m bits is also generally different. This means that the source decoding for these bits is also different, ana the decoding process must therefore be carried out differently depending on the source coding rate being used.
Once channel decoding CD has been carried out, this results in the received mode bits mb and data bits db and source decoding QD is carried out to produce received source signals qs, which are emitted at the information sink S.
In embodiment variants of the invention, other information, in particular control or signaling information, can also be transmitted by means of the mode bits, such as channel status information or responses to the signaling information (back-channel), information relating to the description of the coding used or the decoding to be used or other information which can be used for decoding the first information items.

- 15a -
Figure 6 shows a processor unit PE which may, in particular, be included in a communications device such as a base station BS or mobile station MS. It includes a
AMENDED SHEET

- 16 -
control device STE, which essentially comprises a programmable microcontroller, and a processing device VE, which comprises a processor, in particular a digital signal processor, both of which may have write and read access to memory modules SPE.
The microcontroller controls and monitors all the major elements and functions in a functional unit which includes the processor unit PE. The digital signal processor, a part of the digital signal processor, or a specific processor is responsible for carrying out the voice coding and voice decoding. The choice of voice codec can also be made by the microcontroller, or by the digital signal processor itself.
An input/output interface I/O is used for inputting/outputting user data or control data, for example, to a control unit MMI, which may include a keyboard and/or a display.
AMENDED SHEET

17 We Claim.
1. A method for channel coding of information structured in frames in a mobile radio system, in which first information items (db) and second information items (mb) for describing the coding of first information items are contained within a frame, characterized in that the second information items (mb) and a first portion (db1) of the first information items (db) are uniformly channel-coded irrespective of the nature of the coding for different types of coding.
2. The method as claimed in claim 1, wherein the coding of the information is
carried out such that it is adapted to different types of codes.
3. The method as claimed in one of the preceding claims, wherein the rate of
channel coding of at least a second portion of the first information items
(db) is matched to the quality of the transmission channel and/or to the
network load.
4. The method as claimed in one of the preceding claims, wherein second information items (mb) contain signalling information and /or information for describing the reception quality.

18
5. The method as claimed in one of the preceding claims, wherein
convolution codes are used for channel coding, and wherein the first
portion of the first information items, which is channel coded in a standard
manner, corresponds at least to the length of influence of the convolution
code.
6. A method for decoding of information structured in frames in a mobile
radio system, in which first information items ( db) and second information
items (mb) for describing the coding of first information items are
contained within a frame, characterized in that a first portion of the
channel-coded first information items (db) is used for channel decoding of
the second information items (mb).
7. The method as claimed in claim 6, wherein a first portion of the first
information items (db) uniformly channel-coded for different types of
coding is used.
8. The method as claimed in one of claims 6 to 7, wherein the second
information items (mb) are channel-decoded only once.
9. The method as claimed in one of claims 6 to 8, wherein the information to
be decoded is coded using a method as claimed in one of claims 1 to 5.

119
10. A device for channel coding of information structured in frames, in a
mobile radio system, in which first information items (db) and second
information items (mb) for describing the coding of first information items
are contained within a frame, the device comprising:
a processor unit which is set up in such a manner that a first portion of the first information items can be uniformly channel coded irrespective of the nature of coding for different types of coding.
11. A device for decoding of information structured in frames in a mobile radio
system, in which first information items (db) and second information items
(mb) for describing the coding of first information items are contained
within a frame, the device comprising a processor unit which is set up in
such a manner that a first portion of the first information items (db) can
also be used for channel decoding of the second information items (mb).
12. The device as claimed in claim 11, wherein a first portion of the first
information items (db) uniformly channel coded for different types of
coding can be used.
A method for channel coding of information structured in frames in a mobile radio system, in which first information items (db) and second information items (mb ) for describing the coding of first information items are contained within a frame, characterized in that the second information items (mb) and a first portion (dbl) of the first information items (db) are uniformly channel-coded irrespective of the nature of the cooling for different types of coding.

Documents:

« Previous Patent

Next Patent »

Patent Number

205850

Indian Patent Application Number

IN/PCT/2001/00624/KOL

PG Journal Number

15/2007

Publication Date

13-Apr-2007

Grant Date

13-Apr-2007

Date of Filing

12-Jun-2001

Name of Patentee

SIEMENS AKTIENGESELLSCHAFT

Applicant Address

WITTELSBACHERPLATZ 2, D-80333 MUNCHEN GERMANY, A GERMAN COMPANY.

Inventors:

#	Inventor's Name	Inventor's Address
1	HAGENAUER, JOACHIM	PETER-ROSEGGER-STRASSE 41 D-82229 SEEFELD, GERMANY
2	XU, WEN	BISCHOFSHOFENER STRASSE 11, D-82008 UNTERHACHING, GERMANY
3	HINDELANG, THOMAS	KRUNER STRASSE 17, D-81373 MUNCHEN, GERMANY

PCT International Classification Number

H04L 1/00

PCT International Application Number

PCT/DE99/ 03838

PCT International Filing date

1999-12-01

PCT Conventions:

#	PCT Application Number	Date of Convention	Priority Country
1	198 58 393.1	1998-12-17	Germany