MXPA01005147A - Efficient in-band signaling for discontinuous transmission and configuration changes in adaptive multi-rate communications systems - Google Patents

Abstract

Techniques for discontinuous transmission (DTX) and fast in-band signaling of configuration changes and protocol messages in speech communications systems provide cost efficiency in terms of radio transmission capacity, in terms of fixed line transmission, and in terms of implementation effort. An exemplary method for performing discontinuous transmission (DTX) in a communications system in which source data is interleaved for transmission from a first component in the system to a second component in the system includes the steps of detecting periods of source data inactivity, and transmitting silence descriptor (SID) frames from the first to the second component during the periods of source data inactivity, certain of the transmitted SID frames being interleaved using a different interleaving algorithm as compared to that used for source data. For example, the source data can be block diagonally interleaved, and certain of the SID frames can be block interleaved. An exemplary method for effecting configuration changes in a communications system includes the step of transmitting an escape frame in place of a speech data frame, the escape frame including a gross bit pattern to distinguish the escape frame from speech data frames and conveying a configuration change indication. The escape frame can further include a data field to indicate a particular configuration change to be made. For example, where the communications system is an AMR system, an escape frame can be used to change an active codec mode set. Alternatively, an escape frame can be used to change a phase of codec information.

Description

SIGNAL 1 IN EFFICIENT BAND FOR TRANS MISSION DISCONTINUATION AND CHANGES OF FIGURE IN SYSTEMS OF COMMUNICATIONS ADJUSTABLE MULTIPLE SPEED CROSS REFERENCE TO RELATED REQUESTS The present application claims the benefit of the provisional application of the United States of America No. 60 / 109,694 filed on November 24, 1998, which is hereby incorporated by reference in its entirety.

FIELD OF THE I NVENTION The present invention relates to communication systems and more particularly to discontinuous transmission (DTX) and configuration changes in adaptive multiple speed communication systems.

BACKGROUND OF THE I NVENTION Currently, multiple mode coding systems employ at least different source and channel coding modes that can be used to maintain near-optimal communication quality under varying transmission channel conditions. A mode with a low bit rate source coding and a high degree of channel error protection can be selected for the erroneous channels. On the other hand suitable channels allow the selection of a coding mode with high bit rate source coding and a relatively low degree of error protection. As is well known in the art, multi-mode coding systems must convey (either explicitly or implicitly) the actually selected coding mode towards a receiver decoder to enable proper decoding of the received data. The two-way communication systems with the coding mode adaptation have additionally to transmit the similar information on the return link. That is, either the quantized link measurement data describing the previous channel status to the current one, or a corresponding request / command coding mode that takes into account the state of the channel. Such link adaptation data is known in the art as the coding mode information, which consists of coding mode indications (the coding mode actually selected) and encoding mode requests / commands (the coding mode which it will be used on the transmitter side). The evolving global system for the adaptive multiple velocity standard of mobile communication (GSM) employs the adaptation of coding mode described above. In such systems, AMR, in-band signaling is used to reassign parts of the speech transmission resource to transmit the control information. It is applied when other suitable control channels are not available. The voice coding standard, GSM ARM is an example that makes use of in-band signaling. It uses parts of the GSM voice traffic channel for the transmission of the AMR link adaptation data. More specifically, the GSM ARM standard provides a band channel for the transmission of the encoding mode information. The encoding mode information consists of the encoding mode requests / commands and the encoding mode indications, which are transmitted in every second of structure (every 40 ms), in alternating order. The encoding mode information identifies a coding mode in joint use of up to four encoding modes out of 8 (for Adaptive Total Rate Voice or AFS) or 6 (for Adaptive Average Velocity Voice AHS) available modes. These subsets of coding modes are referred to as active coding sets. In any coding system that includes the GSM system AM R described above, transmission capacity is a limited and expensive resource. For this reason, in order to save transmission capacity, discontinuous transmission (DTX) is widely applied when transmitting voice. Sometimes DTX is referred to as a voice operated transmission (VOX). The basic principle of DTX is to deactivate the transmission during voice inactivity. Instead, tolerable noise parameters are transmitted which allow the decoder to produce the inactivity signal, which is usually of some kind of background noise. CN parameters require much less transmission resource than voice. DT is also an important feature for mobile telephones since it allows to deactivate energy-consuming devices (such as radio transmitters) during inactivity. Doing this helps save battery power and increase phone conversation time. In two-way communication systems that employs DTX, there will typically be an active link as long as the other link is inactive (as one speaker is talking to the other listening). The active link has, with some speed of reduced structure transmission, that transfer the silence descriptor structures (SID) (also referred to as background information, or tolerable noise, descriptor structures) to the receiver. The SI D structures contain parameters CN and allow the receiver to generate the tolerable noise silence signal, for example to reassure that a user who listens to that connection is still active. In the current GSM coding standards FR, H R and EFR DTX is carried out in a similar way. By way of example, the state of the art of the voice communication operated by DTX in the GSM system will be described with respect to the GS EFR coding. For further information, see for example the standards GSM 06.1 1, GSM 06.12, GMS 06.21, GSM 06.22, GMS 06.31, GMS 06.41, GSM 06.61, GSM 06.62 and GSM 06.81. and the related documents. The GSM EFR scheme is characterized as follows: The end of the speech activity is signaled by transmission of a first structure SI D, which is not aligned in phase with SACCH. Instead, it immediately follows the last active voice structure. After the first structure SI D, the SI D structures are updated, with a period of one time per 24 structures (= 480ms). The transmission of update structure SI D is aligned with the time alignment signal (TAF), which is generated in the radio subsystems and which is derived from the frame structure, SACCH. Apart from the SI D structures, other structures are not transmitted during inactivity. Simply by summarizing the transmission of active voice structures the inactive period is ended. RSS handles SI D structures as regular voice structures. This means in particular that the same channel coding and diagonal collation is used for voice structures. A number of forty-three effective net bits (43) is used for the tolerable noise parameters that describe the spectral shape and the gain of the inactivity signal. Ninety-five net bits (95) are used for a special SI D bit pattern to identify the structure as a SI D structure and distinguish it from the voice structures. The CN parameters are differentially coded with respect to the parameters, which are derived from the last transmitted speech structures. The structure transmission SI D described is illustrated in Figure 1 for TCH / FS (ie, total traffic / voice channel) and in Figure 2 for TCH / HS (ie, traffic channel / speed voice). half) . The top row symbolizes the voice structures, as seen in the encoder input. The middle row symbolizes the TDMA structures that transmit the respective voice or SI D bits by means of the radio interface. The lower row symbolizes the tolerable noise structure or voice after the speech decoder. Such voice structure is exactly 20 ms in length. The TDMA structures have an average of 5 ms. The TDMA structures for SACCH and I DLE are not shown. The implementation delays and collateral effects are not shown either. Apart from the regular transmission of SI D structures, synchrony and time alignment for a fixed time structure, ITU-U Recommendation G.729 / Annex B describes a reference DTX method that transmits the SI D structures always that an update of the CN parameters is required because it has changed significantly since the last SI structure transmission D. In the well-known peaceful digital cellular system (PDC) with VOX functionality, special post and preamble structures are used to signal the transitions from the voice to inactivity or, respectively, back from inactivity to the voice (see for example, RCR STD-27D). These structures contain single bit patterns over the entire bit level to identify them. The postamble structures consist of two channel structures of which the first carries information of the identification bit pattern and of which the second carries two tolerable noise parameters that describe the idle signal. During voice inactivity, the postamble structures are sent periodically to allow the receiving end to update the generation of the tolerable noise. Postamble and preamble structures are used interleaving the same for voice structures. The conventional DTX solutions, described above, as achieved in GSM FR, EFR and HR, are not well suited for use in multi-mode coding systems. This results from the fact that SID structure signaling is done on the net bit level. A special bit pattern that identifies the SID structure is part of the net bit stream. The SID structure detection unit in the receiver will be executed after deinterleaving and channel decoding. This approach is inappropriate for multi-mode coding systems with more than one source and channel mode since SID structure identification would depend on the correct selection of coding mode for channel decoding. It is correct coding mode in the receiver can, due to possible mode transmission errors, not always guaranteed. Furthermore, for analogous reasons, the variations of the intercalation scheme, either for the different coding modes as well as for the SI D structures, is also impractical for reasons of complexity. Such approaches require in the worst case to run once intercalation of structure SI D and more strictly, the channel decoding in addition to the de-interleaving of the voice structure and the channel decoding. Additionally, there are at least two main problems in the adoption of the PDC realization. First, since the post-mortem structures consist of two traffic structures, the first transmission mode of relatively inefficient inactivity in terms of transmission energy savings. Each tolerable noise parameter update requires the transmission of two structures. Secondly, since the transmissions from voice inactivity to the activity are signaled by preamble structures, any part of the beginning of the voice can be subjected or the transmission of the voice starts is delayed by the preamble structures. The first case directly affects the quality sides of the reconstructed voice, while the last one increases the voice transmission delay which can cause degradation of the conversation quality. Note also that the application of a common diagonal intercalation scheme over two structures for SI D and voice structures, as is currently done in GSM and PDC, causes additional problems. The application of the diagonal intercalation for the transmission of individual SID structures is inefficient in terms of radio resource use and power consumption since only half of each TDMA structure transmitted carries the SI D information while the other half remains unused and therefore wasted (said bursts of half wasted are marked in figures 1 and 2). This loss of efficiency in the current GSM and PDC systems is less than the transmission of the SI D structure, which is relatively rare. However, it is stricter for some multi-mode communication systems with coding mode adaptation. The high adaptation performance requires much more frequent transmission (adaptation data) on the inactive link compared to the transmission of the SI D structures in the current systems. further, there are certain upper limits of the radio channel activity during inactivity (for example, the AM R system requirement is: TCH / FS: 16 TDMA structures per 480 ms of multiple structures; TCH / AHS: 12 TDMA structures per 480 ms of multiple structures). The waste of the half of the available radio resource would mean that the encoding mode information could only be transmitted half the frequency as initially possible. The result is a loss of potential performance due to less adaptation of the coding mode. A further disadvantage of the application of the same diagonal interleaving for the SI D structures (transporting the encoding mode information) for the voice structures is the delay caused by this type of interleaving. With respect to achieving the best possible performance of the coding mode adaptation of the multiple mode coding systems, the transmission delay of the coding mode information must be kept to a minimum. This prohibits the use of diagonal collation. A particular problem in DTX systems is the detection of voice starts after periods of inactivity. The failure of the start results in the voice output of the decoder being held. On the other hand, if a non-transmitted structure is mistakenly detected as a voice initiation structure, undesirable knocking or knocking sounds can occur which can considerably degrade the quality of the communication. In principle, ARM systems with DTX operation only need to transmit the request from coding mode to the currently active link on the inactive link. There are no coding mode indications for the inactive link that needs to be transmitted. However, when the inactive link is activated again, an appropriate encoding mode must be selected. A solution of how to select the encoding mode for voice initiations after inactivity has to be found to ensure that the transmitting side and the receiving side apply the same mode. In addition, this coding mode must be adequate with respect to the current radio channel conditions. Apart from the signaling method of coding mode in II the AMR standard, until now there are no more control channels available. However, there is a need for such a channel in order to obtain the ability to execute rapid configuration changes (for example, changing a coding set, changing the phase of the encoding mode information in order to minimize the delay of transmission, transfer to an existing GMS encoding such as FR, EFR or HR, and / or switch to a future application such as broadband, voice and data or multimedia coding). Accordingly, there is a need for improved methods and apparatus for executing DTX and configuration changes in adaptive multiple speed systems.t.

BRIEF DESCRIPTION OF THE INVENTION The present invention satisfies the needs described above and others by providing novel solutions for DTX and fast band signaling of configuration changes and protocol messages, as well as the interaction of both operations, in the context of adaptive multiple speed systems. Advantageously, the methods described and the apparatus are efficient in terms of cost in terms of radio transmission capacity, in terms of fixed line transmission and in terms of implementation. An illustrative method for executing discontinuous transmission (DTX) in a communication system in which the source data is interleaved for transmission from a first component in the system to a second component in the system that includes the stages of detecting periods of inactivity of source data and transmit in the structures (SI D) of the silence descriptor from the first to the second component during periods of inactivity of source data, where some of the transmitted SID structures are interspersed using a different collation algorithm in comparison to that used for the source data. For example, the source data can be interleaved diagonally in blocks, and certain SI D structures can be interspersed in blocks. The illustrative method may further include the steps of transmitting a first type of structure SI D to indicate a transition from the source data activity to the inactivity of source data, periodically transmitting a second type of structures SI D during the inactivity of source data and transmit a third type of structure SI D to indicate a transition from the inactivity of source data to the source data activity. Advantageously, when the communication system is an adaptive multiple speed system (AM R), the SI D structures may include the coding mode information in addition to the silence description information. An illustrative method for transmitting the protocol messages from the first component to a second component in a voice communication system includes the step of transmitting an escape structure instead of a voice data structure, the exhaust structure including an integral bit pattern to distinguish the escape structure from the voice data structures and transfer a protocol message. The escape structure may further include a data field for indicating to the second component a particular protocol message. An illustrative method for effecting configuration changes in a communication system includes the step of transmitting an escape structure in place of a voice data structure, the escape structure including an integral bit pattern, the escape structure of voice data structures and transfer an indication of configuration change. The escape structure may further include a data field for indicating to the second component a particular configuration change to be made. For example, when the communication system is an AM R system, an escape structure can be used to change a fixed active coding mode. Alternatively, an escape structure can be used to change a phase of the encoding information. The above characteristics and advantages of others of the invention are explained in detail below by reference to the illustrative examples shown in the accompanying drawings. Those skilled in the art will appreciate that the embodiments described are provided for purposes of illustration and understanding and that numerous equivalent embodiments are contemplated herein.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 illustrates a structure transmission scheme (SID) of total velocity silence descriptor. Figure 2 illustrates a structure transmission scheme (SID) of descriptor. medium speed silence. Figure 3 illustrates an adaptive multiple speed communication system in which the present invention can be implemented. Figure 4 illustrates a SID structure format according to the present invention. Figure 5 illustrates a structure intercalation scheme Total speed SID according to the present invention. Figure 6 illustrates an intermediate velocity SID structure interleaving scheme according to the present invention. Figure 7 illustrates a first SID structure format according to the present invention. Figure 8 illustrates a speech start structure format according to the present invention. Figure 9 illustrates a scheme for inhibiting the first SID structures according to the present invention. Figure 10 illustrates a scheme for inhibiting regular SID structures according to the present invention. Figure 1 1 illustrates a total rate scheme for detecting transitions from voice inactivity to speech activity according to the present invention. Figure 12 illustrates an average rate scheme for detecting transitions from voice inactivity to speech activity according to the present invention. Fig. 1 3 illustrates a total rate scheme for detecting a speech start when a speech initiation indication structure is replaced by a system configuration change structure according to the present invention. Figure 14 illustrates an average rate scheme for detecting a voice start when a speech start indication structure is replaced by a system configuration change structure according to the present invention.

DETAILED DESCRIPTION OF THE INVENTION Although the embodiments of the invention are described hereinafter with respect to the voice transmission in the GSM system, those skilled in the art will immediately appreciate that the techniques described are equally applicable in other contexts. For example, the invention is easily applied in a wireless or fixed line communication system, including TDMA systems (eg, D-AMPS), and the Internet.

Figure 3 illustrates an exemplary AMR system in which the techniques of the present invention can be implemented. The illustrative ARM system includes a transcoding and speed adaptation unit (TRAU) and a base station (BTS) on the network side, as well as a mobile structure (MS). On the network side, a voice coder (SPE) and a channel coder (CH E), as well as a channel decoder (CH D) and a speech decoder (SPD), are connected by means of the Well-known serial A-bis interface. For each link, quality information is derived by estimating the current channel status. Based on the channel status, and also taking into account possible restrictions from network control, the coding mode control, which is located on the network side, selects the encoding modes to be applied . The channel mode to use (TCH / AFS or TCH / AHS) is controlled by the network. The uplink and the downlink always apply the same channel mode. For adaptation of coding mode, the receiving side executes the link quality measurements of the incoming link. The measurements are processed generating a quality indicator. For the uplink adaptation, the quality indicator is fed directly into the mode control unit U L. This unit compares the quality indicator with certain thresholds and generates, also considering the possible restrictions from the network control, a coding mode command that indicates the coding mode to be used in the uplink. The coding mode command is transmitted in band to the mobile side where the incoming speech signal encoded in the corresponding coding mode. For the downlink adaptation, the Generator of DL Mode Request within the mobile unit compares the quality indicator DI with certain thresholds and generates a request for coding mode indicating the preferred coding mode for the downlink. The request for coding mode is transmitted in band to the side of the network where it is fed into the DL mode control unit. This unit generally grants the requested mode. Nevertheless, considering the possible restrictions from network control, you can also cancel the request. The resulting coding mode is applied to encode the incoming speech signal in the downlink direction. For both the uplink and the downlink, the currently applied encoding mode is transmitted in band as an indication of encoding mode together with the encoded voice data. In the decoder, the encoding mode indication is decoded and applied to decode the received voice data. The coding mode selection is made from a set of coding modes (ACS, active coding set), which may include from 1 to 4 AMR coding modes.

Associated with this set is a list of 1 to 3 switching and hysteresis thresholds used by the DL mode request generator and the U L mode control unit to generate the encoding mode requests and the encoding mode commands. These configuration parameters (ACS, thresholds, hysteresis) are defined in the call set-up and can be modified in the transfer or during a call. In accordance with this invention, DTX in a system such as that shown in Figure 3 is based on band signaling with 3 different structure types: SI D_FI RST, S I D regular, and speech initiation structures. These types of structures have in common that they use particular integral bit patterns, which identify them. In addition, they can also carry payload data, which consists of CN parameters and encoding mode information. For example, the implementations according to the invention, see GSM 05.03. Digital cellular telecommunications system (Phase 2+); channel coding (draft copy ETSI EN 300 909 V7.2.0 (1999-1 1) and GSM 06.93: Digital cellular telecommunication system (Phase 2+); Discontinuous transmission (DTX) for multiple speed voice traffic channels Adaptive (AMR) (draft copy ETSI EN 301 707 V.7.2.0 (1999-1 1)), each of which is incorporated herein in its entirety by reference.The SI D structures are identified at the level The SI D structures are defined to be transmitted using k TDMA structures, that is, they consist of k * 1 14 bits, a suitable selection for k, is 4. In this case the SI D structures consist of 456 bits , ie one of the 456-bit channel structure for TCH / AFS and two channel structures of 228 bits each for TCH / AHS Each structure SI D has a structure identification field SI D containing a pattern of single bit and two message fields, the message field is reserved for the channel encodable tolerable noise parameters, the other pair to encoded channel coding mode information. The encoding mode information field can carry the encoding mode requests only, or it can be further subdivided into two parts, one carrying the requests / encoding command and the other conveying the encoding mode indications. An example of the definition of the regular SI D structure format is given in Figure 4. In this example, the SI D structure consists of a 212 D SI D structure identifier, a 212 bit field for the noise parameters tolerable and a 32-bit field for encoding mode information. In this example, it is assumed that the CN parameters are convolutionally encoded and the coding mode information consists of the commands / requests coded by block and indications. In an alternative solution the two message fields can be placed together, if, for example, the CN parameters and the coding mode information are encoded using the same convolutional or block code. According to the invention, the regular SI D structures are interspersed by blocks instead of diagonally. While this undoes the possible interleaving gain (that is, transmissions potentially less robust against transmission errors), the SI D structures generally carry less information than regular voice structures, and therefore may be protected. using channel codes more powerful than those used for voice transmission. This compensates for the loss in the interleaving gain or even makes the transmission of the SI D structure stronger than is possible for the current solutions (GSM FR, EFR or H R). Important information such as the encoding mode information can, for example, be protected by stronger channel codes (compared to the transmission in band of the coding mode information in regular voice structures). further, the CN parameters are usually represented with much fewer bits than the speech parameters. The few CN bits can therefore be protected with lower speed channel codes. As an example, out of a number of 35 CN, all can be protected, first by means of a 14 bit CRC code (which makes possible a very powerful error detection), and then using a quarter speed convolutional code (length of restriction k = 5), In addition, CN parameters and coding mode information are generally information that varies relatively slowly. Also taking into consideration the proposed structure speed SI D (of each octave structure), which is much higher than in the existing solutions, the occasional losses of a SI D structure due to channel errors are even tolerable. As shown in the respective figures 5 and 6, both for TC H / AFS and for TCH / AHS, the SI D structures consist of 4 * 1 14 bits which are mapped, according to the invention, by block interleaving on 4 structures 4 TDMA. The purpose of the interleaver is to distribute the structure bits SI D in such a way on the TDMA structures that the firmness is maximized against transmission errors. The diagonal interleaver for voice structures is used. Since deinterleaving does not have much demand in terms of complexity, this solution with a particular block interleaver for structures S I D is viable. In the worst case, the decoder executes the block deinterleaving of the SI D structure and the diagonal deinterleaving of conventional speech structure although not more than one channel decoder. Advantageously, the problem in the current GSM and PDC systems of bits wasted in the TDMA structures belonging to the SI D structures is therefore solved. For TCH / AFS, the current block intercalation system for the SI D structure is relatively minor. In order to obtain a maximum interleaver gain, the identification marker bits, as well as CN and the coding mode interference bits are distributed as evenly as possible over the TDMA structures used in the transmission. For TCH / AHS, special cases may occur due to the fact that the SID structure is transmitted using 2-channel structures. As described in detail below with respect to the SID inhibition structures, the situation may arise when the first half of the TDMA structures carrying the SID structure have been transmitted and the second half can not be transmitted due to a speech start. For this case, it is important to have the ability to inhibit the SID pattern, which has already been sent. This is ensured by transmitting the second half of the pattern bits over the odd positions of the second half of the TDMA structures. With respect to the encoding mode information it is important that the encoding mode is used to decode the voice start that is available. This can be ensured by also transmitting the second half of the encoding mode indication bits in the odd positions of the second half of the TDMA structures. One possible solution is to map the pattern bits and coding mode indication bits onto the TDMA structures using the diagonal collation. Accordingly, the CN bits and the encoding mode request / command bits are transmitted in the odd positions of the first half of the TDMA structures and in the even positions of the second half of the TDMA structures. The interleaving scheme described for the SID structures on TCH / AHFS is illustrated in Figure 6. According to the invention, the particular SID_FIRST structures are transmitted immediately after the last voice structure when going from activity to inactivity. The solution is only to identify the end of the voice instead of also transmitting the CN parameters. An example solution for TCH / AFS is the use of a 228-bit field consisting of 212 marker bits and 16 bits for coding mode information, as shown in Figure 7. The coding mode information is the request / command or indication, depending on which is in turn (if a voice structure has been transmitted). The type of information of the coding mode transmitted with the SID_FIRST structure therefore depends on the structure number and the transmission phase of the coding mode information. A special interleaver maps the SID_FIRST structure over the 228 bits available in the unused mid-burst. Figure 5 illustrates the transmission scheme of the structure SID_FIRST for TCH / AHS. Note that there are no average bursts wasted. An analogous solution for TCH / AHS would transmit an identification pattern SID_FIRST and the coding mode information on the two average bursts usually not employed. An example that makes detection of SID_FIRST more reliable is, use the following 2 2TDMA structures. This means that two channel structures SID_FIRST and SID_FIRST_2 are transmitted. A possibly identical 228-bit structure as used in the solution of the TCH / AFS sample (consisting of 21 12 marker bits and 6 bits for encoding mode information, see Figure 7) is mapped over the even positions of the TDMA structures , which carry the last voice structure (medium bursts not used), and on the odd positions of the two subsequent TDMA structures. This type of diagonal mapping allows the application of the existing diagonal deinterleaver. The information of the coding mode is request / command or indication, depending on the number of structures and the transmission phase of the coding mode information. That kind of information is transmitted in coding mode, which would have been set in the respective channel structure if the voice had been transmitted. The mapping is done in a way that the equal portions of the pattern bits and the encoding mode information bits are placed in the first two and the two TDMA structures used. Figure 6 illustrates a technique for increasing the reliability of the structure detection SI D_F I RST in an additional manner. According to the invention, the even positions of the two TDMA structures are read with an additional identification pattern. It is also possible to use a part of these medium bursts for the transmission of information in coding mode. The identification pattern could also be the code word of the code mode information repeated so frequently that all available bits are used. If, for example, 14 bits are available and the code word for the coding mode information is 16 bits wide, then it would be repeated 114/16 times. The diagonal intercalation used for voice structures implies that the odd positions of the first half of the TDMA structures carrying the first voice structure after a period of inactivity are free for other purposes. A solution that improves the start detection according to the invention, is to fill those bits with a special start identification pattern. In addition, the parts of these bits can also be used for transmission of a coding mode indication that indicates the coding mode according to which the first speech structure is encoded. One solution that transfers a start bit pattern and the encoding mode indication is to repeat the encoding mode indication code word more frequently than all the available bits that are used, as illustrated in FIG. 8. An example for TCH / AFS it is to repeat the word of bit code 16 of the indication 228/16 times. For TCH / AHS, the codeword of bit 16 is repeated 114/16 times. Such start structure is mapped by a particular interleaver, over the otherwise unused medium bursts. The respective structure transmission schemes of TCH / AFS and TCH / AHS are illustrated in Figures 5 and 6. For TCH / AHS, regular SID structures and SID_FIRST structures are transmitted using 2 channel structures. The situations that can occur therefore in which a high priority voice start is transmitted after the first but before the second of the second channel structure of the SID structure that has been transmitted. In such a case, the error should occur so that the receiver loses the start and instead detects a SID or, respectively, a SID_FIRST structure, even though currently only receives the first half of it. To help avoid this problem, a special SID_FIRST inhibit structure is used instead of a regular start structure when the first half of the TDMA structures carries the SID_FIRST that has been sent first although the second half can not be sent due to a start of voice. The pattern bits belonging to the second half of the SID_FIRST structure, which would have been transmitted, are now inverted. This inhibits the detection of the entire SID_FIRST pattern in the receiver. The information bits of the coding mode remain identical, as from the original SID_FIRST structure. The receiver will obtain a structure that can not be used in the described situation. It is useful to hide this structure using the appropriate error reconciliation (EC) techniques. The described case is illustrated in figure 9. Another special structure, ie a structure of inhibition SID is used in place of a regular SID structure when the first of the TDMA structures carrying the SID has been sent although the second half can not be sent due to a voice start. The pattern bits that belong to the second half of the SID structure that would have been transmitted are now reversed. This inhibits the detection of the entire SI D pattern in the receiver. The information bits of the coding mode, which represent an information of the coding mode remain identical, from the original SI D structure. The receiver will obtain a structure that can not be used in the described situation, for which purpose CN generated will continue using the previous CN parameters. The receiver can also verify the patterns that are transmitted in this special case in order to detect the beginnings of voice with improved reliability. The case described is illustrated in Figure 10. According to the invention, the SI D structures are transmitted during inactivity to each nFR structures (TC H / AFS) and, respectively, each ΔHR structures (TCH / A HS). A suitable option is nFR = nHR = 8. The phase-aligned transmission and the decoding of the SI D structures (alignment deduced from SACCH, as in the current GSM system) is an existing solution in the current GSM system, which helps to achieve a good performance of decoding structure SI D. However, the identification of structure SI D proposed based on the integrated bit patterns provide that high performance of structure detection more than flexible solutions without a fixed base that are possible. An example is to initiate the transmission of the SI D structures with the third structure after the transmission of the SI D_FI RST pattern and then to transmit the SI structures D each eighth structure. An alternative solution is the transmission SI D as incrona (ie not aligned with the fixed time structure). As an example, structures S I D are transmitted whenever a mode request changes, possibly with the restriction that a certain maximum of the TDMA structures transmitted by the 480ms multiple structure has not been exceeded. Another improved solution can transmit a SI D structure if the CN parameters are changed significantly and a maximum of the TDMA structures transmitted by 480 ms of the multiple structure have not been exceeded. Such solutions with the asynchronous transmission of the SI D structure may fall back to the time-aligned transmission provided that certain minimum transmission requirements per time interval have not been met. Note the different bit patterns that are sent to identify the different types of structures can be partially altered by transmission errors. In order to ensure reliable detection of patterns also in the presence of channel errors, correlation techniques can be used. One possible solution is to count the number of coupling bits when comparing the received bits with the patterns. As an example, if 70% bits match, then the receiver may consider the pattern as localized. An alternative solution that uses the uniform bit information is to accumulate the received uniform bits with a positive sign if the corresponding bit of the pattern is one and with a negative sign if the corresponding bit is zero. This accumulated measurement can be normalized by the product of the length of the pattern and the maximum possible uniform bit value. If the normalized measure exceeds a certain threshold, for example, 0.4 the receiver may consider that the pattern was found. An additional criterion, which can be used for SID structures, is the CRC of the CN bits. If there is a CRC error, the structure is not considered as a valid SID structure. For cost reasons, it is desirable that the identification patterns do not require much memory to store. As an example, the identification pattern for SID_FIRST and SID for TCH / AFS can be constructed by repeating the short 9-bit sequences ceil ((228-16) / 9) = 24 times and then discarding the last 4 bits. Such a 9-bit sequence is, for example,. { 0, 1, 0, 0, 1, 1, 1, 1, 0.}. . For TCH / AHS, it is also important to avoid the possible decoding in the structure SID_F1RST as a regular SID structure and vice versa. Therefore, the identification patterns for SID and SID_FIRST are as different as possible. As an example, the pattern for the SID_FIRST structure can be identical to the pattern used for TCH / AFS. The pattern used for regular SID structures can then be constructed by inverting the SID_FIRST pattern. The solution to transmit only a special bit pattern and the coding mode information in the structure SID_FIRST instead of transmitting the CN parameters also helps to maintain the DTX efficiency at a maximum (ie the activity at the air interface is kept to a minimum). At the same time, the reliability of the detection of the identification pattern can be maximum since all the available bits are used by the bit pattern (except those used for the transmission of the information of the coding mode). A problem with this is that the receiver does not obtain a set of CN parameters for the CN generation during the period from the end of the voice until the reception of the first regular SID structure. The solution is to derive the CN parameters locally in the receiver using the voice parameters of the last n structures before the end of the voice. Usually, the encoder operates with persistence, that is, even that the VAD detects voice inactivity as a number of n structures is still encoded as a voice. The decoder can therefore derive the CN parameters locally, by, for example, the average of the gain and the spectral parameters of the persistence structures, ie n = m. Other solutions applied the last received set of CN parameters from a previous period of inactivity. According to the invention, an AMR receiver incorporates a 2-state model with the activity and inactivity of states. The purpose of this state model or to carry the distinction of voice structure / SID / non-transmitted. The transition from activity to inactivity requires the detection of a SID_FIRST structure that follows the voice structures. Moving from inactivity to the activity state requires detecting the speech initiation identification pattern and a valid first speech structure that can be decoded without CRC error and optionally, exhibiting quality measurements are, for example, derived from the decoder of receiver / channel and that exceeds certain thresholds. An example is SFQ measurement (integral bit error estimation), which must be below some threshold. The reliability of this state transition can be increased with the restriction of more than one of the structures must be decodable if CRC error and, optionally, does not exceed a certain SFQ measure. Another criterion, as illustrated in Figure 11, which helps to adequately detect transitions from inactivity to activity is that of the structures received immediately after the SI structures that can never be a voice structure, condition that block interleaving is used for SI D structures that require less delay than diagonal collation for voice structures. Figure 12 illustrates this criteria for the TCH / AHS example. Another way to improve the detection of the first voice structures and help distinguish them from non-transmitted structures is to access measurements from other components of the receiver (for example, the RF receiver or the RF equalizer). Examples for such measurements are estimates of carrier resistance and interference and derived measurements such as the C / I ratio. A further way to improve both SI D_FI RST and the identification performance of the first speech structure is to transmit the TDMA structures that carry them against increased transmission power. According to the invention, the following solutions are suitable for defining the coding mode for the speech starts after a period of inactivity: (a) Selection of the strongest coding mode or, alternatively, with the coding mode firm n-th. The safest solution is to select p = 1. The coding mode indication does not need to be transmitted. The disadvantage for n = 1 is that, for the appropriate channels, a too strong coding mode with low intrinsic speech quality is selected. (b) The selection of the same encoding mode than that of the currently active link. This is motivated by the fact that the uplink and downlink channel qualities are similar. The transmitting side of the link that summarizes the voice transmission applies the encoding mode it is requesting for the currently active input link. The receiving side of the link that becomes active again knows the applied coding mode since it is identical with the encoding mode requests that it is receiving for the application in the currently outgoing active link. The scheme can be made firmer if a mode is selected for voice initiations that is n (for example p = 1) firmer modes than the currently active link mode (provided that this stronger mode exists). (c) Selection of the same coding mode that was selected at the end of the last voice period that precedes the period of inactivity. This is motivated by the fact that radio channel conditions generally do not change very fast. The scheme can be made firmer by selecting a mode for voice initiations that is n (for example n = 1) firmer mode than the mode that was used at the end of the last speech period (provided that said mode firm exists). (d) The selection according to the measurements of the inactive link. As the transmission on the inactive links is not completely stopped, link quality measurement is possible. The corresponding measurement reports or requests / encoding mode commands are transmitted over the active link. When the inactive link summarizes the voice transmission, a coding mode corresponding to the last received mode request is selected. Advantageously, solutions (a), (b) and (c) above can make use of the fact that there are no coding mode requests for the idle link that need to be transmitted. The active link can therefore save transmission capacity for the encoding mode request and use it for some other purpose. As an example is the use of this transmission capability for a better protected transmission of the coding mode indications. In addition to the techniques described above for executing DTX in AM R systems, the invention also provides techniques for executing rapid configuration changes in AMR systems. The purpose of these techniques is to allow quick configuration changes that can not be made using the existing slow control channels. In addition, existing control channels can not ensure those configuration changes that are synchronized with the transmission of voice data. Like the DTX mechanism, described above, the configuration change mechanism is based on in-band signaling. The applications for example, are in connections with free tandem operation (TFO) the change of the active coding set and the change of the phase of the coding mode information (in order to minimize the transmission delay). Additional general applications in transfers to one of the existing GSM encodings (FR. EFR, H R), or to switch to a future application such as, for example, broadband, voice and data or multimedia coding.

Like the DTX mechanism, the configuration change mechanism is described with respect to TCH / AFS and TCH / AHS in the GSM system, although it is equally applicable in other contexts.

The mechanism of configuration change is based on the capture of structure similar to the well-known FAC structure capture. (that is, voice structures that are replaced by configuration change structures), and hence referred to as escape signage. Since the escape signaling mechanism is used only occasionally during a connection and only a few structures can be captured, the error reconciliation unit in the receiver is capable of being virtually undetectable structure capture. According to the invention, the escape structures are of similar format to the SI D structures described above. They are identified at the full bit level by a particular individual identification pattern. Like the SI D structures, these include the pattern and one or more message fields. One field carries the encoded escape message of the current channel and the other the encoding mode information. As an example, the escape structure may include 456 bits and be of exactly the same structure format as the SI D structures (see for example figure 4), where the CN field is replaced by the escape message. The payload to be tratted through the escape mechanism is called the escape message. The escape messages constitute a number of net bits, which can be grouped for logical units. For example, implementations according to the invention, see GSM 05.09: Digital cellular telecommunication system (Phase 2 +); Link adaptation (draft copy ETSI EN 301 709 V7.1 .0 (1999-1 1), which are incorporated herein by reference in their entirety.

Escape messages can be encoded per channel with any suitable channel coding scheme, such as block or convolutional coding. A cost-efficient solution is the use of exactly the same channel coding that was used for the CN parameters in the SI D structure as described above. This means that, following the example solution described above with 35 CN bits, that net 35-bit escape message is protected with a 14 CRC bit and then coded convolutionally with a VA code rate. and the restriction length K = 5. As with the SI D structure, the encoding mode information field can carry the coding mode indication coded by block or convolutionally, and the coding mode command / request. The escape structures are interspersed diagonally by block, like voice structures. This implies, assuming the example solution, with an exhaust structure of 456 full bits, that an escape structure replaces a voice structure in TCH / AFS and two voice structures in TCH / AHS. For TCH / AHS, there are not necessarily two consecutive structures, although it is assumed in the example solution described. It is not advantageous to capture two consecutive structures for error reconciliation in order to hide the capture. On the other hand, the capture of two consecutive voice structures is beneficial in terms of the trassion delay of the escape message. The collation is made of such. so that the first half of the escape structure (228 bits, see figure 4) replaces the first voice structure. It is important that this first half contain the escape identification pattern. This allows the receiver to verify this pattern. After finding the pattern, the receiver is able to locate the second captured voice structure, which carries the second half of the escape structure. In order not to interfere with the regular trassion of the coding mode information, the interleaver can additionally map one of the codeword information of the encoding mode over bit positions of the first captured speech structure. Accordingly, the other codeword information of the encoding mode is mapped into bit positions of the second captured speech structure. In addition, the coding mode information placement, i.e., the coding mode indication and the requests / commands in the coding mode field is made with respect to the coding mode information phase during the transmission of structures of regular voice. If, for example, the first half of the exhaust structure replaces a voice structure that would be transported in a coding mode indication, then this first half of the exhaust structure has yet to transmit an indication of coding mode. Note that the above described escape mechanism can also be used in conjunction with the DTX mechanism described above. Therefore, according to the invention, the escape structures can replace not only voice structures, but also all other types of structures, namely SI D_FI RST, regular SID, NoTX, and voice initiation structures. Consider the case that an escape structure is going to be sent during a period of inactivity, and it is efficient in terms of the use of transmission resources to apply block interleaving, as it is done for SID structures. However, since the escape mechanism is indicated to be used only occasionally, the use of transmission resources is not the most important criterion. Instead, efficient implementation in terms of cost and low complexity are important. Therefore, a beneficial solution to maintain the structure format, channel coding and diagonal block interleaving, which is also used for the escape structures during the speech. Note that using the block diagonal collation for the escape structures during DTX implies that there are medium bursts not defined by the intercalation. For TCH / AFS the odd positions of the first 4 bursts the even positions of the last 4 bursts that carry the escape structure are not defined. Undefined bits are not a problem in themselves, although the following problem can be solved by setting the undefined positions appropriately. Consider the case of a voice initiation. As described above, a speech start structure is marked with a start pattern that allows for the best identification of structure as a start and identification of the coding mode used for the start speech structure. If an escape structure must be sent at the same time, it will replace the start structure. Therefore, for subsequent voice structures, it will be more difficult to identify them as voice structures since the start pattern was captured. According to the invention, this problem is avoided by filling the first half of the undefined bits (odd positions) with the start pattern, regardless of whether there is a start or not. For the case in which there really was no start, it needs to be noted that the inactivity continues. Sending SID_FIRST immediately after the escape structure performs this action. This defines the second half of the otherwise unused bits (even positions). This solution is also beneficial in the implementation costs. It allows the management of the escape structures, apart from the channel coding, exactly as if it were a voice. Figures 13 and 14 illustrate the solution described with respect to TCH / AFS and TCH / AHS respectively. Note that voice structures, which have been captured for escape purposes, can not be reprogrammed for transmission after the escape, since this would increase the speech transmission delay. Nevertheless, the SID structures, which are affected by the transmission of the exhaust structure can be reprogrammed for transmission immediately after the transmission of the exhaust structure. Advantageously, this helps maintain a high. subjective tolerable noise signal quality. Illustrative solutions are provided in GSM 06.93 cited above.

In order to ensure the correct reception of the escape messages and to define the appropriate routines for the error events, an escape protocol is proposed. For example, the solutions are provided in GSM 05.09 cited above. Those skilled in the art will appreciate that the present invention is not limited to the specific illustrative embodiments that have been described herein for purposes of illustration and that numerous alternative embodiments are also contemplated. The scope of the invention is therefore defined by the claims appended hereto, rather than by the foregoing description and all equivalents that are compatible with the meaning of the claims are intended to be encompassed therein.

Claims (28)

1. A method for executing discontinuous transmission (DTX) in a communication system in which the source data is interleaved for transmission from a first component in the system to a second component in the system, the method comprising the steps of: detecting periods of inactivity of source data; and transmitting the structures (SID) of the silence descriptor from the first to the second component during periods of inactivity of the source data, where certain transmitted SID structures are interspersed using a different collation algorithm compared to that used for the data source.

The method according to claim 1, characterized in that the source data are interleaved diagonally by block, and where certain SID structures are interleaved per block.

The method according to claim 1, characterized in that the SID structures include parameters (CN) of tolerable noise.

The method according to claim 1, which includes the steps of: transmitting a first type of SID structure to indicate a transition from the source data activity to the source data inactivity; periodically transmit a second type of SID structure during the inactivity of the source data; and transmitting a third type of structure S I D to indicate a transition from the inactivity of source data to the source data activity.

The method according to claim 1, characterized in that the communication system is an adaptive multiple speed system (AMR) and wherein the SI D structures include the coding mode information in addition to the silence description information.

The method according to claim 1, characterized in that each structure SI D includes a bit pattern to distinguish the SID structure from the source data structures.

The method according to claim 6, characterized in that the bit patterns are integral bit patterns.

8. The method according to claim 1, characterized in that the source data is voice, and wherein the communications system is one of a Time Division Multiple Access Wireless System (TDMA), a Wireless System of Multiple Access by Division of Frequency (FDMA) and a Wireless System of Multiple Access by Code Division (CDMA).

The method according to claim 1, characterized in that the escape structures are transmitted to effect the configuration changes, and wherein an escape structure can replace a source data structure, an SID structure, or a structure ( NoTX) without transmission.

10. The method according to claim 9, characterized in that the SID structures are interleaved per block, and wherein the escape structures are interspersed diagonally per block.

The method according to claim 9, characterized in that the communication system is an adaptive multiple speed system (AMR), and wherein an escape structure is used to change a coding mode set.

The method according to claim 9, characterized in that the communication system is an adaptive multiple rate (AMR) system, and wherein an escape structure is used to change a phase of the encoding information.

The method according to claim 1, characterized in that the active speech source data are interleaved diagonally and that and the unused bits within the interleaving scheme for a last speech structure are used for a specific bit pattern to mark the end of the voice, and where the unused bits within the interleaving scheme for a first structure are used for a specific bit pattern to mark the start of the speech.

14. In a voice communication system in which the voice data is transmitted from a first component and to a second component as a method for transmitting the protocol messages to the second component, comprising the step of: transmitting an escape structure instead of a voice data structure, the escape structure including a full bit pattern to distinguish the escape structure from the voice data structures and conveying a message of protocol.

The method according to claim 14, characterized in that the escape structure further includes a data field for indicating to the second component a particular protocol message.

16. A communication system in which the voice data is transmitted from a first component and towards a second component, a method for effecting configuration changes, comprising the step of: transmitting an escape structure instead of a voice data structure, the escape structure including an integral bit pattern for distinguishing the escape structure from the voice data structures and conveying a configuration change indication.

The method according to claim 16, characterized in that the escape structure further includes a data field for indicating to the second component a particular configuration change to be made.

18. The method according to claim 16, characterized in that the communication system is an adaptive multiple speed system (AMR), and where an escape structure is used to change a coding mode set.

The method according to claim 16, characterized in that the communication system is an adaptive multiple speed system (AMR), and where an escape structure is used to change a phase of the coding structure.

The method according to claim 16, further comprising the step of: transmitting the silence descriptor structures from the first and the second component during periods of source data inactivity, where an escape structure can replace a source data structure, a SI D structure or a structure (NoTX) without transmission. twenty-one .

The method according to claim 20, characterized in that the SI D structures include parameters (CN) of tolerable noise.

The method according to claim 20, characterized in that the source data structures and the escape structures are interleaved diagonally by block, and where some of the transmitted SI D structures are interleaved per block.

The method according to claim 20, which includes the steps of: transmitting a first type of SID structure to indicate a transition from the source data activity to the source data inactivity; periodically transmit a second type of SID structure during the inactivity of source data; and transmitting a third type of SID structures to indicate a transition from the inactivity of source data to the source data activity.

The method according to claim 20, characterized in that the communication system is an adaptive multiple speed system (AMR) and wherein the SID structures include the coding mode information in addition to the information of the silence description .

25. A voice communication system, comprising: a first component that transmits voice data structures interspersed; and a second component that receives the interleaved data structures, wherein the first component detects periods of voice inactivity and transmits the silence descriptor structures instead of the voice data structures during the periods of voice inactivity, and in where at least some of the SI D structures are interspersed using a different collation algorithm compared to that used for voice structures.

26. The method according to claim 25, characterized in that the speech structures are interspersed diagonally by blocks, and wherein certain SI D structures are interleaved per block.

27. A communications system, comprising: a first component that transmits source data; and a second component that receives the source data, wherein the first component transmits an escape structure instead of a source data structure to indicate a configuration change to the second component, and wherein the escape structure includes a bit pattern integral to distinguish the escape structure from the source data structures.

28. The method according to claim 27, characterized in that the escape structure further includes a data field for indicating to the second component a particular configuration change to be made. SUMMARY Techniques for discontinuous transmission and fast-band signaling of configuration changes and protocol messages in the voice communications system provide cost efficiency in terms of radio transmission capacity, in terms of fixed-line transmission and in terms of the implementation effort. An illustrative method for executing discontinuous transmission in a communication system in which the source data is interleaved for transmission from a first component in the system to a second component in the system and includes the steps of detecting periods of inactivity of the source data and transmit the structures of the silence descriptor from the first to the second component during periods of inactivity of the source data, certain transmitted SI D structures that are interleaved using a different collation algorithm compared to that used for the source data. For example, the source data can be interleaved diagonally per block and certain SI D structures can be interleaved per block. An illustrative method for effecting configuration changes in a communication system includes the step of transmitting an escape structure instead of a voice data structure, the escape structure including a full bit pattern to distinguish the structure of the speech structure. Exhausting the voice data structures and conveying a configuration change indication, the escape structure may further include a data field to indicate a particular configuration change to be made. For example, when the communication system is an AMR system, an escape structure can be used to change a set of active coding mode. Alternatively, an escape structure can be used to change a phase of the encoding information.