Patent Application
20060234741
The present invention relates to the control of multichannel or multimedia data communications from a mobile station to a network, particularly in a GSM mobile telephone system, also known as GERAN—GSM/EDGE Radio Access Network
In earlier embodiments of GSM, the channel coding to be used in the uplink is selected by the network. This is true both in the case of the selection of the codec mode for the adaptive multi-rate (AMR) speech codec and in the case of the selection of the coding scheme for GPRS and EGPRS. The selection is made on the basis of the condition of the uplink—that is, the transmission path from the mobile station to the network—and is signalled to the mobile station. If the uplink is in a good condition, large amounts of data may be transmitted. The mobile station did not participate in defining the channel coding scheme to be used in the uplink.
In earlier embodiments of GSM/GPRS, communication between a network and a mobile station is defined as a single or a multitude of temporary block flows (TBFs), a TBF being a set of procedures defining protocols for data transfer, data acknowledge and so on. At any moment in time, only one TBF is allowed to transmit data in the uplink. The network would send a USF (Uplink State Flag) signal to the mobile station. This signal informs the mobile station of which TBF is given permission to use the radio channel. The mobile station did not participate in defining the TBF allowed to transmit data. This is possible to manage for the network because only one TBF can be transmitted in a single radio block.
A recent improvement to the GSM service is the proposed Flexible Layer One (FLO). This improvement will allow data belonging to different TBFs to be transmitted from the mobile station to the network in a single radio block. For example, the different TBFs could correspond to voice data, control data, an image file, video data, interactive web-based service data, or unidentified user data for transfer across the network. Each of these data types will have a defined priority, and tolerated delay. For instance, if transmitting a static image file, it is important that all data is received, but it is of relatively little importance whether the complete file takes a long time to arrive at its destination. On the other hand, when transmitting voice data, it is important that the data arrive with as short a delay as possible, whereas the loss of occasional data is relatively unimportant. Control signalling between the mobile station and the network, on the other hand, must be received quickly and accurately. Similar criteria could be defined for other data types.
For example, control data, voice and video data may be multiplexed together into a single radio block, and all data types may meet the requirements for quality and timely delivery. Should the uplink quality degrade, it may no longer be possible to transmit all of these types of data together. A decision will need to be taken as to whether to transmit, for example, just the control data, the control data and the video data or the control data and the voice data.
In known versions of FLO, the transmitter sends, together with the data, a TFCI (Transport Format Combination Indicator) signal to the receiver, to inform it of the particular Transport Format Combination (TFC), i.e. combination of data types, that has been used during transmission. A major drawback with this system is that the TFC to be used by the mobile station cannot be selected by the network, as the network is unaware of the types and quantity of each data type that the mobile station has to transmit. Therefore, it is desirable that the mobile station should be involved in deciding the data transmission format to be used.
In the UTRAN system, the scheduling of uplink data is under the control of the mobile station. Further details may be found in 3GPP TS 25.133, 3GPP TS 25.321 and 3GPP TS 25.331, available from the internet site www.3gpp.org. The mobile station handles the dynamic control of the uplink formatting, although the base station has the possibility to limit the selection available to the mobile station in a semi-static fashion. The TFC to be used in the uplink will to some extent depend on the conditions of the radio channel. The system is arranged such that the base station received equal signal power form all mobile stations. Voice, data and other services have different power requirements which may be accommodated within a fixed received power level. The power transmitted by the mobile station is adjusted, by means of a feedback loop, so that the power received by the base transceiver station (BTS or Node B) is approximately constant, equal to a value set by the network. For the UTRAN system, this feedback loop may operate at a speed of 1500 Hz. This enables the mobile station to obtain an estimate of the uplink channel conditions from the transmit power commanded by the network, offering effective TFC adaptation to uplink channel conditions. However, the corresponding feedback loop in GSM/GERAN operates only at approximately 2 Hz. This is too slow for effective TFC adaptation to uplink channel conditions. Therefore, a system similar to that used in the UTRAN system could not be used in the GSM/GERAN system. A criterion based on the estimated mobile station transmit power, as used for UTRAN, is suitable for a CDMA system (where power is the common shared resource) but would not work in a TDMA system such as GSM/GERAN. Therefore, different criteria need to be defined for the GSM/GERAN mobile station to decide whether a TFC is available for use or not.
The present invention accordingly seeks a method for providing an effective mechanism to enable a mobile station to select an appropriate transport format combination (TFC) based on uplink channel conditions.
For Flexible Layer One (FLO) operation in GSM/GERAN, it has been agreed in 3GPP that the TFC to be used in the uplink is selected by the mobile station. However, in order to select the best TFC during every Transmission Time Interval (TTI), the mobile station should have an estimate of the radio channel conditions, such as the channel-to-interference ratio CIR. The problem is that the channel conditions in the uplink are not known to the mobile station, but only to the network. Therefore, in the algorithm proposed in this application for the GERAN, the mobile station will need to rely on assistance provided by the network to perform a selection.
A somewhat similar scheme for application to UTRAN communication systems is described in Tdoc R1-02-1277 presented at the TSG-RAN WG1 #29 meeting in November 2002 and available from the internet site www.3gpp.org.
The invention accordingly provides, in a mobile communications system comprising a network and at least one mobile station, a method for selecting a transport format combination TFC to be used for communication from the mobile station to the network, over a channel of variable quality. The method comprises the steps of, in the network:
The step (h) of selecting one of the transport format combinations may be performed with regard to the type of data to be transmitted by the mobile station.
The transport format combinations preferably enable transmission of data blocks containing data from different TBFs in each block.
Calculation of the existing quality of the channel of variable quality may be performed periodically during communication.
The relative channel quality may be calculated as the minimum channel quality required such that data sent on the channel is received with an error ratio below a defined threshold.
The step (c) of indicating transport format combinations and channel quality requirements to the mobile station may include the steps of:
The indication of the existing quality of the channel of variable quality may be communicated to the mobile station by inband signalling, whereby the indication of the existing quality of the channel of variable quality is included in every downlink radio packet, in data locations normally assigned for carrying user information.
The indication of the existing quality of the channel of variable quality may be communicated to the mobile station by inband signalling, whereby the indication of the existing quality of the channel of variable quality is split into sections, respective sections being transmitted in respective successive downlink radio packets, in data locations normally assigned for carrying user information. In certain embodiments, two, or six, or eight data bits of each radio packet are employed for communication of the indication of the existing quality of the channel of variable quality.
The indication of the existing quality of the channel of variable quality may be communicated to the mobile station by inband signalling, whereby the indication of the existing quality of the channel of variable quality is split into sections, respective sections being transmitted in respective successive radio bursts. In certain embodiments, two data bits of each burst are employed for communication of the indication of the existing quality of the channel of variable quality.
The indication of the existing quality of the channel of variable quality may be communicated to the mobile station using the slow associated control channel SACCH, whereby two bits of the SACCH header are employed to transmit the indication of existing quality of the channel of variable quality, over a corresponding number of SACCH messages.
The indication of the existing quality of the channel of variable quality may be communicated to the mobile station using a dedicated channel provided in parallel with the slow associated control channel with embedded enhanced power control SACCH/TP for signalling the indication of existing quality of the channel of variable quality. The signalling may be performed over a number of SACCH/TP bursts, employing twelve bits per SACCH/TP burst. The signalling may recommence at every fourth SACCH/TP burst.
The present invention also provides a communications system arranged to operate according to the method described. The present invention also provides a network arranged to operate within such a communications system. The present invention also provides a mobile station arranged to operate within such a communications system.
The above, and further, objects characteristics and advantages of the present invention will become more apparent from consideration of the following description of certain embodiments, given by way of examples only, with reference to the accompanying drawings, wherein:
A basic description of an example of the invention follows. A background description of TFC selection is available in 3GPP TR 45.902, presented at 3GPP TSG GERAN#15 in June 2003 and available from the internet site www.3gpp.org.
The present invention accordingly provides a method of operation of transmission of data blocks containing data from different TBFs in each block, wherein the transmission channel is of variable quality. The invention is of particular application to the transmission of data from mobile stations by the GSM—or GERAN, system.
The present invention relates to a network assisted uplink TFC (transport format combination) selection. A transport format combination defines the types and rates of different types of data that may be transmitted in a “flexible layer one” by a mobile station to the network. Each transport format combination TFC will have a certain requirement for data rates and delays, and this will define a minimum quality of radio condition required for use of that TFC. TFC's may be ranked according to the radio conditions, that is to say, the signal quality of the uplink, required to use the particular TFC. As illustrated in
In the illustrated example, the signal quality is assumed to be just sufficient to allow TFC#6, the TFC identified by TFCI=5, to be used. Since the TFCs TFC#1 to TFC#5 identified by TFCI=0 to 4 each require a lower signal quality than TFC#6, then each of these TFCs is also available for use. Conversely, since the TFCs TFC#7 to TFC#10 identified by TFCI=6 to 9 each require a higher signal quality than TFC#6, then none of these TFCs is available for use. This situation is illustrated in
The ranking of the TFCs is performed by the network. The ranking must be communicated to the mobile station. This may be performed by transmitting the definitions of the TFCs to the mobile station in order of ascending or descending TFCI, that is to say, in order of ascending or descending required quality of the required radio uplink, so that the mobile station may store the TFC definitions in the correct position in the stack shown in
In operation, part of the network, such as the base station subsystem BSS, determines the current radio conditions of the uplink. This is performed in a manner known in itself to those skilled in the art. The network determines the TFC of value of pointer 10 indicating the highest allowed TFC which would be capable of being transmitted effectively with the prevailing radio channel conditions. The network transmits this TFCI to the mobile station on the downlink. This transmission may be made using one of the methods described later on the application.
The network—typically in the base transceiver station—will calculate the highest allowed TFC by applying an algorithm to the measured values representing the quality of the radio transmission. For example, the measurements may represent the received signal strength (RXLEV), the bit error probability (BEP), the block error ratio (BLER). This calculation may be performed by the Medium Access Control (MAC) layer of the base station, that is, part of the control software employed by the base station to control signalling to the mobile station.
The base station may perform similar calculations in order to select the TFC to be used in the downlink, that is, in the transmissions to the mobile station. For these calculations, the measurements used will need to be provided by the mobile station to indicate the radio conditions—signal quality—of the downlink.
The mobile station is then free to select among the allowed TFCs—the shaded ones in the example of
The signalling method used by the network should be part of the 3GPP standards for compatibility between mobile stations and network equipment from different manufacturers. The rules and criteria used by the MAC layer should also be standardised.
An aspect of the present invention relates to the methods used to signal the highest allowed TFC to the mobile station. The present invention provides numerous alternative methods for signalling the value of TFCI indicating the highest allowed TFC to the mobile station.
Three principal methods for the network to signal the value of TFCI indicating the highest allowed TFC to the mobile station are: inband signalling; using bits from the SACCH channel associated with the dedicated traffic channel; and use of a new signalling channel, associated with the slow associated control channel with embedded enhanced power control SACCH/TP, to replace the EPCCH. Each of these alternatives will be discussed in turn, along with alternative ways of communicating the TFCI, either by absolute value or by relative value.
“Inband signalling” means that the TFCI corresponding to the value of TFCI indicating the highest allowed TFC is included in every downlink radio packet, in data locations normally assigned for carrying user information, such as voice or video signals. One radio packet is sent every 20 ms. Using inband signalling has the advantage that a new value of TFCI indicating the highest allowed TFC could be signalled to the mobile station every 20 ms, and therefore the adaptation to the uplink channel conditions is very fast. One disadvantage of this arrangement is that signalling in each radio packet will consume radio resources and will lead to a degradation of performance. For example, a higher carrier to interference ratio (CIR) will be required in order for the transport blocks carried in the radio packet to achieve the same block error rate (BLER).
For certain applications, fast adaptation of the value of TFCI indicating the highest allowed TFC may not be necessary, in which case inband signalling may lead to a waste of resources.
A possible compromise could be to spread the bits of the value of pointer 10 indicating the highest allowed TFC over several radio packets. Assuming that an encoded TFCI sequence is made up of N bits, and that n of those bits are transmitted in each radio packet, then the adaptation period (i.e. the time required to transmit a new value of the TFCI) is:
┌N/n┐×20 ms, that is, r×20 ms, where (r−1)·n<N≦r·n, and r is an integer.
For the purposes of the present description, it will be assumed that N is always equal to the number of encoded bits corresponding to the 5-bit TFCI (e.g. 36 bits in the case of GMSK full-rate channels); however, it could also be decided that N should be equal to the size of the TFCI used in the downlink. Concerning the value of n, one possible arrangement is such that only two bits are “stolen” from each radio packet (n=2). This means that 18 radio packets are needed to send a 36-bit encoded TFCI, which corresponds to 360 ms. Alternatively, either 6 or 8 bits could be “stolen” from every radio packet. This would make the signalling faster. In the case of GMSK full-rate channels, the case of n=8 could also be achieved by stealing two bits from each radio burst. These two bits could be—for example—the bits adjacent to the TSC (i.e. the stealing flags). However, this would require some changes to the current interleaving algorithm for the flexible layer one (FLO).
Inband signalling has the disadvantage that that fewer bits in a radio packet are used to carry useful information, and therefore the physical layer performance will be somewhat degraded. However, if the number of bits used to signal the value of TFCI indicating the highest allowed TFC is small, then the degradation will be limited.
An alternative arrangement for signalling the value of pointer 10 indicating the highest allowed TFC to the mobile station is by using bits from the SACCH (slow associated control channel) associated with the dedicated traffic channel. In a SACCH message, 2 spare bits are currently available; they are contained in the SACCH header, as shown in subclause 7.1.1 (for the A/Gb mode of the GERAN) and 7.1.2 (for the Iu mode of the GERAN) of 3GPP TS 4.004, available from the internet site www.3gpp.com. Each SACCH message is therefore limited to carry only 2 bits. On the other hand, signalling a TFCI value would typically require 5 bits (when sent uncoded). Therefore it is not possible to signal the value of TFCI indicating the highest allowed TFC in a single SACCH message. One possibility could be to use 3 SACCH messages. In this case the update period, being the time that elapses between the reception of two successive commands, is 3×480 ms=1440 ms.
According to a further alternative, the slow associated control channel with embedded enhanced power control SACCH/TP may be used, similarly to the known operation of Enhanced Power Control (EPC). Instead of using the EPCCH associated with it, a new channel may be provided in parallel with the EPCCH for signalling the TFCI indicating the highest allowed TFC to the mobile station. Twelve bits are available in each SACCH/TP burst, and the transmission of a TFCI requires 36 bits when encoded. It would take three SACCH bursts to signal a TFCI value. Two alternative choices are available for signalling these bursts. Firstly, a new value of the TFCI corresponding to the highest allowed TFC may be sent every three SACCH bursts, i.e. every 360 ms. This would provide the shortest update period. Secondly, a new value of the TFCI corresponding to highest allowed TFC may be sent every four SACCH bursts, i.e. every 480 ms. This would involve a longer update period, but would have the advantage of aligning the TFCI transmission with a SACCH block period.
The major problem with either of these arrangements is a relatively lengthy update period, which may delay the provision of a value of TFCI indicating the highest allowed TFC indication to the mobile stations. If channel conditions change quickly, the adaptation period of 360 ms may be too slow, and the performance of the system may be unsatisfactory. Another disadvantage of using SACCH/TP is that Enhanced Power Control (EPC) could not be used with Flexible Layer One. This is presently regarded as a serious drawback, and such arrangements are not presently preferred.
In all of the schemes proposed above, the major problem is the time delay for the base transceiver station (BTS) or other part of the network to perform the measurements and for the network to signal the value of TFCI indicating the highest allowed TFC to the mobile station. The actual delay will depend to a certain extent on the particular scheme used to signal it. The adaptation rate may be too slow, resulting from a long update period. If the channel conditions vary rapidly, the performance of this procedure may not be very good.
An alternative approach is that, instead of signalling the complete, absolute TFCI value each time, an incremental value could be sent. For example, the network may simply signal “UP” or “DOWN” to cause the value of the TFCI corresponding to the highest allowed TFC, represented by pointer 10 in
In an example, the two spare bits in the SACCH header could be used. In such an example a command varying the value of pointer 10 indicating the highest allowed TFC identity by a relative value could be encoded using these two bits as follows:
The SACCH header is transmitted every 480 ms, meaning that the update period is quite long in a multiple of 480 ms, depending on the magnitude of the change.
In the case of the use of a channel in parallel with the SACCH/TP (alternative to the EPCCH channel), an uncoded message of 2 bits could be encoded using the 12 bits available in each SACCH burst, each SACCH burst having 120 ms burst length. If two bits are stolen from each radio packet and an encoded message is made up of 12 bits, then one command can be sent every 120 ms. Furthermore, if six bits are stolen from every radio packet, for example two bits per burst, then a 12 bit message can be sent in two Transmission Time Intervals (TTIs), i.e. every 40 ms.
This operation scheme whereby the relative value is signalled may be the most suitable if inband signalling were chosen. Again, if two bits are stolen from each radio packet and an encoded message is made up of 12 bits, then one command can be sent every 120 ms. But if 6 bits are stolen from every radio packet, then a 12-bit message can be sent in two transmission time intervals (TTIs), i.e. 40 ms.
The advantage of signalling the relative value is that the adaptation to the channel is faster, however the value of TFCI indicating the highest allowed TFC can be only varied in small steps. While a shorter response time is possible, it may take even longer to signal very sharp changes in the highest allowed TFCI.
A summary of possible alternatives for signalling the value of TFCI indicating the highest allowed TFC in GMSK full-rate channels described in this section is given in Table 1. In the table, the update period is given for each alternative scheme.
Some simulation results comparing the performance for the described signalling alternatives are presented below.
Although the preceding description has been expressed in terms of GMSK full-rate channels, similar methods may also be applied to 8-PSK full-rate channels, as will now be described, by way of example. As explained in subclause 7.5 of 3GPP TR 45.902, for 8-PSK modulated basic physical subchannels the number of bits in an encoded TECI sequence is twice as those of a GMSK channel. Therefore, the values given in Table 1 need to be revised as shown in Table 2. However, the SACCH is always transmitted using GMSK; therefore it is only the inband solutions that will be different: the two solutions that use either the SACCH header or a new channel sent in parallel to the SACCH/TP are not affected (it is assumed that in these two cases, the TFCI sequences defined for GMSK full-rate channels would be used).
It has been assumed that in order to signal a relative value inband, a 24-bit sequence is used, which corresponds to twice the length as in the case of GMSK full-rate channels. By doing so, the adaptation rate is the same.
Table 2 shows a summary of possible alternatives for signalling the value of pointer 10 indicating the highest allowed TFC to the mobile station using 8-PSK full-rate channels.
Note that in the case of the inband solutions, in order to keep the signalling rate the same as for GMSK full-rate channels, 4 coded bits are “stolen” either from a radio packet or a burst, rather than 2 symbols (6 bits) as in the case of GMSK full-rate channels.
In the case of GMSK half-rate channels, the coding of the TFCI is obtained by using only the middle segment of the coding defined for GMSK full-rate channels (see subclause 7.5 of 3GPP TR 45.902). This means that the TFCI sequences contain only half the number of bits as in the case of full-rate channels. Given that also half the number of bits is available in a radio packet, it is proposed that, for inband solutions, only 1 bit is “stolen” from each radio packet. By doing so, the signalling rate does not vary with respect to the case of full-rate channels, as shown in Table 3.
When signalling a relative value inband, it has been assumed that a 6-bit sequence is used. This would allow maintaining the same adaptation rate as in the case of full-rate channels; however, the error rate for the commands may be higher than in the case of GMSK full-rate channels, and this aspect may be investigated through simulation.
Table 3 shows a summary of possible alternatives for signalling the value of pointer 10 indicating the highest allowed TFC to the mobile station using GMSK half-rate channels.
In the case of 8-PSK half-rate channels, the coding of the TFCI is obtained by using only the middle segment of the coding defined for 8-PSK full-rate channels (the number of bits is the same as in the case of GMSK full-rate channels). The update times for this case are given in Table 4.
Simulated Results
Possible methods for selecting and signalling the Transport Format Combination (TFC) to be used in the uplink for Flexible Layer One are discussed above. There follow simulation results comparing the alternatives and comparisons of which of the schemes offers optimal performance. Both acknowledged and unacknowledged radio link control (RLC) modes are considered. The simulator used for these simulations is similar to the one used in the investigations described in GP-031389, “Performance of FLO with one transport block per transport channel per TTI”, Siemens, GERAN#15, Florida (USA), 23-27 Jun. 2003, assuming only one transport Block per Transport Channel per Transmission Time Interval (TTI).
A summary of the simulated alternatives for TFC signalling is given in Table 5. In the table, the adaptation period and signalling method is given for each alternative scheme. All TFCs are 8PSK modulated.
Table 6 summarises the simulation parameters used for all the simulated alternatives.
Table 7 indicates the Transport Format Combination Set used. All TFCs are 8PSK modulated.
For each simulation of the TFC update period of the link adaptation algorithm and the averaging period for the channel measurements is summarised in Table 8.
Good throughput performance is achieved with signalling periods of 120 ms or less. The fluctuations in the throughput and service data unit frame erasure rate (SDU FER) for the schemes with update periods greater than 120 ms are believed to be due to poor decisions made by the link adaptation algorithm which results using a non-optimal TFC for the prevailing channel conditions.
As shown in
Based on the results of these simulations, a preferred embodiment of the present invention is one in which the absolute TFCI value is signalled, inband, with an update period of 100 ms.
While the present invention has been described with reference to a limited number of particular embodiments, given by way of examples only, it will be clear to those skilled in the art that other alternative methods may be employed in the performance of the methods of the present invention, as defined in the appended claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| 0302024.5 | Jan 2003 | GB | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/GB04/00369 | 1/29/2004 | WO | 3/3/2006 |
