Communications system

ES 2 219 635 T3

DESCRIPTION

Telecommunications system with multi-frame structure and variable data rate channel.

The present invention relates to radio access bearers which are aligned with both the GSM / EDGE RAN (GERAN) and the UMTS RAN (UTRAN).

Generally speaking, telecommunications services are divided into two categories which are carrier services and teleservices. Bearer services allow a user to access various forms of communication such as an asynchronous circuit-switched data service that interworkes with the public switched telephone network (PSTN) or a synchronous packet-switched data service that interworks with the network. Public Packet Switched Data (PSPDN). On the other hand, teleservices allow a user to access various forms of applications such as voice transmission, short message services and facsimile transmissions. Such bearer services are currently adopted in the Universal Mobile Telecommunications System (UMTS). This UMTS network is made up of four subnets, the access network, the core network, the service mobility control network and the telecommunications management network. Among these, the access network is responsible for the basic transmission and switching functions required to enable a mobile station (MS) to access a fixed network resource through the radio interface (Um interface).

The bearer services (the bearers) that allow a user to access various forms of communication through the UMTS radio access network (RAN) are already well defined.

An alternative to the UTRAN is the GERAN. As GERAN develops, new radio access bearers are defined. As the GERAN will connect to a common core network with the UMTS it is required that the bearers offered by the GERAN are aligned with those of the UTRAN. The following traffic classes must then be able to operate to meet the service requirement. These traffic classes are the types of traffic that will occur across the RAN between the access network and the core network of the mobile phone system. Conversational traffic

Real-time conversation schemes are characterized by the fact that the transfer time must be low due to the conversational nature of the scheme and at the same time because the temporal relationship (variation) between information entities of the data stream must be preserved. in the same way as for real-time data streams. For this reason the limit for an acceptable transfer delay is very strict as failure to achieve a sufficiently low transfer delay will result in an unacceptable lack of quality. For this reason the transfer delay requirement is significantly lower and at the same time more stringent than the round trip delay of the interactive traffic case discussed below.

Influent traffic (streaming)

This one-way scheme is characterized by the fact that temporal relationships (variation) between information entities (i.e. samples, packets) within a flow must be preserved, although it does not present any requirement on low transfer delay. . End-to-end flow delay variation should be limited to preserve the temporal relationship (variation) between information entities in the data stream.

Interactive traffic

This scheme applies when the end user is online requesting data from a remote computer. Interactive traffic is characterized by the end-user request response pattern. In the destination of the message there is an entity that waits for the message (response) in a certain time. For this reason the round trip delay time is one of the key attributes. Another feature is the fact that the content of the packets must be transferred transparently (with a low bit error rate).

Deferred traffic (background)

This scheme applies when the end user sends and receives deferred data files. Examples include deferred delivery of emails, SMS, database download, and receipt of metered records. Deferred traffic is characterized by the fact that the destination is not waiting for the data in a certain time. In this way this scheme is more or less insensitive to delivery time. Another feature is that the contents of the packet must be transferred transparently (with a low bit error rate).

The main differentiating factor between these various classes of traffic is the level of sensitivity to traffic delay. Conversational class traffic is intended for traffic that is delay sensitive while deferred class traffic is the most delay insensitive class of traffic. The conversational and affluent classes are primarily intended to be used to transport traffic streams in real time. Interactive class traffic and lazy traffic are primarily intended to be used by traditional Internet applications such as WWW, email, telnet, FTP, and news. Due to the more relaxed delay requirements in comparison

ES 2 219 635 T3 between the talk and tributary classes both provide better error rates through channel coding and retransmissions. These traffic classes are further detailed in UMTS 23.107.

In view of the common use of the UMTS core network in the communication protocols used to create the GERAN, radio access bearers should also be constructed as in UMTS where combinations of different protocol modes in a single stack they provide a large set of carriers.

Communication protocols are the sets of rules that users adopt when establishing services and transferring data. The protocols allow the establishment and management of connections and are also necessary to enable reliable communications. The functions that communication protocols provide are well described, although their implementation is not. A model that describes the functions provided by communication protocols contains several layers. These are called protocol stacks.

Figure 1 shows a user plane protocol stack 10 suitable for use with GERAN in which each layer includes different modes. The stack includes a physical layer 11 that is analogous to the physical layer of a protocol stack of the UMTS access network, a media access control (MAC) layer 12 that corresponds to the data link layer of a standard UMTS stack, a radio link control (RLC) layer 13 corresponding to the network layer of the UMTS stack, and a packet data convergence protocol (PDCP) layer 14 corresponding to the model application layer UMTS stack.

If the MS is not totally based on the internet protocol (IP) or it is desired to use the GSM circuit mode, an element should deal with the translation of data in circuit mode to / from IP packets / User Datagram Protocol (UDP) / Real Time Protocol (RTP) and the translation of 04.08 signaling to / from certain IP-based signaling (eg H.323). This feature is most likely required only for the conversational and affluent traffic classes. Consider an example where a stream of data is transmitted between the endpoints of a connection in packet data. The data blocks produced by an application can be encapsulated in data packets of certain transmission protocols. The Real Time Protocol (RTP) is an example of a packet data protocol that can be used for applications that do not tolerate delays. The data blocks are encapsulated in RTP protocol packets by placing the data blocks themselves in a payload of the packets and adding appropriate headers to the data blocks. Some protocols may require certain information at the end of the protocol packet as well.

RTP data packets can be transmitted using the User Datagram Protocol (UDP), which can run on the Internet Protocol (IP). UDP and IP add their own headers to data packets. For this reason the data packet delivered to a link layer protocol consists of the original payload and many headers. The link layer protocol can perform header extraction, for example protocol headers typically contain multiple fields whose content does not vary from packet to packet. The result of the header extraction is called the header extraction residue, and it is the information that needs to be transmitted for a certain packet or group of packets to allow the receiving end to rebuild the packet headers. The extraction of the headers can be performed on each data packet in a similar way, or it can be performed, for example, on the first data packet and then the content of the headers of the following data packets is determined using the information from the headers of the first data packet.

For the RTP / UDP / IP protocol combination the header extraction result typically contains at least the sequence number (SN) of the RTP packet, the timestamp (TS) of the RTP packet, and the marker bit (M) of the RTP packet. It may be necessary to transmit only a certain compensation of such data for the update. Information related to the UDP and IP headers can be determined directly after the first UDP / IP packets of the connection have been transmitted to the receiving end. Once the residue from the extraction of the headers and the payload of the data packets have been transmitted through the radio access network, a network element on the other side of the radio access network can reconstruct RTP / UDP / IP packets using the residue from extracting the headers and transmitted payloads. Typically the protocol packets are transmitted without the headers over the radio interface, the network element that reconstructs the headers and the protocol packet can be, for example, either a mobile station or a base station controller (BSC ), depending on the transmission direction. Especially in a receiving mobile station, which typically does not forward the data packets to other network elements, the header reconstruction does not necessarily mean that a data structure corresponding to the header is explicitly constructed. It may be sufficient if the residue from the extraction of the headers and the payload of the data packet is forwarded through the IP / UDP protocol layer to the RTP layer. For example, in the IP / UDP layers, only some counters related to the sequence number of the IP / UDP protocol packets can be incremented.

It would also be advantageous if several radio access bearers were allowed which could be used simultaneously with a single user equipment. This can be used to provide support for multiple Quality of Service (QoS) profiles in parallel. This helps maintain the quality of communication in various traffic conditions.

ES 2 219 635 T3

A number of multiplexing scenarios should also be considered in obtaining radio access bearers for the GERAN. They are set out below.

Operational Scenario 1 (OS1)

Permanent assignment of a channel to a voice call (conversational) without any multiplexing capabilities. Operational Scenario 2 (OS2)

Permanent allocation of a channel to a voice call (class of conversational traffic) and multiplexing of data of the best effort of the same user (class of deferred traffic).

Operational Scenario 3 (OS3)

Permanent assignment of a channel to a voice call (conversational traffic class) and multiplexing of best-effort data from different users (deferred traffic class).

Operational Scenario 4 (OS4)

Assigning a channel to more than one voice user (and / or data users) in a dynamic way.

Several attempts have already been made to provide radio access bearers aligned with both GERAN and UTRAN. However, these systems have experienced a series of drawbacks.

One proposed solution provides a system that does not reuse circuit-switched traffic channels. The differentiating characteristic of a circuit switched system is the exclusive use of a channel with previously set bandwidths which is dedicated to the use of two users for the duration of a call. For example, in the Global System for Mobile communications (GSM) radio access network the two-way circuit-switched channel is reserved for each call. The transmission capacity of the bidirectional channel is the same in both directions, that is, the uplink and the downlink. Since during a voice call the channels are active for only about 40-50% of the time, this represents inefficient channel utilization.

Also in the information transfer no diagonal interleaving has been provided. This reduces the effectiveness of error correction codes and makes data loss more likely.

Furthermore the proposed solutions do not provide a half speed packet switched channel. Packet switching is based on the idea of ​​message switching. A message or data group is made up of a header and an end-of-message part. The message is stored in a buffer at each switch where the header is decoded and the next node in a route is determined. A half-rate packet-switched channel allows each channel to be divided into two sub-channels, thereby providing increased traffic potential. It makes use of so-called half-rate codecs (i.e. a codec that provides quality 8 kb / s trunk circuit-type voice) which helps to improve spectral efficiency or user density for the channel spectrum. assigned.

Similarly, no quarter-speed circuit-switched channel has been provided. This situation has the disadvantage that the advantages of the quarter speed codecs that have been developed cannot be used.

Another drawback of previous systems has been the lack of associated control channel (ACCH) considerations. These control channels carry signaling or timing data and are well known in telecommunications systems. Four categories of control channels are used. These are known as a broadcast control channel (BCCH), common control channel (CCCH), autonomous specialized control channel (STDCCH), and associated control channel (ACCH). These ACCH channels will be described in more detail later.

For this reason it is an objective of the present invention to provide GERAN radio access bearers which at least partially conform to the requirements listed above. Advantageously the present invention has the further objective of at least partially avoiding the drawbacks provided by other prior GERAN radio access bearers.

Preferably the data channel for circuit-switched communications and the data channel for packet-switched communications are assigned the same time slot numbers in each frame. Alternatively, half or a quarter of the number of slots that are assigned to the data channel for packet-switched communications can be assigned to the data channel for circuit-switched communications.

The data channel for circuit switched communications may be a half speed data channel or a quarter speed data channel. The data channel for packet switched communications may be a half rate data channel.

ES 2 219 635 T3

Control data for controlling the data channel for packet-switched communications is preferably carried on the data channel for circuit-switched communications. Said control data may be for the control of the transmission power and / or the handover of the channel. The control data may comprise a fast access control channel and / or a slow access control channel.

The data channel for circuit switched communications may be a conversational channel. The data channel for circuit-switched communications may be a deferred channel. The data channel for packet-switched communications may be assigned time slots during periods when the data channel for circuit-switched communications is relatively idle, for example, during pauses in voice data being carried by middle of the data channel for circuit-switched communications.

In the above aspects of the invention a data channel for circuit-switched communications may carry data in the form of a circuit-switched connection or otherwise. Preferably the circuit-switched channel is capable of operating through a core circuit-switched network of the communication system.

According to a third aspect of the invention there is provided a communication system comprising a first station capable of communicating with a second station through a wireless channel, the data being transported through the wireless channel in superframes, each superframe comprising a plurality of frames and each frame comprising a plurality of time slots;

presenting the system:

a first mode of operation in which a full rate data channel for packet switched communications is defined by assigning to that data channel corresponding time slots in each frame;

a second mode of operation in which two full rate data channels are defined for packet switched communications by assigning each of those data channels the same number of corresponding time slots of frames in each super frame.

The channel or each channel of full or half rate data for packet switched communications may be a tributary, interactive or delayed channel. The channel or each channel of full, half or quarter rate data for circuit switched communications may be a conversational channel.

The system can be operated according to the GSM specification or a derivative thereof, such as the GERAN system.

Preferably the wireless channel carries data by means of 8-state phase shift keying (8PSK) modulation.

Embodiments of the present invention provide several advantages over previous solutions. First of all radio access bearers are compatible with and therefore meet the design requirements of version 2000. This represents the next generation of telecommunication networks.

Second, reuse of the already specified channel coding of adaptive multi-rate (AMR) voice traffic channels for conversational traffic classes and circuit-switched data traffic channels for tributary traffic classes is envisaged.

Third, the embodiments of the present invention allow for circuit-switched and packet-switched channel multiplexing within the same time interval. This enables the conversational and interactive traffic classes to coexist within the same time interval.

Fourth, the embodiments envisage a quarter-speed circuit-switched traffic channel thus taking advantage of the quarter-speed codecs that are available.

Fifth, the embodiments of the invention allow the reuse of the already specified associated control channel of the circuit mode (in particular the slow associated control channels (SACCH) and the fast associated control channels (FACCH)) for classes of conversational traffic. and affluent.

Furthermore, the embodiments foresee that when packet data from the same user is multiplexed within the silence periods of a voice traffic channel (conversational traffic class), the packet data will also use the SACCH and FACCH channels of the voice traffic channel.

Still other embodiments provide half rate packet switched traffic channels to increase multiplexing capabilities.

ES 2 219 635 T3

Embodiments of the present invention will now be described with reference to the following drawings in which:

Figure 1 shows a user plane protocol stack suitable for use in GERAN; Figure 2 shows a full rate traffic channel; Figure 3 shows a half speed traffic channel;

Figure 4 shows a quarter speed traffic channel;

Figure 5 illustrates FACCH mapping on full rate channels;

Figure 6 shows the FACCH mapping on half-rate channels;

Figure 7 shows FACCH mapping on quarter speed channels;

Figure 8 shows a full rate packet channel;

Figure 9 shows a half rate packet channel;

Figure 10 illustrates conversational radio access bearers;

Figure 11 illustrates tributary radio access bearers;

Figure 12 illustrates interactive radio access bearers; and Figure 13 illustrates deferred radio access bearers.

In the drawings, like reference numerals refer to the same parts.

The protocols used to create the radio access bearers are built in UMTS in which combinations of different modes of protocols in a single stack provide a large set of bearers. The protocol stack to be used is represented graphically in Figure 1, including each layer in different ways. The different modes for each layer are identified below.

Packet Data Convergence Protocol (PDCP)

Transparent with RTP / UDP / IP header removal. Bearer services can be transparent or non-transparent. Transparent services provide protection against errors only through receive error correction (FEC). On the other hand, non-transparent services have the additional protection of the automatic repeat request (ARQ). This is provided in the radio link protocol which features improved data integrity.

Non-transparent with header adaptation (header extraction or header compression).

Non-transparent without adaptation of headings.

Radio Link Control (RLC)

Transparent

No confirmation of receipt

With confirmation of receipt Media access control (MAC)

Specialized: no user identification is included allowing only one user per channel. However whenever this continuous transmission (DTX) occurs, data packets can be transmitted from the same user. The function of the DTX is to suspend the radio transmission during silent parts on a voice channel. Typically this is used to help prevent interference and increase system capacity. By transmitting data packets during silent parts the capacity of the system can be further increased.

Shared: the same channel can be shared between several users.

ES 2 219 635 T3

Physical (PHYS)

Modulation: A modulation process is used to convert the voice or data encoded in the channels into a type suitable for transmission over the radio channel. Indeed, modulation enables the transmission of binary information on analog carriers. During modulation one bit or group of bits results in rapid state changes such as amplitude or frequency changes. Gaussian Least Shift Keying (GMSK) and Eight-State Phase Shift Keying (8PSK) are currently defined for use with GERAN. Voice transmission uses only GMSK while data can be transported using 8PSK or GMSK modulation. In phase shift modulation the phase of a signal is shifted differently from the previous phase (eg, plus 90% for zero and plus 270% for one).

Channel coding: since voice and data signals encoded with electromagnetic interference transmitted through the radio interface must be protected against errors. Convolutional encoding and block interleaving are used to achieve this protection. In particular, within the GSM specification, there are two different error protection mechanisms which perform convolutional coding. Unequal Error Protection (UEP) which treats the bits of a signal with a different channel encoding depending on the bit class (class 1a bits are the most sensitive to bit errors, class 1b bits are moderately sensitive while Class II bits are the least sensitive to bit errors). Equal Error Protection (EEP) uses the same channel encoding for all data information.

Channel speed - A traffic channel is used to carry voice and data traffic. Traffic channels are defined using a 26-frame multiframe as will be described in more detail hereinafter. Of the 26 frames 24 are used for traffic. These are the full speed traffic channels. Some half speed and quarter speed channels are also provided. It will be understood that the present invention is not limited to frames and multiframes of this configuration.

Interleaving: As mentioned above, interleaving is used to protect data from errors that occur during transmission. After encoding, interleaving steps are carried out to interleave the various bits of the signal with the index coding to form an interleaved sequence. If part of that sequence fails, the rest can be used to reconstruct the correct data. Interleaving can be diagonal (diag) or rectangular (rect) and different depths of interleaving can be used (19, 8, 4, 2). The greater the interleaving depth, the better the link level performance, although the delay will be greater.

Radio access bearers according to the present invention are selected from combinations of the different layers that are offered.

The mapping of radio access bearers on the physical layer can use two types of traffic channels as described above. These are Packet Channels (PCH) and Circuit Switched Channels (TCH). User data is not only the information that must be transported through these channels through the air interface. Signaling messages must also be carried. These allow the network and the MS to discuss the management of various issues such as resources and handovers. When traffic is in progress, this signaling is done through the associated control channel (ACCH). However due to different requirements the way in which ACCH channels are implemented is different for packet or circuit switched traffic channels. Several ACCH channels are well defined for circuit and packet switched channels and some of these are identified and described below. Furthermore, ACCH channels for GERAN radio access bearers implemented according to the present invention are described.

ACCH channels are bidirectional channels. On the downlink they carry control commands from the base station to the mobile station (MS) to control their transmitted power level. On the uplink they carry the state of the MS to the base station. SACCH is used in layer signaling at least for measurement results during transmission from MS to network. SACCH has the particularity that continuous transmission must occur in both directions. To this end in the direction of the MS to the network messages of measurement results are sent on every possible occasion when nothing else is to be sent. Similarly system information type 5, 6 and optionally messages 5 bis and 5 ter as known in the art are sent in the network direction to the MS in UI frames when nothing else is to be sent. SACCH is used for non-urgent procedures, mainly for the transmission of radiocommunication measurement data required for handover decisions.

In each SACCH downlink block there is commanded MS power level and commanded timing advance information. In each SACCH uplink block there is real MS power level and real time lead information.

In addition, SACCH carries messages detailed in Annex A. Each SACCH block contains 184 bits of information that are 456 bits coded and interleaved for four bursts. One SACCH cycle is 480 ms. In others

ES 2 219 635 T3 words time advance, power level and measurement reports can be updated every 280 ms. It will be understood that the present invention is not limited to blocks and bits of this configuration.

The FACCH (also known as the main specialized control channel (DCCH)) facilitates urgent action such as handover orders and channel reassignment in intra-cell handovers. It is transmitted by acquiring half or all the information bits of the bursts of the traffic channel (TCH) to which it is associated.

There are four alternative varieties of bursts used for transmission in GSM. These are the normal burst, the F burst, the S burst, and the access burst. Of these, the normal burst is used to carry data and most of the signaling. It has a total length of 156.25 bits consisting of two 57-bit chunks of information, a 26-bit training sequence used to synchronize the receiver with incoming information and to avoid the negative effects of multipath propagation, 1 theft bit for each information block (which indicates to the receiver if the information carried by a burst corresponds to traffic or signaling data), 3 tail bits at each end (used to cover the up and down ramp periods of a mobile's power) and an 8.25-bit guard sequence (used to avoid a possible overlap of two mobiles during the ramp time ). The FACCH is used for various purposes such as call setup progress, handover, subscriber authentication, DTMF, notification (for VGCS and VBS-instead of NCH), and paging (instead of PCH).

The FACCH can carry messages that are described in Annex A. Each FACCH block contains 182 bits of information (or data bursts), these are 256 bits encoded as SACCH, depending on the interleaving of its associated channel (full speed or half speed) .

The Enhanced Fast Associated Control Channel (E-FACCH) is a Fast Associated Control Channel introduced for the ECSD. Each E-FACCH block contains the same information as the FACCH (182 bits) and uses GMSK modulation. However, the E-FACCH mapping is established in full consecutive bursts instead of eight half bursts for full rate FACCH.

Enhanced in-band associated control channel (E-IACCH) is the in-band E-TCH / F associated control channel introduced for fast power control (FPC) in the ECSD. The BSS indicates to the MS through the SACCH channel the use of the FPC. Power control information is sent every FPC report period of length 2 TDMA frames (20 ms). The three bits of information are encoded into 22 bits which are mapped to the steal symbols of four consecutive normal bursts.

Even if fast power control is activated, normal power control (via SACCH) is always in operation. However, in this case the MS ignores the power level commands from the SACCH.

The ACCHs mentioned above are associated with circuit-switched traffic channels. The next two ACCHs are associated with packet traffic channels.

The Packet Associated Control Channel (PACCH) carries signaling information related to a given MS. The signaling information includes for example acknowledgments and power control information. The PACCH also carries resource allocation and reallocation messages, which comprise the allocation of a capacity for PDTCH channels and for other PACCH phenomena. The PACCH shares resources with PDTCH channels that are currently assigned to an MS. Additionally you can search for an MS that is currently involved in transferring packets for circuit switched services on the PACCH. The messages that can be sent in a PACCH are listed in Annex A.

The PACCH is bi-directional. Each block contains 182 bits of information that are 256 bits encoded and interleaved for four bursts (same encoding as SACCH). However, the PACCH does not have continuous transmission such as the SACCH.

Due to this continuous transmission, a continuous update timing advance mechanism has been defined in GPRS. The time advance can be updated through its own channel. This is called the packet timing advance control channel (PTCCH). An MS in packet transfer mode will be regularly requested to send random access bursts to the uplink to allow timing advance estimation. The PTCCH is then used on the downlink to transmit time-ahead information updates to various MSs. The following Table 1 lists the various control channels.

ES 2 219 635 T3

TABLE 1 ACCH functions

<td colspan="2"></td><td colspan="2">Circuit channels</td><td colspan="3">packages</td>

<td rowspan="2">Measurement reports</td><td>Channel</td><td colspan="2">SACCH</td><td colspan="3">PACCH</td>

<td>Upgrade</td><td colspan="2">480 ms</td><td colspan="3">only if the MS must do it (that is, NETWORK_CONTROL_ORDER = NC1 or NC2) controlled in speed by the NC_REPORTING_PERIOD_T (min = 480 ms / max = 6144 ms)</td>

<td rowspan="2">Time advance</td><td>Channel</td><td colspan="2">SACCH</td><td>PTCCH</td><td colspan="2">PACCH</td>

<td>Upgrade</td><td colspan="2">480 ms</td><td>1920 ms</td><td colspan="2">20 ms-free</td>

<td rowspan="2">Power control</td><td>Channel</td><td>SACCH</td><td>E-IACCH</td><td colspan="2">PDTCH (RLC / MAC Header)</td><td>PACCH</td>

<td>Upgrade</td><td>480 ms</td><td>20 ms</td><td colspan="2">20 ms4free</td><td>20 ms4free</td>

<td colspan="2">T raspaso</td><td colspan="2">FACCH</td><td colspan="3">PACCH not handover as such but new cell selection can be controlled by network or by MS</td>

The table shows the associated control channels and update times for the various control procedures for both circuit-switched and packet-switched traffic channels.

In a somewhat similar way to the existing examples mentioned above, GERAN radio access bearers make use of two different types of traffic channels. These are the circuit-switched and packet-switched channels.

Circuit-switched channels can be used for conversational and affluent traffic classes where a constant stream of real-time data is required. Obviously there is some difference between the delay requirements of these two classes as the type of inflowing traffic presents more relaxed requirements. From a physical layer point of view it means that the type of inflowing traffic allows the use of a longer interleaving.

The way in which the SACCH is mapped onto a physical channel does not depend on the modulation used for data transfer, nor does it depend on the class of traffic. As mentioned above in relation to existing traffic channels (TCH) the SACCH mapping will be established for four GMSK bursts.

The proposed SACCH mapping is graphically represented in Figure 2 which follows well known mapping procedures. The data burst modulation can be either GMSK or 8PSK.

Figure 2 depicts a multiframe (or superframe) 20 defining the full rate traffic channel (TCH / F). Each multiframe comprises a group of 26 TDMA 210-25 frames. As the radio spectrum is a limited resource the bandwidth is divided through frequency division multiple access (FDMA) and time division multiple access (TDMA) as is well known in the art. In particular the FDMA involves the division by division of the bandwidth of 25 Mhz in 124 carrier frequencies separated from each other by 200 khz. Each of these is then divided into time through a TDMA scheme. The basic unit of time in the TDMA scheme is indicated as a burst period and lasts approximately 0.577 ms. Each TDMA frame 210-25 is divided into eight of these burst periods 22. Thus each TDMA frame 210-25 consists of eight burst periods 22 which form a basic unit for logical channels. A physical channel is one burst period 22 per TDMA frame 21. Channels are defined by the number and position of that corresponding burst period. Throughout the description that follows, the term "multiframe" will be used and should be understood as a superframe, that is, a frame made up of multiple TDMA frames. Similarly, the term "burst period" will be understood to represent a time interval in the TDMA frame.

Each of the eight burst periods 22 that make up a TDMA frame comprises a normal 156.25-bit burst that includes two bursts of data as previously described herein.

Of the 26 frames 21, 24 are used for traffic and can transmit data, one, SACCH frame 23 is used for SACCH. The final frame 24 is not used and is inactive. In speech applications, digitized speech is typically compressed using a certain speech coding method before it is transmitted over the radio interface. The amount of encoded speech depends on the quality of the target speech and the efficiency of the speech encoding method. Typically coded speech is transmitted in speech frames, and typically one speech frame roughly corresponds to the length of four TDMA frames. In a full rate channel 6 speech frames (120 ms) correspond to the duration of 26 TDMA frames (24 for the

ES 2 219 635 T3 voice + 1 for SACCH + 1 for inactive). The speech frames are encoded on the channels with a suitable channel encoding method; the choice of channel coding method is usually influenced by the data transmission speed of the communication channel reserved for the call. Typically for the full rate channel the number of bits in a voice frame encoded in the channels is equal to or less than the number of bits carried by four radio bursts. The interleaving depth, which means about how many radio bursts a certain encoded data frame is mapped to, also typically depends on the data transmission rate of the communication channel.

Known half rate traffic channels (TCH / H) are graphed in Figure 3 which also follows the existing SACCH mapping. Two subchannels 30, 31 are shown each provided through a respective multiframe 32, 33. Each of these multiframes (or superframes) includes 26 TDMA frames although the subchannel in each of them is provided over a period Burst (T) in every two TDMA frames

twenty-one. In this case the SACCH for subchannel 31 makes use of the 25^to plot 21₂₅ otherwise it would be inactive.

A quarter rate (TCH / Q) traffic channel for use with circuit switched traffic channels is graphically depicted in Figure 4. Four subchannels 40, 41, 42, 43 are provided each of which is formed by means of a burst period T approximately every four TDMA frames. To provide a SACCH for each of the subchannels, a burst period is reserved once in every two multiframes. Because of this, the conditions required to transmit a satisfactory data rate over the air interface is preferably used in indoor and micro-cell environments. It will of course be understood that the present invention is not limited to such environments. Naturally in such an environment the mobility of the user is reduced and therefore the speed of the SACCH can be decreased without any detrimental effect on performance.

As seen in Figure 4 the SACCH for subchannel zero 40 is provided in TDMA frame 2112 of multiframe 440. The next multiframe 44 of TDMA frames for that channel does not include a SACCH burst period. Similarly for subchannel 1.41 which is formed by the multiframes 450 and 451 that provide TDMA frames 0 to 51 the SACCH period is in TDMA frame 2138. For subchannel 2, 42 the SACCH period occurs in the TDMA frame 2125 of multiframe 460. In multiframe 461 no SACCH period is required. In subchannel 3, 43 the SACCH period occurs in TDMA frame 2151 in multiframe 471. In multiframe 470 no SACCH period is provided.

Obtaining these four subchannels does not require the allocation of any additional TDMA frames other than the previously existing SACCH and other idle channels.

As the FACCH is embedded in delay sensitive mechanisms such as allocation, notification, paging, handover, or even the transmission of ETMF signals, the delay requirements cannot be relaxed. For example, even if the probability of handover is quite low (for example, in a good environment and with a user with reduced mobility), this does not mean that the FACCH delays can be increased. In fact, other mechanisms that use the FACCH must continue to be carried out and longer delays could cause problems in these situations. In this way the FACCH is based on an existing theft mechanism in which the acquisition can take place at two different levels. These are the frame level where each FACCH block replaces data frame (s) and the burst level where each FACCH block replaces four consecutive data bursts with four GMSK bursts (ECSD only).

The way in which the traffic is carried depends on the interleaving used. In ECSD where the relaxed delay requirements allow for long interleaving, the theft mechanism occurs at a burst level (four consecutive bursts stolen). In this case there is only a slight influence on each data frame while the fast adjective of the FACCH still makes sense. When voice is transported the theft mechanism occurs at a frame level. The data frame (s) are then simply lost.

The following Table 2 establishes a brief comparison between the two possibilities of the theft mechanism.

TABLE 2

Theft mechanisms

<td></td><td>Burst robbery</td><td>Plot theft (s)</td>

<td>Interleaved in the FACCH</td><td>Fixed-4 bursts</td><td>the same as the TCH one</td>

<td>Modulation in the FACCH</td><td>GMSK</td><td>the same as the TCH one</td>

<td>Effect on data</td><td>crop / reduced quality</td><td>cutout</td>

<td>FACCH delay</td><td>Permanent</td><td>depends on TCH Interleaving</td>

The method of obtaining the FACCH depends on the type of channel from which the theft mechanism operates. These will be either data channels or voice channels.

ES 2 219 635 T3

A full rate data channel could use either 8PSK or GMSK modulation. For both, existing solutions are included in the GSM specifications and for this reason they are reused for GERAN. Note that when using 8PSK modulation, the question arises as to which modulation to use to transmit FACCH. ECSD studies have shown that considering the performance results and robustness of the FACCH identification the preferred solution is to map the FACCH over four full consecutive GMSK bursts.

A half-rate data channel can use only GMKS modulation to reuse existing solutions included in GSM specifications. 8PSK half speed data channels could be used but are not preferable. On the other hand, a full-speed voice channel can use either 8PSK or GMSK modulation. For GMSK modulation the FACCH mapping follows the existing solutions described in the GSM (frame stealing) specifications. For 8PSK modulation the theft mechanism can take place at two different levels (burst or frame) as shown in Figure 5. Table 3 makes a comparison of both mechanisms.

TABLE 3

FACCH Theft Mechanism Comparison for FR 8PSK Channels

<td>FACCH</td><td>Burst robbery</td><td>Plot theft</td>

<td>Modulation</td><td>GMSK</td><td>8PSK</td>

<td>Code speed</td><td> 0,4</td><td> 0,14</td>

<td>Interleaving depth</td><td> 4</td><td> 8</td>

<td>Effect on voice</td><td>40 ms of reduced quality (enough channel coding to recover data)</td><td>20 ms clipping</td>

Figure 5 shows a portion of multiframe 50 for a full rate speech channel consisting of frames

Consecutive TDMA 510-17. Each one made up of eight burst periods 52 or time intervals. Each burst period consists of 156.25 bits as described above. These include two 57-bit information chunks also known as two 53-bit frames or 57-bit data bursts. Thus each time slot 52 includes two 57-bit data bursts 53 each positioned in a corresponding part of the time slot 52. In other words, each 156.25-bit burst period includes two 57-bit frames 53. When an urgent action requires a fast channel handover or reassignment the FACCH can either steal four consecutive burst periods to provide the data to control that urgent action or it can steal eight consecutive burst period bit frames. In the case of stealing bit frames, a diagonal interleaving policy is adopted to maintain the integrity of the information. In this way, by stealing bit frames (or data bursts) instead of full burst periods (or time intervals) the effect of the audible voice transferred on the open channel can be minimized as can be seen more clearly in Table 3.

Figure 6 illustrates a steal mechanism to be used with a half speed voice channel. Either 8PSK or GMSK modulation techniques are available for this channel. For GMSK modulation the FACCH mapping can follow existing mapping solutions as described in GSM specifications as is well known.

For 8PSK modulation the stealing mechanism necessary to provide the FACCH can take place at two different levels (burst or bit frame) as shown in Figure 6. Figure 6 shows a part of the multiframe 60 consisting of a data stream of consecutive TDMA frames 610-17 each of which includes eight burst periods 62 (or time slots). For a half rate channel the channel will be divided into subchannels each consisting of burst periods in the same time interval in approximately every two frames

TDMA. In Figure 6 the channel transfers voice using burst periods 610-8. When an urgent action occurs requiring fast channel handover or reassignment the FACCH can optionally steal four consecutive 630-0 bursts in consecutive frames or non-consecutive frames. Consecutive bit frame stealing uses both frames from each of two consecutive burst periods. In the case of frame theft, a diagonal interleaving policy is adopted whenever possible. Table 4 shows the effects on the voice of the three independent theft mechanisms and also shows their characteristics.

ES 2 219 635 T3

TABLE 4

FACCH Theft Mechanism Comparison for HR 8PSK Channels

<td>FACCH</td><td>Burst robbery</td><td>Frame theft</td><td>Theft of non-consecutive frames</td>

<td>modulation</td><td>GMSK</td><td>8PSK</td><td>8PSK</td>

<td>code speed</td><td> 0,4</td><td> 0,14</td><td> 0,14</td>

<td>interleaving depth</td><td> 4</td><td> 6</td><td> 8</td>

<td>effect on voice</td><td>60 ms clipping (not enough channel encoding to retrieve data)</td><td>clipping of 40 ms</td><td>20 ms cut + 20 ms cut</td>

<td>Others</td><td></td><td></td><td>delay + 20 ms</td>

Figure 7 illustrates the stealing mechanism for a quarter speed voice channel. The preferable modulation that accommodates two quarter speed channels is 8PSK modulation. The theft mechanism can take place at two different levels (burst or frame) as shown in Figure 6. To increase the depth of the interleaving (hence the link level performance) a solution to consider is to steal two frames not consecutive. Table 5 makes a comparison of the three mechanisms.

TABLE 5

FACCH Theft Mechanism Comparison for the 8PSK QR Channel

<td>FACCH</td><td>Burst robbery</td><td>Frame theft</td><td>Theft of non-consecutive frames</td>

<td>modulation</td><td>GMSK</td><td>8PSK</td><td>8PSK</td>

<td>code speed</td><td> 0,4</td><td> 0,14</td><td> 0,14</td>

<td>interleaving depth</td><td> 4</td><td> 5</td><td> 8</td>

<td>effect on voice</td><td>100 ms clipping (not enough channel encoding to retrieve data)</td><td>80 ms clipping</td><td>20 ms cut + 20 ms cut + 20 ms cut + 20 ms cut</td>

<td>others</td><td></td><td></td><td>delay + 60 ms</td>

Figure 7 shows a portion of multiframe 70 that is part of an ongoing information data stream carrying voice traffic. The multiframe consists of a data stream of consecutive TDMA frames 710-17. For a quarter rate channel the channel will be divided into subchannels each consisting of burst periods in the same time interval in approximately every four TDMA frames (in fact in TDMA frames 710,4,8,13, 17). When urgent action requires fast channel handover or reassignment the FACCH can optionally steal four consecutive bursts from the subchannel (i.e. the burst periods of the TDMA 710,4,8,13 frame) or consecutive frames of the consecutive burst (i.e. the second frame of the burst period in the TDMA 710 frame, both frames of the burst periods in the TDMA 714,8,13 frame, and the first frame of the burst period in TDMA frame 7117 or non-consecutive frames of consecutive burst periods (which would require more TDMA frames than shown in Figure 7). Table 5 shows effects and characteristics provided by the FACCH steal mechanism for the quarter speed voice channel.

ACCHs associated with packet traffic channels (PACCHs) are different from ACCHs associated with circuit-switched traffic channels. The PACCH requires an explicit allocation of resources while the SACCH is implicitly allocated a time slot every 120 ms (26 TDMA frames). Furthermore, no FACCH approach is necessary as each individual packet can carry either user data or signaling, the difference being indicated through the RLC / MAC headers.

ES 2 219 635 T3

For lazy and interactive traffic classes where a constant stream of real-time data is not required, PACCH blocks can be inserted anywhere. However, when it comes to the conversational and affluent traffic classes a constant flow of data is required. Unfortunately, due to the 52 multiframe structure, matching such traffic will not provide any free blocks for PACCH purposes. As an example, consider a full rate voice packet traffic channel. On the one hand, every 52 TDMA frames 12 blocks are available. On the other hand, every 52 TDMA frames (220 ms) it is necessary to transmit 12 voice frames (20 ms). For this reason each block will carry a voice frame. Consequently there is no block available for the ACCH. The same is true when two half-rate packet voice users are multiplexed on the same packet traffic channel.

However, the timing advance and power control mechanisms do not use the PACCH. Furthermore, since the re-selection of cells can be controlled by the MS, it is not always necessary to transmit measurement reports on the uplink. For this reason, an option is a mechanism whereby an MS sends a list of desired cell candidates only when a handover is required. Consequently in packet mode a PACCH rate as high as one in every 280 ms may not be necessary. Thus for the conversational and affluent traffic classes the PACCH should be able to steal a voice block when needed. To reduce the effects on the end user perceived quality, the PCU could try to fill the quiet periods with PACCH blocks.

However, it is impractical to always have to steal voice packets to transmit control information. For this reason, for the conversational and affluent traffic classes, the circuit-switched approach should be followed as described below.

Figure 8 shows a full rate packet channel (PCH / F) 80 which consists of two multiframes 810.1. Each multiframe includes 26 TDMA 820-25 and 8226-51 frames. Each of the TDMA frames includes eight burst periods which are used to carry data (D). A data channel is provided for a corresponding burst period in each of the TDMA frames. In each multiframe 22, TDMA frames are used to transfer D data by packet switching. A TDMA frame is used as the packet switched traffic control channel (PTCCH) while the remaining burst period is left idle.

Figure 9 illustrates a half rate packet channel (PCH / H). Two subchannels 90, 91 are shown each of which is provided through a pair 920.1 and 930.1 of multiframes. Subchannel 90 is made up of D burst periods in approximately every two TDMA 920-51 frames. Similarly subchannel 91 is formed through corresponding D burst periods in approximately every two TDMA 950-51 frames. The two subchannels are constructed so that the burst periods in each of them are offset from each other. Thus TDMA frame 920 is used for subchannel 90, TDMA frame 951 is used for subchannel 91, TDMA frame 922 is used for subchannel 90, and TDMA frame 953 is used for subchannel 91 and so on.

PTCCH is provided for subchannel 90 in TDMA frames 9212 and 9238. PTCCH is provided for subchannel 91 in TDMA frames 9525 and 9551. Those skilled in the art will understand that although for illustrative purposes subchannels 90 and 91 are shown as four independent multiframes 920.1 and 930.1, they actually represent only two consecutive interconnected multiframes.

The use of such a half rate packet channel (PCH / H) allows multiplexing in the same time slot with a half rate circuit switched channel (TCH / H).

Another way to consider a half rate packet channel would be to allocate one for every two blocks (for bursts) within a PCH / F. However, from the physical layer point of view it would look like a PCH / F and for this reason it could not be multiplexed with a TCH / H. The packets are matched following a granularity of four consecutive bursts. In other words, the packets can be either four or eight bursts in length.

With the full rate, half rate and quarter rate channels mentioned above the following are the possible ways in which the channels can be combined into basic physical channels. The numbers in parentheses after the channel designations indicate subchannel numbers.

i) TCH / F ii) PCH / F iii) TCH / H (0) + TCH / H (1) iv) TCH / H (0) + PCH / H (1)

v) PCH / H (0) + TCH / H (1) vi) PCH / H (0) + PCH / H (1)

ES 2 219 635 T3 vii) TCH / Q (0) + TCH / Q (1) + TCH / Q (2) + TCH / Q (3) viii) TCH / Q (0) + TCH / Q (1) + TCH / H (1) ix) TCH / H (0) + TCH / Q (2) + TCH / Q (3)

x) TCH / Q (0) + TCH / Q (1) + PCH / H (1) xi) PCH / H (0) + TCH / Q (2) + TCH / Q (3)

Figure 10 shows the various modes of a user plane protocol stack suitable for conversational traffic and used with GERAN are configured. The protocol stack 100 includes a packet data convergence protocol (PDCP) layer that corresponds to the application layer of the well-known UMTS stack model and contains three modes 102, 103, and 104 that are non-transparent with elimination. of headers, non-transparent with header adaptation and frame structuring and non-transparent with frame structuring respectively. Transparent modes provide error protection only through Receive Error Correction (FEC). On the other hand, non-transparent modes provide additional protection through ACK (acknowledgment mode). The RTP / UDP / IP header can be removed or adapted.

The protocol stack 100 also includes a radio link control (RLC) layer 105 that corresponds to the network layer of the UMTS stack and includes modes 106, 107 and 108 that are transparent with LA encryption, without confirmation of reception with segmentation, link adaptation (LA) and encryption and without confirmation of reception with segmentation, link adaptation (LA), reception error correction (FEC) and encryption respectively.

The protocol stack also includes a media access control (MAC) layer 109 which includes two modes 110 and 111 which are respectively for dedicated and shared channels. For specialized channels no user ID is included allowing only one user per channel although when DTX occurs data packets of the same user can be transmitted. In shared mode the same channel can be shared between several users.

The protocol stack also includes a physical layer (PHYS) 112 that includes two modes 113 and 114 which are respectively for circuit-switched (TCH) and packet-switched (PCH) channels. The physical layer allows GMSK or 8PSK modulation to convert channel-encoded voice or data into a type suitable for transmission over the radio channel. Various channel coding strategies can also be implemented to protect data integrity such as UEP and EEP. Rectangular and diagonal interleaving can also be entered at a depth of 2, 4, 8, or 19 to promote data integrity.

(Table goes to next page)

ES 2 219 635 T3

TABLE 7

Conversational radio access bearers

<img file="ES2219635T3_D0001.tif" />

The first radio access bearer A can operate with operational scenario (OS) 1 which is the permanent assignment of a channel to a voice call (conversational traffic class) without multiplexing capabilities. This provides an optimized adaptive multi-write (AMR) voice reusing the data link layer of the GSMCS mode. The mapping follows Figures 2, 3 or 4 depending on the channel rate, ie full rate TCH / F, half rate TCH / H or quarter rate TCH / Q. Various coding strategies such as UEP, TCH / AFS, E-TCH / AFS, E-TCH / AHS and ETCH / AQS can also be provided. This radio access bearer uses the FACCH and SACCH signaling mapping as previously described herein.

ES 2 219 635 T3

The second radio access bearer B in Table 1 can work with OS1 and also with OS2 which is the permanent assignment of a channel to a voice call (conversational traffic class) and the multiplexing of the best effort data from the same user (deferred traffic class). This bearer B is provided using transparent mode 102 at PDCP layer 101 with header stripping, transparent mode 106 at RLC layer 105 with link adaptation (LA), and specialized mode 110 of encryption at MAC layer 109 and the circuit-switched mode 113 at physical layer 112. The bearer provides an optimized AMR voice. The coding and signaling are equivalent to bearer A although the protocol stack is different, allowing operation with OS2 thanks to the MAC layer. The mapping follows Figures 2, 3 or 4 depending on the channel speed. It is possible to accommodate best effort data packets from the same user in quiet periods.

Similarly the third radio access bearer C of Table 6 can work with OS1 and OS2. This bearer is provided using the non-transparent mode 103 in the PDCP layer 101 with header extraction as adaptation and including frame structuring which includes segmenting and adding a header. Transparent mode 106 at RLC layer 105 with LA and encryption and specialized mode 110 at MAC layer 109 are also used. Circuit switched mode 113 is used in the physical layer with either full, half or quarter speed (TCH (F / H / Q)) depending on the required channel speed. The bearer provides an optimized AMR voice with header extraction. In addition to the SACCH and FACCH control channels, the carrier uses an embedded associated control channel (MACH) as described in Finnish patent application number 20000415 registered on 02/23/2000, which is incorporated herein by reference. . The mapping follows Figures 2, 3 or 4 depending on the channel speed. It is possible to accommodate best effort data packets from the same user in quiet periods.

The fourth radio access bearer D of Table 6 can work with OS3 which is the permanent assignment of a channel to a voice call (conversational traffic class) and the multiplexing of the best effort data from different users. It can also work with OS4, which is the assignment of a channel to more than one voice user (and / or data users) in a dynamic way. The bearer is provided by non-transparent mode 103 with header extraction and frame structuring from the PDCP layer 101. The non-acknowledgment mode 107 of the RLC layer 105 is also used which provides segmentation, LA and encryption. The shared mode 111 of the MAC layer 109 is used as well as the packet-switched mode 114 of the physical layer 112. Configuring the protocol stack in this way produces a generic conversational radio access bearer D. The mapping follows the scheme shown in Figures 8 and 9 depending on the required channel speed. To take advantage of the longer interleaving, two voice frames are encapsulated in one radio block.

Figure 11 shows the protocol stack 100 for tributary radio access bearers. The protocol stack includes the same modes and layers as those in Figure 10, although the routing and mode selection is different. Blocks shown through a dashed line are not used. The data link layer 115 is taken from the GSMCS mode and thus allows the use of existing circuit-switched data channels. Table 7 details the paths through the protocol stack as indicated by the arrows in Figure 11. The operational scenarios are not applicable in the context of tributary radio access bearers.

(Table goes to next page)

ES 2 219 635 T3

TABLE 7

Tributary radio access bearers

<td rowspan="2">PDCP</td><td rowspan="2">RLC</td><td rowspan="2">MAC</td><td colspan="3">PHY</td><td rowspan="2">Coding</td><td rowspan="2">Signage map</td><td rowspan="2">YOU</td><td rowspan="2"></td>

<td>Channel</td><td>Intercal.</td><td>Mod.</td>

<td rowspan="2">^X</td><td rowspan="2">data link (from the GSM CS)</td><td rowspan="2">data link (from the GSM CS)</td><td rowspan="2">TCHZF</td><td rowspan="2">19 diag</td><td>GMSK</td><td>EEP TCH / F14.4 TCH / F9.6</td><td>FACCH + SACCH</td><td>Not inte- hangover</td><td rowspan="2">TO</td>

<td>8PSK</td><td>EEP E-TCH / F28.8 E-TCH / F32.0 E-TCH / F43.2</td><td>FACCH + SACCH + E-IACCH / F</td><td>Not inte- hangover</td>

<td rowspan="2">Transparent No heading</td><td rowspan="2">Transparent THE Encryption</td><td rowspan="2">Specialized</td><td rowspan="2">TCH / F</td><td rowspan="2">19 diag</td><td>GMSK</td><td>EEP TCH / F14.4 TCH / F9.6</td><td>FACCH + SACCH</td><td>Not inte- hangover</td><td rowspan="2">B</td>

<td>8PSK</td><td>EEP E-TCH / F28.8 E-TCH / F32.0 E-TCH / F43.2</td><td>FACCH + SACCH + E-IACCH / F</td><td>Not inte- hangover</td>

<td rowspan="4">Not transparent Header extraction Frame structuring</td><td rowspan="4">No conf. Segmentation THE Encryption</td><td rowspan="4">Specialized</td><td rowspan="2">TCH / F</td><td rowspan="2">19 diag</td><td>GMSK</td><td>EEP</td><td>FACCH + SACCH</td><td>Not inte- hangover</td><td rowspan="4">C</td>

<td>8PSK</td><td>EEP</td><td>FACCH + SACCH + E-IACCH / F</td><td>Not interest sa</td>

<td rowspan="2">TCH / H</td><td rowspan="2">19 diag</td><td>GMSK</td><td>EEP</td><td>FACCH + SACCH</td><td>Not inte- hangover</td>

<td>8PSK</td><td>EEP</td><td>FACCH + SACCH + E-IACCH / F</td><td>Not inte- hangover</td>

<td rowspan="4">Not transparent Header compression Frame structuring</td><td rowspan="4">No conf. Segmentation THE Encryption</td><td rowspan="4">Specialized</td><td rowspan="2">TCH / F</td><td rowspan="2">19 diag</td><td>GMSK</td><td>EEP</td><td>FACCH + SACCH</td><td>Not inte- hangover</td><td rowspan="4">D</td>

<td>8PSK</td><td>EEP</td><td>FACCH + SACCH + E-IACCH / F</td><td>Not inte- hangover</td>

<td rowspan="2">TCH / H</td><td rowspan="2">19 diag</td><td>GMSK</td><td>EEP</td><td>FACCH + SACCH</td><td>Not interest sa</td>

<td>8PSK</td><td>EEP</td><td>FACCH + SACCH + E-IACCH / F</td><td>Not inte- hangover</td>

<td rowspan="4">Not transparent Header compression Frame structuring</td><td rowspan="4">No conf. Segmentation THE Encryption</td><td rowspan="4">Shared</td><td rowspan="2">PCH / F</td><td rowspan="2">8 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Not inte- hangover</td><td rowspan="4">b</td>

<td>8PSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Not inte- hangover</td>

<td rowspan="2">PCH / H</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Not inte- hangover</td>

<td>8PSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Not inte- hangover</td>

Five radio access bearers A to E are defined as tributary radio access bearers. The first one labeled A is provided for optimized tributary reuse of the GSMCS mode data link layer 115. Carrier A uses a depth 19 diagonal interleaver for a full rate circuit switched traffic channel that can be modulated with either GMSK or 8PSK. The coding scheme for these two alternatives is different as are the signaling mapping schemes. When GMSK modulation is used, the FACCH and SACCH control channels are used in conjunction with TCH / F14.4 and F9.6 encoding. This is a traffic channel for data transmission specified in the GSM 05.02 specification. The numbers correspond to the bit rate: 14.4 kbit / s and 9.6 kbit / s respectively. When 8PSK modulation is used on the traffic channel, the FACCH and SACCH control channels can be operated together with E-IACCH / F. These allow the use of E-TCH / F28.8, 32.0 or 43.2 encoding. In this case the numbers correspond to the bit rate of each coding scheme, that is, 28.8 kbit / s, 32 kbit / s and 43.2 kbit / s respectively. These coding schemes are used for the ECSD (EDGE Circuit Switched Data service) as equal error protection.

The second tributary radio access bearer B uses transparent mode 102 at the PDCP layer 101 of the protocol stack. The transparent mode 106 of the RLC layer 105 is also used in conjunction with the specialized mode 110 in the MAC layer 109. The physical layer 112 is configured to provide channels over

ES 2 219 635 T3 circuit switching using a depth 19 diagonal interleaving policy. Using either GMSK or 8PSK modulation on the channel to maintain data integrity, various signaling encoding and mapping policies can be implemented such as can be seen in Table 7. Encoding and signaling are equivalent to A although the protocol stack is configured differently. The correspondence of the signals follows Figures 2, 3 and 4 depending on the channel speed.

The third tributary radio access bearer C uses the non-transparent mode 103 of the PDCP layer of the protocol stack. Additionally the headers are adapted by extraction and then frame structuring is carried out. The protocols path is then configured to use the unacknowledged mode 107 at the RLC layer 105 including segmentation, LA, and encryption. Layer 109 specialized mode 110 is also used. Several options are available below for channel operation as set forth in Table 7. This provides optimized stream with header extraction. The correspondence follows Figures 2, 3 and 4 depending on the channel speed.

The fourth tributary radio access bearer D provides an optimized influx with header compression. Bearer D uses non-transparent mode 103 in the PDCP layer of the protocol stack including header compression and frame structuring. The unacknowledged mode 107 is also used from the RLC layer 105 along with segmentation, LA, and encryption. MAC layer 109 is configured to operate in specialized mode 110 while physical layer 112 is configured to operate in circuit-switched mode 113. Table 7 shows various interleaving, modulation, encoding, and mapping protocols that can be implemented.

The fifth tributary radio access bearer E provides a generic tributary radio access bearer. The protocol stack is configured as shown in Table 7 and Figure 11. The non-transparent mode 103 in the PDCP layer 101 is selected and configured for header compression and frame structuring. The unacknowledged mode 107 is used at the RLC layer 105 in conjunction with segmentation, LA, and encryption. Shared 111 is used from MAC layer 109. Packet-switched mode 114 is selected from the physical layer. By configuring the protocol stack in this way, the various options for traffic channels listed in Table 7 become available. This bearer uses the PACCH and PTCCH control channels as previously described herein. The mapping follows Figures 2, 3 or 4 depending on the channel speeds. To benefit from longer interleaving, two voice frames are encapsulated within a packet. However, only one data frame can be encapsulated.

Figure 12 shows the protocol stack for interactive radio access bearers. The protocol stack includes the same modes and layers as those corresponding to Figure 10 although the routing and selection of the modes is different as indicated by the arrows that indicate the path of the possible bearers. Blocks or modes shown through a dashed line are not used. The paths indicated by the arrows are detailed in Table 8. Only two radio access bearers are provided and they are labeled A and B.

(Table goes to next page)

ES 2 219 635 T3

TABLE 8

Interactive radio access bearers

<td rowspan="2">PDCP</td><td rowspan="2">RLC</td><td rowspan="2">MAC</td><td colspan="3">PHY</td><td rowspan="2">Encoded cation</td><td rowspan="2">Signaling zation</td><td rowspan="2">YOU</td><td rowspan="2"></td>

<td>Channel</td><td>Intercal.</td><td>Mod.</td>

<td rowspan="4">Not transparent Header compression Frame structuring</td><td rowspan="4">Conf. Segment- tation THE Encryption BEC</td><td rowspan="4">Shared</td><td rowspan="2">PCH / F</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Do not care</td><td rowspan="4">TO</td>

<td>8PSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Do not care</td>

<td rowspan="2">PCH / H</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Do not care</td>

<td>8PSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Do not care</td>

<td rowspan="4">Not transparent Frame structuring</td><td rowspan="4">Conf. Segment- tation THE Encryption BEC</td><td rowspan="4">Shared</td><td rowspan="2">PCH / F</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Do not care</td><td rowspan="4">B</td>

<td>8PSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Do not care</td>

<td rowspan="2">PCH / H</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td>Do not care</td>

<td colspan="2">1 8PSK, EEP j</td><td>PACCH + PTCCH</td><td>Do not care</td>

The first of these A's occurs through mode 103 of the PDCP layer 101, which is a non-transparent mode that adapts the header through compression and frame structuring techniques. The acknowledged mode 108 is chosen from the RLC layer 105 along with segmentation, LA, and reverse error correction (BEC) and encryption. Shared mode 111 of MAC layer 109 is also implemented in the protocol stack. Packet-switched traffic channels are used with full-rate or half-rate channels depending on the required channel rate as shown in Figures 2, 3, or 4. PACCH and PTCCH channels can be used as described. described earlier herein. The reference to operational scenarios is not relevant for interactive access bearers.

The second interactive bearer B is implemented in a similar way although the adopted PDCP mode does not use header compression. This bearer provides a generic interactive radio access bearer. The correspondence of the channels follows Figures 2, 3 or 4 depending on the speed of the channels.

Figure 13 illustrates the protocol stack for deferred radio access bearers. The protocol stack includes the same modes and layers as the corresponding ones shown in Figures 10, 11 and 12 although it uses different modes thereof through a different routing method as shown through the arrows. Blocks shown through a dashed line are not used. The paths shown by the arrows in Figure 13 are described in more detail in Table 9. Four deferred radio access bearers A to D are defined.

ES 2 219 635 T3

TABLE 9

Deferred radio access bearers

<td rowspan="2">PDCP</td><td rowspan="2">RLC</td><td rowspan="2">MAC</td><td colspan="3">PHY</td><td rowspan="2">Encodes- tion</td><td rowspan="2">Signaling zation</td><td rowspan="2">YOU</td><td rowspan="2"></td>

<td>Channel</td><td>Intercal.</td><td>Mod.</td>

<td rowspan="4">Not transparent Header compression Frame structuring</td><td rowspan="4">Conf. Segment- tation THE Encryption BEC</td><td rowspan="4">Specialized</td><td rowspan="2">TCH / F</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>SACCH + FACCH</td><td> 2</td><td rowspan="4">TO</td>

<td>8PSK</td><td>EEP</td><td>SACCH + FACCH</td><td> 2</td>

<td rowspan="2">TCH / H</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>SACCH + FACCH</td><td> 2</td>

<td>8PSK</td><td>EEP</td><td>SACCH + FACCH</td><td> 2</td>

<td rowspan="2">Not transparent Frame structuring</td><td rowspan="4">Conf. Segment- tation THE Encryption BEC</td><td rowspan="4">Specialized</td><td rowspan="2">TCH / F</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>SACCH + FACCH</td><td> 2</td><td rowspan="4">B</td>

<td>8PSK</td><td>EEP</td><td>SACCH + FACCH</td><td> 2</td>

<td rowspan="2"></td><td rowspan="2">TCH / H</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>SACCH + FACCH</td><td> 2</td>

<td>8PSK</td><td>EEP</td><td>SACCH + FACCH</td><td> 2</td>

<td rowspan="4">Not transparent Header compression Frame structuring</td><td rowspan="4">Conf. Segment- tation THE Encryption BEC</td><td rowspan="4">Shared</td><td rowspan="2">PCH / F</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td> 3-4</td><td rowspan="4">C</td>

<td>8PSK</td><td>EEP</td><td>PACCH + PTCCH</td><td> 3-4</td>

<td rowspan="2">PCH / H</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td> 3-4</td>

<td>8PSK</td><td>EEP</td><td>PACCH + PTCCH</td><td> 3-4</td>

<td rowspan="4">Not transparent Frame structuring</td><td rowspan="4">Conf. Segment- tation THE Encryption BEC</td><td rowspan="4">Shared</td><td rowspan="2">PCH / F</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td> 3-4</td><td rowspan="4">D</td>

<td>8PSK</td><td>EEP</td><td>PACCH + PTCCH</td><td> 3-4</td>

<td rowspan="2">PCH / H</td><td rowspan="2">4 rect</td><td>GMSK</td><td>EEP</td><td>PACCH + PTCCH</td><td> 3-4</td>

<td>8PSK</td><td>EEP</td><td>PACCH + PTCCH</td><td> 3-4</td>

ES 2 219 635 T3

The first of them A in Table 9 is provided by selecting the non-transparent mode 103 of the PDCP layer 101 along with header compression and frame structuring. The RLC layer 105 is configured using the acknowledged mode 108 that enables segmentation, LA, encryption, and BEC. MAC layer 109 is implemented using a specialized channel structure by selecting mode 110. The circuit-switched channels are then used by selecting the TCH modes. This satisfies OS2 and provides silent packet transmission of the circuit-switched channels. Best effort (or lazy) data is provided with header compression in OS2. The control associated with the packet data is carried out by means of the associated control channels of the voice traffic channel (FACCH and SACCH). The best effort data packets are mapped over four consecutive bursts.

The second deferred radio access bearer (B of Table 9) is implemented as shown in Table 9 using non-transparent mode 104, acknowledgment mode 108, dedicated mode 110, and switched mode 113 of circuits. This also provides silent packet transmission though best effort (or lagged) data with no header compression on OS2. The control associated with packet data is carried out through the associated control channels of the voice traffic channel (FACCH and SACCH). The best effort data packets are mapped over four consecutive bursts.

The third deferred radio access bearer (C of Table 9) is implemented using the non-transparent mode 103 of the PDCP layer 101, the acknowledgment mode 108 of the RLC layer 105, the shared mode 111 of the MAC layer 109 and packet switched mode 114 of physical layer 112. The bearer implements OS3 and OS4 and provides a deferred radio access bearer with header compression.

The fourth deferred radio access bearer (D of Table 9) provides a generic deferred radio access bearer. This is implemented using the non-transparent mode 104 of the PDCP layer 101, the acknowledgment mode 108 of the RLC layer, the shared mode 111 of the MAC layer 109, and the packet-switched mode 114 of the physical layer 112. The mapping follows Figure 2, 3 or 4 depending on the channel speed and the bearer can work with OS3 and OS4.

So far, the possible associated control channels required for GERAN have been described. These depend on the type of traffic channel used through the interface. For packet traffic channels the PACCH clearly meets the signaling requirements for the deferred and interactive traffic classes. However when considering the classes of conversational and affluent traffic the only way to transmit the PACCH is by stealing voice packets. The influence on the voice quality could be reduced. However as the TA and PC updates do not use the PACCH and as the measurement reports may be limited, the PACCH traffic could be reduced. However, it is advantageous to reuse existing circuit-switched traffic channels in which more efficient associated control has been defined.

For circuit switched traffic channels the SACCH and FACCH are tailored to the signaling requirements of the affluent and conversational traffic classes.

The embodiments of the present invention take place in the GERAN which means that the physical layer is mainly connected to the core network by packet switching although it can also be connected to the core network by circuit switching. Previously, a circuit-switched air interface (TCH + SACCH + inactive) connected to a core network by circuit-switched (through interface A) and on the other side of a circuit-switched air interface has been previously arranged. packets (PDTCH + PTCCH + idle, i.e. PDCH) connected to a core network by packet switching (via the Gb interface). The embodiments of the present invention allow the circuit-switched air interface to be connected to a core packet-switched network (via the Gb or Iu-ps interfaces), and allow the circuit-switched air interface it can work with packet data (not only the TCH) and therefore it is also connected to a core network by packet switching (through the Gb or Iu-ps interfaces). Thus a possible combination across the circuit switched air interface will be PDTCH + SACCH + inactive. In the case of OS2 a possible combination will be TCH + PDTCH + SACCH + inactive. Combinations in which a communication system according to the present invention can be implemented.

The GERAN is used as an example of a system in which a communication system according to the present invention can be implemented. However, the systems and methods described herein according to the invention are not limited to those used in GSM or EDGE; a system or method according to the invention can also be applied in other radiocommunication networks.

The GERAN is used as an example of a system in which a communication system according to the present invention can be implemented.

Those skilled in the art will understand that the present invention is not limited to the above examples but instead modifications could be made without departing from the scope of the present invention.

ES 2 219 635 T3

Annex A-Content of the associated control channels

<td></td><td>U; * 'F'^r Ϊ, 'Ί ·?', · ·. I · '74 í '* Íí.i' · «-&gt; - Ψ '* 6' ......... *» '* ¿I «. ,. ·, , Messages . -. ._t, j</td>

<td>SACCH</td><td>Measurement Report-Uplink System Information Type 5-Downlink System Information Type 6-Downlink System Information Type 5 bis-Downlink System Information Type 5b-Downlink Measurement Order Extended-Downlink Extended Measurement Report-Uplink SID frames in the case of DTX</td>

<td>FACCH</td><td>Additional allocation-downlink Assignment order-downlink Assignment Complete-Uplink Assignment failure-uplink Channel-Downlink Mode Modification Channel-uplink mode change acknowledgment Channel-downlink release Encryption Mode Order-Downlink Encryption mode completed-uplink Class rebrand-uplink Class brand query-downlink Configuration change order-downlink Configuration Change Acknowledgment-Uplink Configuration change rejection-uplink Frequency redefinition-downlink Handover access Handover order-downlink Handover completed-uplink Handover Failure-Uplink Notification / FACCH-downlink Cell-RR-downlink change order Search response-uplink Partial release-downlink Partial release completed-uplink Physical information-downlink RR initialization request-uplink Speaker Indication-Uplink Busy uplink-downlink-only VGCS Free Uplink-Downlink-VGCS Only VGCS uplink release-only</td>

<td>PACCH</td><td>Packet access rejection-downlink Packet Control Acknowledgment-Uplink Packet Cell Change Order-Downlink Packet Cell Switch Failure-Uplink Packet-Uplink Downlink Conf / Nconf Packet Downlink Conf / Nconf EGPRS-Uplink Packet Downlink Assignment-Downlink Downlink Mapping by EGPRS Packet-Downlink Packet-Uplink Downlink Padding Control Block Packet-Downlink Downlink Allocation Packet-Uplink Uplink Padding Control Block Packet-Uplink Measurement Report</td>

ES 2 219 635 T3

Packet measurement order-downlink

TBF status of mobile packet-uplink

Packet Search Request-Downlink

Packet PDCH request-downlink

Packet polling request-downlink

Power Control / Packet Time Advance-Downlink

Packet-uplink resource request

EGPRS Packet Resource Request-Uplink

EGPRS Packet Resource Request-Uplink

System information type packet 1-downlink

2-downlink packet system information type

System information type packet 3-downlink

3 bis-downlink packet system information type

System information type packet 4-downlink

13-downlink packet system information

TBF packet release-downlink

Packet Uplink Conf / Nconf-Downlink

EGPRS Packet Uplink Conf / Nconf-Downlink

Packet-Downlink Uplink Assignment

Uplink Mapping by EGPRS Packet-Downlink

Reconfiguration of packet timeslots-downlink

Reconfiguration of EGPRS-downlink packet timeslots