Data transfer method and transceiver equipment
Field of the Invention
The invention relates to a data transfer method and a transceiver equipment in a digital mobile network, in which method an information to be transmitted is channel coded for the transmission.
Background of the Invention
Demands on data transfer systems increase continuously. This concerns particularly wireless data transfer systems, such as cellular radio sys¬ tems, which are supposed to comprise more and more versatile services, such as data services of different kinds.
Traditionally, wireless data transfer systems have been used only for speech transmission. An increasing number of different services to be trans- mitted means especially in wireless systems that the system shall be capable of transferring signals having different capacities over the radio path. Accordingly, the data transfer system should be capable of operating effectively in an envi¬ ronment where transmissions of many different service types are transferred.
Data transfer over limited bandwidth radio channel is sort of a com- promise between bit error ratio describing the quality of the transfer and data transfer rate experienced by the user. Bit error ratio can be decreased by in¬ creasing channel coding, which introduces into the information to be transferred extra information less essential for the user, i.e. redundancy. If the number of the bits to be transferred per time unit is limited, the data transfer rate experi- enced by the user decreases with the redundancy.
In GSM system, for instance, a full rate channel has a bit rate of 22.8 kbit/s on the radio path. The used coding methods reduce the bit rate down to the values 12 kbit/s and 6 kbit/s, which correspond to user data rates 9.6 kbit/s and 4.8 kbit s, i.e. data services TCH/F9.6 and TCH/F4.8. Outgoing data transferred over the radio path are transmitted further from base stations to base station controllers and to a mobile exchange and, on the other hand, incoming data are transmitted from the mobile exchange to a base station controller and
further to a base station to be transferred over the radio path.
Besides the above intentional redundancy, the current GSM data services have overhead in the user information. In transparent (T) service, the overhead consists of flow control signalling and in non-transparent (NT) service of radio link protocol (RLP) frame headers and L2R flow control. In both cases the user gets a data rate of 9.6 kbit/s or 4.8 kbit/s at the most for his/her own use, depending on whether TCH/F9.6 or TCH/F4.8 service is concerned. At the moment, the user has no opportunity for a higher data rate in networks of GSM type, though there is a great need for this with data services becoming general. On the other hand, should a service use a user data rate of 7.2 kbit/s, it must at present be transmitted over TCH/F9.6, due to which also the bit error ratio is the same as the bit error ratio of a conventional TCH/F9.6. This makes the service unreliable in places where the signal strength is usually weak or the interference level is high, e.g. at the border areas of a cell, inside build- ings. Continuous retransmissions make the data flow slower and a transfer of large files uncomfortable. All this is time-consuming and reduces the total ca¬ pacity of the mobile network.
It would be beneficial, however, if the user data rate of 7.2 kbit/s could be provided with a better bit error ratio than is achieved by TCH/F9.6. There are several equipments supporting the 7.2 kbit/s user data rate, for ex¬ ample modems, such as ITU V.32 bis and V.34 modems, as well as group 3 telefax terminals. It would also be advantageous, if this new data service pro¬ vided an open platform for future link protocols and were compatible with the current link protocols. On the other hand, it should be possible to introduce a new data rate into an existing mobile network with as small modifications as possible.
Disclosure of the Invention
The object of the invention is to provide a 7.2 kbit/s data transfer in a mobile network in such a way that the above requirements are met.
This is achieved by a data transfer method in a digital mobile net¬ work, the method comprising the following steps: receiving user information bits,
which should be transmitted over the radio interface, block coding of the infor¬ mation bits into a block of 152 bits, convolutional coding of a block of 152 bits at the ratio 1/3 in such a way that a block of 456 bits is formed, interleaving, burst building, modulating, transmitting a modulated burst onto the air interface. Ac- cording to the invention, the method is characterized by receiving said user in¬ formation bits at a nominal user data rate of 7200 bit/s, and forming said block of 152 bits by providing it with 144 information bits and 8 other bits.
The object of the invention is also a transceiver equipment of a digital mobile system, the equipment comprising an interface, over which infor- mation bits to be transmitted are inputted to the transceiver equipment, a block coding means for block coding said information bits into blocks of 152 bits, a convolutional coding means for convolutional coding the block of 152 bits at the convolutional ratio 1/3 in such a way that a block of 456 bits is formed, means for interleaving, burst building and modulation, a transmitter for transmitting a modulated burst onto the radio path. The equipment is according to the inven¬ tion characterized in that the nominal user data rate of said interface is 7200 bit/s, the block coding means is arranged to form said block of 152 bits in such a way that it comprises 144 information bits and 8 other bits.
The present invention utilizes the fact that, in a TCH/F4.8 service according to the recommendations of the GSM mobile system or another digital mobile system having channel and frame structure and channel coding similar to those of the GSM system, fill bits are inserted among the user data information bits during block coding, which fill bits have no effect on error correction per¬ formance. This results in that the data rate before convolutional coding is 7600 bit/s, which means an overhead of 1600 bit/s, the nominal bit rate of the traffic channel on the radio interface being 6000 bit/s. In the present invention, this overhead capacity is used for supporting the 7200 bit/s user rate. In other words, the number of information bits is increased and the number of fill bits is de¬ creased in such a way that the user rate increases, but the data rate before convolutional coding still is 7600 bit/s. Consequently, the same convolutional coding, i.e. 1/3, can be used for the 7200 bit/s user rate as is used for the 4800 bit/s user rate. The 7200 bit s service according to the invention remains com-
patible with the TCH/F4.8 service and it has the same low bit error ratio. There¬ fore, the new 7200 bit/s service can be used in areas where radio conditions are too poor for the current TCH/F9.6 service data rates, for which is used 1/2 con¬ volutional coding. In connection with multislot bearer services, the invention may also be utilized at data rates being multiples of 7200 bit/s, such as 14400 or 28800 bit s.
Several other advantages are achieved by the method of the inven¬ tion. A desired data rate is achieved by the method without any substantial changes in the current mobile networks. Because the user rate and better bit er¬ ror ratio shorten the transfer time, the resources and capacity of the mobile net¬ work are used efficiently.
Brief Description of the Drawings In the following, the invention is described in more detail with refer¬ ence to the examples according to the attached drawings, in which
Figure 1 illustrates a cellular radio system, to which the method of the invention can be applied,
Figures 2a to 2c show different alternatives to the location of a transcoding unit,
Figures 3 and 4 show functional block diagrams for a GSM trans- mitter and receiver, respectively, which implement both the current TCH/F4.8 service and the TCH/F7.2 service according to the invention,
Figure 5 illustrates block coding according to TCH/F4.8, Figure 6 illustrates block coding according to the invention.
Preferred Embodiments of the Invention
Preferred embodiments of the invention will be described in the fol¬ lowing in connection with the European digital mobile system GSM. However, the invention can preferably be applied to a digital cellular radio system having channel and frame structure as well as channel coding similar to those of the GSM system. To these belong e.g. DCS1800 (Digital Communication System)
and the digital cellular system PCS (Personal Communication System) of the United States.
The invention will be described below by using the GSM mobile system as an example. The configuration and operation of the GSM system are well-known by one skilled in the art and defined in the GSM specifications of ETSI (European Telecommunications Standards Institute). In addition, reference is made to the book "The GSM System for Mobile Communications", M. Mouly and M. Pautet, Palaiseau, France, 1992; ISBN:2-9507190-0-7.
Figure 1 illustrates the configuration of a cellular radio system of GSM type. The system comprises a number of terminal equipments 202 to 206 having connections 208 to 212 to a base station 200. The base station 200 is connected via digital transmission links 218 to a base station controller 214, to which are subordinated one or several base stations. On the other hand, the base station controller 214 is connected via digital transmission links 220 to a mobile exchange 216, which has a further connection 222 to the other parts of the network.
The interface 218 between the base station 200 and the base sta¬ tion controller 214 is called Abis interface. Correspondingly, the interface 220 between the base station controller 214 and the mobile exchange is called A interface. Two common ways of implementing said interfaces are in use. Essen¬ tial for both ways is the transfer rate used at the Abis interface, this rate being either 64 kbit/s or 16 kbit/s. Due to the 64 kbit/s transfer rate employed by the exchange 216 in its switching, the signal must be subjected to transcoding, and therefore, the location of a transcoding unit TRAU in the network depends on the transfer rate used at the Abis interface. Figures 2a to 2c illustrate different alternative network structures at different transfer rates.
Figure 2a illustrates an alternative, in which the Abis interface 218 between the base station 200 and the base station controller 214 is imple¬ mented at 64 kbit/s rate. Then the transcoding unit 300 is located in association with the base station 200. The connection 220 between the base station con¬ troller 214 and the mobile exchange 216 then also has the rate of 64 kbit/s.
Figure 2b illustrates an alternative, in which the Abis interface 218 between the base station 200 and the base station controller 214 is imple¬ mented at 16 kbit/s rate. Then the transcoding unit TRAU 300 is located in as¬ sociation with the base station controller 214. The connection 220 between the base station controller 214 and the mobile exchange 216 then has the rate of 64 kbit/s.
Figure 2c illustrates a further alternative in which the Abis interface 218 between the base station 200 and the base station controller 214 is imple¬ mented at 16 kbit/s rate. In this case, the transcoding unit TRAU 300 is located in association with the mobile exchange 216. The connection 220 between the base station controller 214 and the mobile exchange 216 then has the rate of 16 kbit/s.
However, the invention relates to channel codings performed in the mobile station and the base station, and the location of the transcoding unit is not essential for the invention. The solutions required by the invention are identi¬ cal to those of the current TCH/F4.8 service.
In the method of the invention, the object of which is to enable a higher user data transfer rate than earlier in a cellular radio system, a new way of implementing channel coding on the radio path is presented. Changes caused by the new channel coding in the current systems are minor, but they enable the 7.2 kbit/s transfer rate for the user at the same bit error ratio as the 4.8 kbit/s transfer rate has. The method according to the invention is at first ex¬ amined in connection with transmission.
Figures 3 and 4 show functional block diagrams for a GSM trans- mitter or receiver, which implement the current TCH/F4.8 service and the 7.2 kbit s service according to the invention.
In Figure 3, a channel encoder 300 receives data bits from an in¬ terface 30 at 4.8 kbit s or 7.2 kbit/s user rate in data blocks of 60 bits or 36 bits. At the 4.8 kbit s data rate, the channel encoder 300 performs the measures ac- cording to the GSM Specification 05.03, version 4.1.1 : block encoding 31 , con¬ volutional encoding 32, interleaving 33 and mapping 34 on a burst (burst build¬ ing). At the 4.8 kbit/s user rate, the channel encoding 300 agrees with the GSM
Specification 05.03, as will be explained below. The bursts formed by the chan¬ nel encoder 300 are modulated in a modulator 35 and the modulated signal is transmitted onto the radio path by a transmitter 36 via a transmitting antenna 37. In Figure 4, a signal at radio frequency is received via a receiving antenna 46 by a radio receiver 41 , which modulates the signal to base band fre¬ quency, and subsequently, the signal is demodulated in a demodulator 42 and fed to a channel decoder 400. The channel decoder performs a deinterleaving 43, convolutional decoding 44 and block decoding 45 of the signal, and then the signal having 4.8 or 7.2 kbit/s user rate is fed to an interface 40 similar to the in- terface 30. The channel decoder 400 performs reversed operations with respect to the channel encoder 300.
Firstly, the 4.8 kbit s user rate is examined with reference to the Figures 3 and 5. The block encoding 31 in the channel encoder 300 receives user data in data blocks of 60 bits at the rate of 1 data block in 10 ms from the interface 30. The interface 30 and the 60 bit frames are modified CCITT V.110 frames according to the GSM Specification 04.21 , version 4.4.0. In case if the user unit transmits to the encoder a bit stream organized in blocks of 240 infor¬ mation bits, these blocks occurring every 40th millisecond, the bits are pro¬ cessed in four blocks of 60 bits. Such blocks of 240 bits can be for example RLP (Radio Link Protocol) frames used in non-transparent GSM data services. These blocks of 60 bits are illustrated generally by blocks 50A and 50B in Figure 5.
The block encoder 31 of the channel encoder 300 performs a block coding of the 60 bit blocks according to the GSM Specification 05.03, version 4.1.1 , chapter 3.4.2, whereby 16 bits equal to 0 are inserted in addition to these 60 information bits, the result being a block of 76 bits, such as blocks 51 A and
51 B in Figure 5. To be more precise, the information bits are provided with 12 further bits, indicated by symbol x in Figure 5, and 4 tail bits. In Figure 5, the block 51 A of 76 bits is formed of the block 50A of 60 bits and the block 51 B of 76 bits is formed of the block 50B of 60 bits, respectively. In a later coding proc- ess, two such blocks of 76 bits, 51 A and 51 B, are dealt with as one block 52 of
152 bits. The bit rate after block coding is thus 156 bits/20 ms, which corre¬ sponds to the rate of 7.6 kbit/s.
The block 52 of 152 bits is applied to the convolutional encoder 32, which performs a convolutional coding according to the GSM Specification
05.03, version 4.1.1, chapter 3.4.3 at the ratio 1/3, the result being a block 53 comprising 3 x 152 = 456 coded bits. This corresponds to a transfer rate of 22.8 kbit/s on the radio path.
At present, blocks of 456 bits are interleaved according to the GSM Specification 05.03, version 4.1.1 , chapter 3.3.4, in the interleaver 33. On inter¬ leaving, the 456 encoded and reorganized bits from four predetermined data blocks are divided into 22 bursts in such a way that 6 bits are divided evenly to the first and the 22nd burst, 12 bits are divided to the second and the 21st burst, 18 bits are divided to the third and the 20th burst and 24 bits are divided to the remaining 16 bursts. The coded data block is interleaved "diagonally", whereby a new block of coded data begins in every 4th burst and is divided over 22 bursts. The mapper 34 maps the coded bits on bursts according to the
GSM Specification 3.1.4., version 4.1.1 , chapters 3.1.4 and 4.2.5. The encoder 300 feeds the bursts to the modulator 35.
In the present invention, the fact is utilized that in the TCH/F4.8 service according to the specifications, which service was described above in connection with Figure 5, the fill bits x inserted among the information bits on block coding have no effect on error correction performance. Additionally, the data rate before convolutional coding is 7600 bit/s, which means an overhead of 1600 bit/s, the nominal bit rate of the traffic channel on the radio interface being 6000 bit/s. In the present invention, this overhead capacity is used for supporting the 7200 bit s user rate. In other words, the number of information bits is in¬ creased and the number of fill bits is decreased in such a way that the user rate increases, but the data rate before convolutional coding 32 remains unchanged. Consequently, the same convolutional coding, i.e. 1/3, can be used for the 7200 bit/s user rate as is used for the 4800 bit s user rate. Thus the 7200 bit/s service according to the invention remains compatible with the TCH/F4.8 service and it has the same low bit error ratio. In the transmitter and receiver of the Figures 3 and 4, the present invention relates only to the interfaces 30 and 40 and also to
the block coding 31 and the block decoding 45. In other respects, the operation of the transmitter and receiver remains compatible with the TCH/F4.8 service.
In the following, the block coding 31 according to the invention is described with reference to Figure 6. The block encoder 31 in the channel encoder 300 receives informa¬ tion bits from the interface 30 at 7.2 kbit/s user rate. The information bits are grouped into data blocks of 36 bits, which blocks are received at the rate of 1 block/10 ms. The interface 30 and the 36 bit frames are modified CCITT V.110 frames according to the GSM Specification 04.21 , version 4.4.0. Figure 6 shows four blocks 60A to 60D of 36 bits.
The block encoder 31 of the channel encoder 300 performs a block coding, in which two successive blocks of 36 bits, 60A and 60B, are combined and four tail bits are provided. As a result, a block 61A of 76 bits is formed. Be¬ cause the combination of the blocks 60A and 60B produces 72 information bits, no fill bits indicated by symbol x in Figure 5 are introduced. Instead, only four tail bits equal to zero are inserted. In this way is obtained a block of 76 bits com¬ patible with the TCH/F4.8 service, but transferring user data at the 7200 kbit/s user rate. Correspondingly, the blocks 60C and 60D are combined to a block 61 B of 76 bits. The block coding 31 according to the invention can be described also more generally by means of the following equations. The changes required by the invention can be observed by comparing the block coding described be¬ low with the block coding according to the GSM Specification 05.03, chapter 3.4. In the present invention, the blocks consisting of 72 information bits {d(0),...,d(71)} are transmitted by the channel coder every 10th second. To these 72 information bits are added four bits equal to zero, the result being a block containing 76 bits {u(0),...,u(75)}, where u(k)=d(k), when k=0,...,71 u(k)=0, when k=72 75. In a later coding process, two such blocks of 76 bits, 61 A and 61 B, are dealt with as one block 62 of 152 bits. Consequently, the bit rate after the block coding is 156 bits/20 ms, corresponding to the rate of 7.6 kbit/s. More
generally, two blocks of 76 bits form a block comprising 152 bits {u'(0),...,u'(151)} as follows: u'(k)=u1(k), when k=0,...,75(u1= 1st block) u'(k+76)=u2(k), when k=0,...,75(u2= 2nd block). The block 62 of 152 bits is brought to the convolutional encoder 32 performing the convolutional coding described in connection with Figure 5 at the ratio 1/3, the result being a block 63 comprising 3x152=456 coded bits. This cor¬ responds to the transfer rate of 22.8 kbit/s on the radio path.
Blocks 63 of 456 bits are interleaved and mapped on bursts, modulated and transmitted, as described above in connection with Figure 5.
The block encoder 31 may also insert overhead control information in some (at least one) of the eight fill bit locations in the block 62 of 152 bits. In other words, the fill bits or some of them may be utilized to carry inband control information between the transmitter and receiver. The receiving antenna 46 receives a radio frequency signal, which is processed in the radio receiver 41 , demodulator 42, deinterleaver 43 and convolutional decoder 44, as explained earlier in connection with the Figures 4 and 5. The convolutional decoder 44 produces for the block decoder 45 a block of 156 bits, which block corresponds to the blocks 62 of Figure 6. From this block of 156 bits, the block decoder 45 separates two blocks of 76 bits, which blocks correspond to the blocks 61 A and 61 B of Figure 6. Subsequently, the block decoder 45 removes four tail bits from each block 61 A and 61 B splits each block 61 A and 61 B into two blocks of 36 bits, 60A, 60B and 60C, 60D, respec¬ tively. From these the block decoder 45 feeds blocks of 36 bits to the interface 40 at the rate of 1 block/5 ms, which rate corresponds to the 7.2 kbit/s user rate. The invention has been described above in a case when a connec¬ tion employs one traffic channel. However, the present invention is directly ap¬ plicable also to high speed data services HSCSD, in which one connection uses several parallel traffic channels. Then, for instance, the maximum transfer rate of a data connection using two parallel traffic channels is 14.4 kbit/s. In this case, however, both parallel traffic channels are block coded independently according to the invention. Such a high speed data service based on multislot access is
described for instance in copending PCT Application W095/31878 of the Appli¬ cant. In a multichannel case, there is a separate channel encoding unit 300 for each parallel channel, as illustrated in Figure 3, and a separate channel decod¬ ing unit 400 for each parallel channel, as illustrated in Figure 4. Altematively, one unit can process each channel sequentially, one at a time.
The attached figures and the specification relating to them are only intended to illustrate the present invention. As to the details, the present inven¬ tion may vary within the scope and spirit of the attached claims.