Universal mobile telephone system network with improved rate matching method

0001This invention relates to a Universal Mobile Telephone System (UMTS) network and more especially to a method of calculating the number of bits to be punctured or repeated to achieve effective rate matching.

0002In a UMTS system, the technique of time multiplexing is used to combine different services required by one user. The services have different Quality of Service (QoS) requirements, such as signal power with respect to noise and interference power, and latency restrictions which lead to different requirements of transmission power; transmission power must therefore be adjusted so that in any one transmission channel the QoS is maintained for the service making the highest demand, but simultaneously transmission power is minimised.

0003To achieve this, and also to adapt the symbol data rate to the physical channel data rate, the method of rate matching is used, ie, some bits are punctured or repeated. The principle of rate matching is described by the Ericsson company in the paper "Description of Rate Matching in UTRA/FDD," TDoc SMG2 UMTS L1 235/98 and by the Nokia company in the paper "Mapping Rule for semi static rate matching," TDoc SMG2 UMTS L1 476/98. The publication by Nokia also proposes a method of deriving the number of bits to be punctured or repeated to achieve the required QoS for two different services by interpolation methods. However, interpolation cannot be applied to more than two services, which is a substantial limitation on the network and also the algorithm is complex.

0004According to the invention, in a UMTS network in which a plurality of services of a single user having different transmission power requirements are multiplexed in one channel and the technique of rate matching is applied, a method of determining for each service the number of bits to be punctured or repeated characterised by the steps of deriving for each service the Energy per Bit per Noise density E_B/N_O required to achieve a desired Bit Error Rate; for each E_B/N_O deriving a value of the Energy per coded Symbol per Noise density E_S/N_O; obtaining for each service the number of data bits given by the data rate; selecting for each service a spreading factor SF using the number of data bits; deriving the semi static rate matching factors SRF_i, where <maths id="math0001"><math display="block"><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub><mo>=</mo><mfrac><msub><mi>N</mi><mrow><mi>R</mi><mo>,</mo><mi>i</mi></mrow></msub><mrow><msub><mi>N</mi><mrow><mi>S</mi><mo>,</mo><mi>i</mi></mrow></msub><mo>⋅</mo><mi>G</mi><mfenced><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub></mfenced></mrow></mfrac><mo>,</mo></math><img file="EP1069798B1_D0001.tif" /></maths> and <i>N_Rj</i> is the number of code symbols after semi static rate matching <i>N_S,i</i> is the number of code symbols before rate matching, and <i>G</i>(<i>SRF_i</i>) is the coding gain associated with the level of semi-static rate matching; calculating the rate matching factor RF_i for each service from the semi static rate matching factor SRF_i and a dynamic rate matching factor DRF; and calculating the number of bits to be punctured or repeated from the rate matching factor RF_i.

0005The invention will now be described by way of example only with reference to the accompanying drawings in which:- <ul id="ul0001" list-style="none"><li><figref idref="f0001">Figure 1</figref> is a schematic view of a UMTS network at network level;</li><li><figref idref="f0002">Figure 2</figref> shows the conventional steps in UMTS transport channel multiplexing for the uplink;</li><li><figref idref="f0003">Figure 3</figref> shows rate matching gain;</li><li><figref idref="f0003">Figure 4</figref> shows the major steps in a first example of a method according to the invention; and</li><li><figref idref="f0003">Figure 5</figref> shows a relationship between channel coding and rate matching in a second example of a method.</li></ul>

0006In <figref idref="f0001">Figure 1</figref>, a UMTS network 10 comprises a number of mobile users 12, 14 each connectable to a base station BTS/Node B 16. The BTS 16 is connected through a Base Station Controller BSC 18 in a Radio Network Controller (RNC) 20 to a Mobile Switching Controller (MSC) 21 in the UMTS Core Network (CN) 22. The CN 22 provides connections to the public switched telephone network PSTN 24, integrated services digital network ISDN 26 and the internet 28.

0007The Radio Access Network (RAN) comprises the mobile users 12, 14 and the BTS 16. To multiplex signals in the RAN in the uplink direction (from the mobile 12,14 to the BTS 16), the arrangement of <figref idref="f0002">Figure 2</figref> is used; the steps are performed in the appropriate mobile 12 or 14. <figref idref="f0002">Figure 2</figref> shows transport channel multiplexing for the uplink and is taken from the standard TS 25.212 V1.1.0 (1999-06).

0008In <figref idref="f0002">Figure 2</figref>, the steps to encode services with identical QoS requirements are shown within box 30, and identical steps to encode a set of different services are performed within box 31 (steps not shown). The steps comprise Cyclic Redundancy Coding (CRC) steps 40 for identical services, (but in this invention a single service at step 40 is considered), a first multiplexing step 41, channel coding step 42, first interleaving step 43. radio frame segmentation step 44 and rate matching step 45.

0009The equivalent rate matching step 55 is shown in box 31.

0010The outputs of the two boxes 30, 31 are combined in a second multiplexing step 60 followed by a physical channel segmentation step 61, second interleaving step 62, and physical channel mapping step 63; the outputs are then provided to respective Dedicated Physical Data Channels DPDCH 64.

0011The reverse steps are performed for uplink de-multiplexing at the BTS 16.

0012This entire procedure exists also in the downlink direction, ie from the BTS 16 to a mobile 12 or 14.

0013While the rate matching steps 45. 55 are indicated as separate, they are in fact interrelated because rate matching is based on the relative QoS requirements of different sets of services, and because the channel must be able to support the combined rate matching output of all services, and this interrelationship is indicated at 56 but, as stated above, the interrelationship has previously been specified only for rate matching of two services by interpolation.

0014By application of the method of the invention, rate matching of more than two different services is achievable.

0015In the rate matching steps 45, 55, a set of puncturing or repetition factors is determined based on the requirements that:- <ol id="ol0001"><li>a) desired transmission quality requirements of each transport channel is fulfilled and not significantly exceeded. This means that required transmission power to meet quality requirements for all transport channels is as low as possible.</li><li>b) on the uplink, the total bit rate after transport channel multiplexing is identical to the total channel bit rate of the dedicated physical channels allocated.</li><li>c) on uplink and downlink, the total allocated code resource should be minimised.</li><li>d) the puncturing factors should not exceed a certain maximum puncturing factor, specific for each transport channel.</li></ol>

0016In the current UMTS concept there are two types of rate matching specified: <ol id="ol0002"><li>(i) Semi-static Rate Matching: this is used to balance the transmission power requirements of different services, which are multiplexed to one Common Composite Traffic Channel (CCTrCH). This function is also known as 'ES/NO-balancing' - balancing the ratio of symbol energy to noise density. Because the elements related to service multiplexing are decided by the Radio Resource Control (RRC) layer, this physical layer function is controlled by RRC via the semi-static part of TFCI.</li><li>(ii) Dynamic Rate Matching: this is used in the uplink to adapt the symbol data rate to the data rate of the CCTrCH in order to preserve continuous transmission. It is an autonomous function of the physical layer. Dynamic rate matching is performed through discontinuous transmission (DTX) in the downlink.</li></ol>

0017In a single channel, several services (as indicated by boxes 30, 31) must be multiplexed which have different bit rates and different types of coding. For example a low rate video using 144 kbps and turbo coding; a voice service at 13kbps using convolutional coding, plus signalling information, also using convolutional coding.

0018Each service will have a QoS requirement, and an important part of that requirement is the Energy per Bit per Noise Density E_B/N_o required to achieve the desired Bit Error Rate BER. The E_B/N_O for a service will apply only to specific measurement conditions, such as the environment (pedestrian, indoor, vehicular etc), the level and type of coding (turbo, conventional), the level of physical layer control signalling (Dedicated Physical Control Channel DPCCH overhead) and the spreading factor. Conventionally, the requirements have been calculated by simulation or by measurement of a service in a particular environment.

0019When services having different E_B/N_O requirements are to be multiplexed, certain differences can be adjusted by comparing E_B/N_O requirements; eg, the DPCCH consists of pilot symbols, power control and Transport Format Combination Indicator (TFCI) symbols only, with the amount of energy required being relatively independent of the service and dependent only on environment. Also, most UMTS services use rate 1/3 coding schemes with similar levels of extra overhead such as tail bits. Adjustment of power is therefore relatively easy.

0020Other differences cannot be derived from E_B/N_O values. For example, when services are multiplexed, the Coded Composite Transport Channel CCTrCH will use a spreading factor which is different to the spreading factor that each service would use individually, and variations may be wide. A voice service normally uses a spreading factor of 128 and could be mapped onto a CCTrCH using a spreading factor of 4, when sent together with other services.

0021The method of the invention solves this problem.

0022In the invention, a look up table of E_B/N_O is created; E_B/N_O is expressed as Energy per coded Signal versus noise density, ie E_S/N_O. It must be understood that E_B/N_O is not the bit energy when the data is uncoded, but is the equivalent energy per bit assuming a certain coding scheme, spreading factor and environment. Therefore for every value of E_B/N_O there is a unique value E_S/N_O which in this context does not include the DPCCH overhead.

0023Using the E_s/N_o value for each service I, (E_S/N_O)_i, a required level of semi-stactic rate matching can be calculated. Since it is impossible for the E_S/N_O for the CCTrCH to satisfy the (E_S/N_O)_I, rate matching is needed, such that each service achieves its desired (E_S/N_O)_i.

0024The E_S/N_O for the CCTrCH is set as E'_S/N_O, which is likely to equal the E_S/N_O for one of the constituent services; all other services are therefore forced to use semi-static rate matching. In some circumstances however the E'_S/N_O does not equal the E_S/N_O for a constituent service, in which case all services will be required to use rate matching. The relationship between (E_S/N_O)_n and E'_s/N_o is given by <maths id="math0002" num="(1)"><math display="block"><mfrac><msubsup><mi>E</mi><mi>s</mi><mo>"</mo></msubsup><msub><mi>N</mi><mn>0</mn></msub></mfrac><mo>=</mo><msub><mi>N</mi><mrow><mi>R</mi><mo>,</mo><mi>i</mi></mrow></msub><mo>=</mo><mi>G</mi><mfenced><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub></mfenced><mo>⋅</mo><msub><mfenced><mfrac><msub><mi>E</mi><mi>S</mi></msub><msub><mi>N</mi><mn>0</mn></msub></mfrac></mfenced><mi>i</mi></msub><mo>⋅</mo><msub><mi>N</mi><mrow><mi>S</mi><mo>,</mo><mi>i</mi></mrow></msub></math><img file="EP1069798B1_D0002.tif" /></maths>

0025In order to be satisfied with E'_S/N_O, the service must use a semi static rate matching factor, SRF<i>_i</i>, of <maths id="math0003"><math display="inline"><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub><mo>=</mo><mfrac><msub><mi>N</mi><mrow><mi>R</mi><mo>,</mo><mi>i</mi></mrow></msub><mrow><msub><mi>N</mi><mrow><mi>S</mi><mo>,</mo><mi>i</mi></mrow></msub><mo>⋅</mo><mi>G</mi><mfenced><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub></mfenced></mrow></mfrac><mo>,</mo></math><img file="EP1069798B1_D0003.tif" /></maths> where <i>N_R,i is</i> the number of coded symbols after semi static rate matching <i>N_S,i</i> is the number of code symbols before rate matching, and <i>G</i>(<i>SRF_i</i>), illustrated in <figref idref="f0003">Figure 3</figref>, is the coding gain associated with the level of semi-static rate matching.

0026Thus any service i must use a semi-static matching factor SRF_i.

0027Generally, it can be assumed that adjusting the energy by adding extra symbols is equivalent to adjusting the E_S/N_O. Therefore if SRF; is limited to greater than -1 dB, <i>G(SRF_i</i>) is 1 (0dB). However, puncturing alters the coding scheme used, which results in a loss greater than the decrease in energy. Moreover, repetition may result in extra time diversity, thereby providing coding gain beyond the increase in energy. Repeating some of the bits in a transport block may not be equivalent to increasing the energy of <i>all</i> of the bits. Finally, the gain achieved by repetition depends on the combining method implemented.

0028An additional requirement is that the semi-static rate matched transport block (on leaving box 30 or box 31 in <figref idref="f0002">Figure 2</figref>) must fit into a physical channel having bits per frame N_Frame. One time frame is 10 milliseconds, and contains N_S symbol bits where <maths id="math0004"><math display="inline"><msub><mi>N</mi><mi>S</mi></msub><mo>=</mo><mn>16</mn><mo>⋅</mo><munderover><mo>∑</mo><mrow><mi>j</mi><mo>=</mo><mn>1</mn></mrow><mi>m</mi></munderover><msub><mi>N</mi><mi mathvariant="italic">dataj</mi></msub></math><img file="EP1069798B1_D0004.tif" /></maths> where m=1 for single code transmission, and m is greater than 1 for multicode transition, ie where more than one DPDCH is assigned to one CCTrCH, as in the method according to the invention.

0029In the uplink the number of data bits N_data of one DPDCH in a time slot (0.625 ms) is simply given by N_data = 10·2^k, k = 0...6. In the downlink, due to time multiplex of DPDCH and DPCCH it is given by N_data = 20·2^k-N_pilot-N_TPC-N_TFCI, k = 0...6. The spreading factor of the related DPCH in uplink and downlink is SF = 256/2^k, thus it may range from 256 down to 4 . The rate matching factor RF_i is given by the product of the dynamic rate matching factor DRF and the semi static rate matching factor SFR_i by RF_i = DRF · SFR_i. The dynamic rate matching factor DRF is equal for all multiplexed services, but the semi static rate matching factor SFR_i depends on the service. Two basic conditions must be fulfilled by the rate matching principle: <maths id="math0005" num="(2)"><math display="block"><mfrac><msubsup><mi>E</mi><mi>s</mi><mo>"</mo></msubsup><msub><mi>N</mi><mn>0</mn></msub></mfrac><mo>=</mo><msub><mfenced><mfrac><msub><mi>E</mi><mi>S</mi></msub><msub><mi>N</mi><mn>0</mn></msub></mfrac></mfenced><mi>i</mi></msub><mo>⋅</mo><mfrac><mn>1</mn><msub><mi mathvariant="italic">RF</mi><mi>i</mi></msub></mfrac></math><img file="EP1069798B1_D0005.tif" /></maths><maths id="math0006" num="(3)"><math display="block"><mstyle displaystyle="true"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover></mstyle><msub><mi>N</mi><mi mathvariant="italic">Ri</mi></msub><mo>=</mo><mstyle displaystyle="true"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover></mstyle><msub><mi>N</mi><mi mathvariant="italic">Si</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">RF</mi><mi>i</mi></msub><mo>=</mo><mi mathvariant="italic">DRF</mi><mo>⋅</mo><mstyle displaystyle="true"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover></mstyle><msub><mi>N</mi><mi mathvariant="italic">Si</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">SFR</mi><mi mathvariant="italic">i</mi></msub><mo mathvariant="italic">=</mo><msub><mi mathvariant="italic">N</mi><mi mathvariant="italic">total</mi></msub></math><img file="EP1069798B1_D0006.tif" /></maths> where E"_S/N_O is the E_S/N_O set point for the CCTrCH. (Through dynamic rate matching each service will achieve its individual target of (E_S/N_O)_i) <i>n</i> is the number of services multiplexed onto the CCTrCH. In the downlink, DRF is set to 1. Discontinuous transmission (DTX) is used instead of dynamic rate matching. DTX means that when N_frame > N_total, the difference in the number of bits is simply not transmitted. Hence N_total is simply the sum of all N_Ri, as defined for semi-static rate matching. Therefore, generally E"_S/N_O = E'_S/N_O in the downlink. However, if N_total is just slightly larger than N_frame it is possible to set DRF to between 0.8 and 1.0, thereby applying puncturing instead of DTX to adjust the CCTrCH rate. This method avoids using the next lower spreading factor and the resulting high level of repetition. In the uplink, N_frame must be equal to N_total, since DTX cannot be used. Therefore, DRF is defined by equations 2 and 3, as well as: <maths id="math0007"><math display="block"><mtable><mtr><mtd><msub><mi mathvariant="italic">N</mi><mi mathvariant="italic">frame</mi></msub><mo>=</mo><msub><mi mathvariant="italic">N</mi><mi mathvariant="italic">total</mi></msub></mtd></mtr></mtable></math><img file="EP1069798B1_D0007.tif" /></maths>

0030<figref idref="f0003">Figure 4</figref> indicates five major steps of the method according to the invention; as indicated, steps 1, 2 and 3 are performed in the BSC (reference 18 in <figref idref="f0002">Figure 2</figref>) while steps 4 and 5 are performed in the BTS 16 and mobiles 12, 14. <dl id="dl0001"><dt><b>Step 1:</b></dt><dd>In step 1, Inputs to the rate matching algorithm are the values (E_S/N₀)_i which are determined for each specific service and the number of data bits which is explicitly given by the data rate N_si.</dd><dt><b>Step 2:</b></dt><dd>During this step the rate matching factors, SRF<i>_i</i> and DRF<i>_i</i>, are calculated (equations 1 and 3). First an initial spreading factor (SF) estimate is chosen based such that <maths id="math0008"><math display="inline"><msub><mi mathvariant="italic">N</mi><mi mathvariant="italic">frame</mi></msub><mo>></mo><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover><msub><mi mathvariant="italic">N</mi><mi mathvariant="italic">Si</mi></msub><mn>.</mn></math><img file="EP1069798B1_D0008.tif" /></maths> Using this initial estimate for SF, the set of (E_S/N_O)_i are identified from the look up table for the relevant environment and level of coding. N_total is calculated . If N_frame < 0.8·N_total, the value of SF is too large and Step 2 is performed again using the new value for the spreading factor of SF_new = SF/2.</dd><dt><b>Step 3:</b></dt><dd>The outputs from the RRC part of algorithm are the semi static rate matching factors for each DCH, which are element of the semi-static part of TFCI and the spreading factors for each service combination, which may be part of the dynamic part of the TFCI. The TFCI is used to inform the physical layer of both, the mobiles 12, 14 and the NodeB 16, what physical transmission parameter has to be chosen.</dd><dt><b>Step 4:</b></dt><dd>This is the first step of the execution part of the rate matching algorithm. Only the TFCI is necessary to derive all other parameters for rate matching. The dynamic rate matching factor is automatically given by knowing the semi static rate matching factor for a certain service combination. An example is given in the paper by Mitsubishi Electric and Siemens: "Determination of Rate Matching Parameters for Service Multiplexing," Tdoc 3GPP TSG-RAN WG1 710/99. Now the rate matching factor for each service is calculated by RF_i = DRF · SRF_i. From this, the number of punctured / repeated bits is simply N_RMi = (RF_i-1) . N_codi.</dd><dt><b>Step 5:</b></dt><dd>The last step performs puncturing / repetition of the bits using the value of N_RMi. Several techniques can be applied and an example is given in the paper by Philips "Service Multiplexing," TDoc SMG2 UMTS L1 229/98.</dd></dl>

0031Advantages of the method according to the invention are that the final calculation of the number of bits to be repeated or punctured, and the puncturing or repetition themselves, are carried out in the physical layer in an autonomous arrangement; that the semi-static rate matching is explicitly given; and that the method derives a spreading factor for each service combination for use by the dynamic part of the TFCI. The last two advantages are in contrast to the proposal by the Nokia company (reference above) in which the TFCI elements are not explicitly given but must be calculated separately from the number of bits.

0032Overall, the method of the invention leads to a significant decrease in transmitted power and hence an improvement in overall system capacity.

Example

0033In the following example, three services are multiplexed for a user in a vehicle. In the downlink a spreading factor of 4 is selected resulting in CCTrCH rate of 512ksps. The services are as follows:- <ul id="ul0002" list-style="none" compact="compact"><li>144 kbps low rate video, using turbo coding, requiring E_S/N_O = 5 dB.</li><li>13 kbps voice, using convolutional coding, requiring E_S/N_O = 8 dB.</li><li>Signalling information, using convolutional coding, requiring E_S/N_O = 9 dB.</li><li>E'_S/N_O = 5 dB, hence rate matching is required by each service to achieve the desired (E_S/N_O)_i is:</li><li>video: 0 dB</li><li>voice: 3 dB</li><li>signalling: 4 dB.</li></ul>

0034This is shown in table 1, under the rate matching row. There is no Dynamic rate matching, Discontinuous transmission is used to match T_frame and T_total. <tables id="tabl0001" num="0001"><table frame="all"><title><b>Table 1. Multiplexing example: Downlink.</b></title><tgroup cols="4"><colspec colnum="1" colname="col1" colwidth="81mm" /><colspec colnum="2" colname="col2" colwidth="14mm" /><colspec colnum="3" colname="col3" colwidth="21mm" /><colspec colnum="4" colname="col4" colwidth="19mm" /><thead><row><entry valign="top">Combined Service</entry><entry namest="col2" nameend="col4" align="center" valign="top">Voice, Low Video + Signalling</entry></row><row><entry valign="top">Component Services</entry><entry valign="top">Voice</entry><entry valign="top">Low Video</entry><entry valign="top">Signalling</entry></row></thead><tbody><row><entry>Transport Block size</entry><entry align="right">320</entry><entry align="right">320</entry><entry align="right">80</entry></row><row><entry>Transmission interval and first interleaver duration (ms)</entry><entry align="right">20</entry><entry align="right">10</entry><entry align="right">10</entry></row><row><entry>Number of Transport Blocks / frame</entry><entry align="right">1</entry><entry align="right">5</entry><entry align="right">1</entry></row><row><entry>Number of Bits/transmission interval</entry><entry align="right">320</entry><entry align="right">1600</entry><entry align="right">80</entry></row><row><entry>Add CRC</entry><entry align="right">336</entry><entry align="right">1680</entry><entry align="right">96</entry></row><row><entry>Coding Interval (ms)</entry><entry align="right">20</entry><entry align="right">10</entry><entry align="right">10</entry></row><row><entry>Number of bits/coding interval</entry><entry align="right">336</entry><entry align="right">1680</entry><entry align="right">96</entry></row><row><entry>Add Tail bits / Trellis Termination</entry><entry align="right">344</entry><entry align="right">1692</entry><entry align="right">104</entry></row><row><entry>Convolutional Encoding</entry><entry align="right">1032</entry><entry align="right" /><entry align="right">312</entry></row><row><entry>Turbo Encoding</entry><entry align="right" /><entry align="right">5076</entry><entry align="right" /></row><row><entry>Rate Match factor, dB (-20%= -1dB)</entry><entry align="right">3.01</entry><entry align="right">0.00</entry><entry align="right">3.98</entry></row><row><entry>Rate Matching / Trans. Interval</entry><entry align="right">1032</entry><entry align="right">0</entry><entry align="right">468</entry></row><row><entry>Add DTX indication / Trans. Interval</entry><entry align="right">2096</entry><entry align="right">12</entry><entry align="right">1780</entry></row><row><entry>First interleaver size</entry><entry align="right">4160</entry><entry align="right">5088</entry><entry align="right">2560</entry></row><row><entry>Interleaver check mod(1st IL/ 32)</entry><entry align="right">0</entry><entry align="right">0</entry><entry align="right">0</entry></row><row><entry>second interleaver duration (ms)</entry><entry colsep="0" align="center">10</entry><entry namest="col3" nameend="col4" align="left" /></row><row><entry>second interleaver size</entry><entry colsep="0" align="center">9728</entry><entry namest="col3" nameend="col4" align="left" /></row><row><entry>number of physical channels</entry><entry colsep="0" align="center">1</entry><entry namest="col3" nameend="col4" align="left" /></row><row><entry>Pilot bits /slot</entry><entry colsep="0" align="center">16</entry><entry namest="col3" nameend="col4" align="left" /></row><row><entry>TFCI bits per slot</entry><entry colsep="0" align="center">8</entry><entry namest="col3" nameend="col4" align="left" /></row><row><entry>TPC bits / slot</entry><entry colsep="0" align="center">8</entry><entry namest="col3" nameend="col4" align="left" /></row><row><entry>Total DPCCH field / frame</entry><entry colsep="0" align="center">512</entry><entry namest="col3" nameend="col4" align="left" /></row><row><entry>Total bits / physical channel frame</entry><entry colsep="0" align="center">10240</entry><entry namest="col3" nameend="col4" align="left" /></row><row><entry>Physical Channel rate (ksps)</entry><entry colsep="0" align="center">512</entry><entry namest="col3" nameend="col4" align="left" /></row></tbody></tgroup></table></tables>

0035In the uplink a spreading factor of 4 was chosen resulting in a CCTrCH rate of 1024 ksps. The services are defined by: <ul id="ul0003" list-style="none" compact="compact"><li>144 kbps low rate video, using turbo coding, requiring E_S/N_O = 4 dB.</li><li>13 kbps voice, using convolutional coding, requiring E_S/N_O = 7 dB.</li><li>Signalling information, using convolutional coding, requiring E_S/N_O = 8 dB.</li><li>If E'_S/N_O = 4 dB, then the rate matching is required by each service to achieve the desired (E_S/N_O)_i is:</li><li>video: 0 dB voice: 3 dB signalling: 4 dB.</li></ul> This is shown in table 2, under the rate matching row. Additionally, dynamic rate matching is used to match T_frame and T_total resulting is a further gain due to repetition of 1.7 dB. Hence E"_S/N_O = E'_S/N_O - 1.7 = 2.3 dB. The number will vary slightly in tables 1 and 2 in order to achieve number of bits that agree with realistic interleaver size; interleaver sizes must be divisible by the number of slots. <tables id="tabl0002" num="0002"><img file="EP1069798B1_D0009.tif" /></tables>

Second Method

0036In an alternative approach to the channel coding and rate matching required to multiplex several services, the arrangement shown in <figref idref="f0003">Figure 5</figref> is used; in a small variation to <figref idref="f0003">Figure 4</figref>, the inputs at Step 1 in the BSC 18 are values of E_b/N_O for each service and the number of bits in one time frame, N_biti; the current data rate is R_B = N_biti · 10 ms.

0037(E_B/N_o)_i indicates a QoS of service i which is handled by the BSC 18. After the channel coding step, which is specific to the service i and is described by the coding factor (coding gain) CF_i, when the number of coded bits N_codi = N_biti . CF_i

0038This value is the input to a rate matching step, the output of which is (E_s/N_O)_i, the QoS after the coding and rate matching, and the number of symbols N_Si = N_codi·RF_i

0039This second method depends on the assumption <maths id="math0009" num="(4)"><math display="block"><mrow><mo>(</mo><mfrac><msub><mi>E</mi><mi>B</mi></msub><msub><mi>N</mi><mn>0</mn></msub></mfrac><mo>⁢</mo><msub><mrow><mo>)</mo></mrow><mi>i</mi></msub><mo>⋅</mo><msub><mi>N</mi><mi mathvariant="italic">bilt</mi></msub><mo>=</mo><mrow><mo>(</mo><mfrac><msub><mi>E</mi><mi mathvariant="italic">cod</mi></msub><msub><mi>N</mi><mn>0</mn></msub></mfrac><mo>⁢</mo><msub><mrow><mo>)</mo></mrow><mi>i</mi></msub><mo>⋅</mo><msub><mi>N</mi><mi mathvariant="italic">codi</mi></msub></mrow><mo>=</mo></mrow><mo>(</mo><mfrac><msub><mi>E</mi><mi>S</mi></msub><msub><mi>N</mi><mn>0</mn></msub></mfrac><mo>⁢</mo><msub><mrow><mo>)</mo></mrow><mi>i</mi></msub><mo>⋅</mo><msub><mi>N</mi><mi mathvariant="italic">Si</mi></msub></math><img file="EP1069798B1_D0010.tif" /></maths>

0040Eq. (4) may be rewritten using the coding factor CF_i and the rate matching factor RF_i(cf. eq.(5)). The number of repeated / punctured bits is simply given by N_RMi = N_Si<ul id="ul0004" list-style="dash" compact="compact"><li>N_codi = (RF_i-1)· N_codi. Depending on the rate matching factor RF_i there are three possibilities: <ul id="ul0005" list-style="bullet" compact="compact"><li>Repetition: RF_i > 1, i.e. N_Si > N_codi, N_RMi > 0</li><li>Puncturing: RF_i < 1, i.e. N_Si < N_codi, N_RMi < 0</li><li>No rate matching: RF_i = 1, i.e. N_Si = N_codi N_RMi = 0</li></ul> The linear approximation in eq. (4) is only valid if the puncturing does not exceed a certain ratio. Currently, in ETSI there is the working assumption, that puncturing should not exceed 20%, i.e. DRF_i ≥ 0.8! The rate matching factor RF_i is given by the product of the dynamic rate matching factor DRF and the semi static rate matching factor SRF_i by RF_i = DRF · SFR_i. The dynamic rate matching factor DRF is equal for all multiplexed services, and the semi static rate matching factor SRF_i depends on the service. Two basic conditions must be fulfilled by the rate matching principle: <maths id="math0010" num="(5)"><math display="block"><mrow><mo>(</mo><mfrac><msub><mi>E</mi><mi>S</mi></msub><msub><mi>N</mi><mn>0</mn></msub></mfrac><mo>⁢</mo><msub><mrow><mo>)</mo></mrow><mi>i</mi></msub><mo>=</mo><mrow><mo>(</mo><mfrac><msub><mi>E</mi><mi mathvariant="italic">cod</mi></msub><msub><mi>N</mi><mn>0</mn></msub></mfrac><mo>⁢</mo><msub><mrow><mo>)</mo></mrow><mi>i</mi></msub><mo>⋅</mo><mfrac><mn>1</mn><msub><mi mathvariant="italic">RF</mi><mi>i</mi></msub></mfrac><mo>=</mo></mrow></mrow><mo>(</mo><mfrac><msub><mi>E</mi><mi>B</mi></msub><msub><mi>N</mi><mn>0</mn></msub></mfrac><mo>⁢</mo><msub><mrow><mo>)</mo></mrow><mi>i</mi></msub><mo>⋅</mo><mfrac><mn>1</mn><mrow><mi mathvariant="italic">DRF</mi><mo mathvariant="italic">⋅</mo><msub><mi mathvariant="italic">SRF</mi><mi mathvariant="italic">i</mi></msub><mo mathvariant="italic">⋅</mo><msub><mi mathvariant="italic">CF</mi><mi mathvariant="italic">i</mi></msub></mrow></mfrac><mo>≈</mo><mi mathvariant="italic">const</mi><mn>.</mn><mo>∀</mo><mi>i</mi><mo>=</mo><mn>1</mn><mo>,</mo><mo>…</mo><mo>,</mo><mi>n</mi></math><img file="EP1069798B1_D0011.tif" /></maths><maths id="math0011" num="(6)"><math display="block"><mstyle displaystyle="true"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover></mstyle><msub><mi>N</mi><mi mathvariant="italic">Si</mi></msub><mo>=</mo><mstyle displaystyle="true"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover></mstyle><msub><mi>N</mi><mi mathvariant="italic">codi</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">RF</mi><mi>i</mi></msub><mo>=</mo><mi mathvariant="italic">DRF</mi><mo>⋅</mo><mstyle displaystyle="true"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover></mstyle><msub><mi>N</mi><mi mathvariant="italic">codi</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">SFR</mi><mi mathvariant="italic">i</mi></msub><mo mathvariant="italic">=</mo><msub><mi mathvariant="italic">N</mi><mi mathvariant="italic">S</mi></msub></math><img file="EP1069798B1_D0012.tif" /></maths></li></ul>

0041The number of symbol bits in one time frame (10 ms) is derived by <maths id="math0012"><math display="inline"><msub><mi>N</mi><mi>S</mi></msub><mo>=</mo><mn>16</mn><mo>⋅</mo><mstyle displaystyle="false"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>m</mi></munderover></mstyle><msub><mi>N</mi><mi mathvariant="italic">dataj</mi></msub></math><img file="EP1069798B1_D0013.tif" /></maths> The value of m > 1 is used for multi-code transmission, i.e. more than one DPDCH is assigned to one CCTrCH. Single code transmission is described by m = 1. In the uplink the number of data bits N_data of one DPDCH in a time slot (0.625 ms) is simply given by N_data = 10·2^k, k = 0...6. In the downlink, due to time multiplex of DPDCH and DPCCH it is given by N_data = 20·2^k-N_pilot N_TPC-N_TFCI, k = 0...6. The spreading factor of the related DPCH in uplink and downlink is SF = 256/2^k, thus it may range from 256 down to 4.

0042With this approach and referring once more to <figref idref="f0003">Figure 4</figref>, after the slightly different inputs to Step 1, as explained above, Step 2 is different in that the rate matching factors are calculated as follows:- <dl id="dl0002"><dt><b>Step 2:</b></dt><dd>First, the following transformation is performed (cf. eq.4): (E_cod/N₀)_i = (E_B/N₀)_i /CF_i. Then a service i=i0 with certain E_codN₀)_i0 requirement is chosen as reference, which has a semi static rate matching factor SRF_i0 = 1. From this all semi static rate matching factors for the remaining services can be derived by SRF_i = (E_cod/N₀)_i0 / (E_cod/N₀)_i, i # i0. Such factors do not depend on the currently used data rates, they are a function of E_B/N₀-requirement and the coding method. Thus, they are part of the semi-static part of the TFCI. Next, each service combination is taken into account. The spreading factor SF which is explicitly given by the number of symbol bits N_S must be chosen for each combination of data bits N_biti according to <maths id="math0013"><math display="block"><mstyle displaystyle="false"><munderover><mo mathvariant="italic">∑</mo><mrow><mi mathvariant="italic">i</mi><mo mathvariant="italic">=</mo><mn>1</mn></mrow><mi mathvariant="italic">n</mi></munderover></mstyle><msub><mi mathvariant="italic">N</mi><mi mathvariant="italic">codi</mi></msub><mo mathvariant="italic">⋅</mo><msub><mi mathvariant="italic">SRF</mi><mi mathvariant="italic">i</mi></msub><mo mathvariant="italic">=</mo><mstyle displaystyle="false"><munderover><mo mathvariant="italic">∑</mo><mrow><mi mathvariant="italic">i</mi><mo mathvariant="italic">=</mo><mn>1</mn></mrow><mi mathvariant="italic">n</mi></munderover></mstyle><msub><mi mathvariant="italic">N</mi><mi mathvariant="italic">bilt</mi></msub><mo mathvariant="italic">⋅</mo><msub><mi mathvariant="italic">CF</mi><mi mathvariant="italic">i</mi></msub><mo mathvariant="italic">⋅</mo><msub><mi mathvariant="italic">SRF</mi><mi mathvariant="italic">i</mi></msub><mo mathvariant="italic">≤</mo><msub><mi mathvariant="italic">N</mi><mi mathvariant="italic">S</mi></msub><mo mathvariant="italic">/</mo><msub><mi mathvariant="italic">DRF</mi><mi>min</mi></msub><mspace width="1em" /><mrow><mo>(</mo><mi>cf</mi><mn>.</mn><mspace width="1em" /><mi>eq</mi><mn>.</mn><mrow><mo>(</mo><mn>6</mn><mo>)</mo><mo>,</mo><msub><mi>DRF</mi><mi>min</mi></msub><mo>=</mo></mrow></mrow><msub><mi>RF</mi><mi>min</mi></msub><mo>/</mo><msub><mi>SRF</mi><mi>min</mi></msub><mo>,</mo><msub><mi>RF</mi><mi>min</mi></msub><mo>=</mo><mn>0.8</mn><mo>)</mo><mn>.</mn></math><img file="EP1069798B1_D0014.tif" /></maths> Because N_S and hence SF depends on the currently used data rate, the spreading factor SF may be an element of the dynamic part of TFCl.</dd><dt>Step 3</dt><dd>is identical to that performed in the first method according to the invention.</dd><dt><b>Step 4</b></dt><dd>The dynamic rate matching factor is automatically given by <maths id="math0014"><math display="inline"><mi mathvariant="italic">DRF</mi><mo>=</mo><msub><mi>N</mi><mi>s</mi></msub><mo>/</mo><mrow><mo>(</mo><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover><msub><mi>N</mi><mi mathvariant="italic">codi</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub><mo>)</mo><mo>=</mo><msub><mi>N</mi><mi>S</mi></msub><mo>/</mo><mfenced><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover><msub><mi>N</mi><mi mathvariant="italic">hiti</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">CF</mi><mi>i</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub></mfenced></mrow></math><img file="EP1069798B1_D0015.tif" /></maths> for a certain service combination. Note: N_biti is element of the dynamic part of TFCI (transport block set size), N_S is given by SF and CF_i comes from the used coding scheme. Now the rate matching factor for each service is calculated by RF_i = Dry SRF_i. From this, the number of punctured / repeated bits is simply N_RMi = (RF_i-1) · N_codi.</dd><dt>Step 5</dt><dd>is identical to the first method.</dd></dl>

Example 2

0043In the following example of the second method, a combination of 2 services assumed with the following parameters:- <ul id="ul0006" list-style="bullet" compact="compact"><li>DCH₁: LCD data service with E_B/N₀ = 5dB, 1/3 conv. coding + 8/9 Reed Solomon (e.g. video)</li><li>DCH₂: UDD data service with E_B/N₀ = 3dB, 1/3 conv. coding (e.g. packet data)</li></ul> The following three service combinations are permitted: {DPCH₁, DCH₂}={144/32, 64/64, 64/32} kBit/s. In table 3, the result for the semi static rate matching is shown. For illustration the values are given logarithmic in dB. <tables id="tabl0003" num="0003"><table frame="all"><title><b>Table 3:</b> Determination of the semi static rate matching factor</title><tgroup cols="4"><colspec colnum="1" colname="col1" colwidth="24mm" /><colspec colnum="2" colname="col2" colwidth="32mm" /><colspec colnum="3" colname="col3" colwidth="34mm" /><colspec colnum="4" colname="col4" colwidth="35mm" /><thead><row><entry align="center" valign="top" /><entry align="center" valign="top">DCH₁: LDD-service</entry><entry align="center" valign="top">DCH₂: UDD-service</entry><entry align="center" valign="top">Comments</entry></row><row><entry align="center" valign="top">E_B/N₀</entry><entry align="center" valign="top">5 dB</entry><entry align="center" valign="top">3 dB</entry><entry align="center" valign="top" /></row></thead><tbody><row><entry align="center">coding factor</entry><entry align="center">5.2827 dB (27/8)</entry><entry align="center">4.7712 dB (3/1)</entry><entry align="center">CF_dB = 10log₁₀(CF)</entry></row><row><entry align="center">E_cod/N₀</entry><entry align="center">-0.2827 dB</entry><entry align="center">-1.7712 dB</entry><entry align="center">ΔE_cod/N₀ = 1.4885 dB</entry></row><row><entry align="center">SRF_i</entry><entry align="center">1.4088</entry><entry align="center">1</entry><entry align="center">SRF = 10^{(ΔEcod/N0)/10}</entry></row></tbody></tgroup></table></tables>

0044In the current example, the service with the highest requirement on E_cod/N₀ was chosen for reference. It will be evident, that with the proposed rate matching algorithm at least this service needs no semi static rate matching. Table 4 shows the results for the dynamic part of the rate matching for the different assumed service combinations and the calculated number of bits. <tables id="tabl0004" num="0004"><table frame="all"><title><b>Table 4:</b> Dynamic Rate Matching for different Service Combinations</title><tgroup cols="5"><colspec colnum="1" colname="col1" colwidth="18mm" colsep="0" /><colspec colnum="2" colname="col2" colwidth="25mm" /><colspec colnum="3" colname="col3" colwidth="32mm" /><colspec colnum="4" colname="col4" colwidth="32mm" /><colspec colnum="5" colname="col5" colwidth="48mm" /><thead><row><entry valign="top" /><entry valign="top" /><entry align="center" valign="top">DCH₁: LDD-service</entry><entry align="center" valign="top">DCH₂: UDD-service</entry><entry align="center" valign="top">Comments</entry></row></thead><tbody><row><entry>144/32</entry><entry>N_biti</entry><entry align="center">1440</entry><entry align="center">336 (16 bit MAC)</entry><entry align="center">TB set size</entry></row><row><entry /><entry>N_codi</entry><entry align="center">4860</entry><entry align="center">1008</entry><entry rowsep="0" align="center">N_S = 10240, SF = 4</entry></row><row><entry /><entry>N_sixteenths</entry><entry align="center">8926</entry><entry align="center">1314</entry><entry align="center">DRF = 1.3037</entry></row><row><entry /><entry>E_S/N₀</entry><entry align="center">-2.9230</entry><entry align="center">-2.9230</entry><entry morerows="1" align="center">ΔE_S/N₀= 0.222 dB</entry></row><row><entry /><entry>E_S/N₀ w/o RM</entry><entry align="center">-2.7008</entry><entry align="center">-4.1893</entry></row><row><entry>64/64</entry><entry>N_biti</entry><entry align="center">640</entry><entry align="center">772 (32 bit MAC)</entry><entry align="center">TB set size</entry></row><row><entry /><entry>N_codi</entry><entry align="center">2160</entry><entry align="center">2016</entry><entry morerows="1" align="center">N_S = 5120, SF = 8 DRF = 1.0121</entry></row><row><entry /><entry>N_sixteenths</entry><entry align="center">3080</entry><entry align="center">2040</entry></row><row><entry /><entry>E_S/N₀</entry><entry align="center">-1.8234</entry><entry align="center">-1.8234</entry><entry morerows="1" align="center">ΔE_S/N₀ = 0.6556 dB</entry></row><row><entry /><entry>E_S/N₀ w/o RM</entry><entry align="center">-1.1678</entry><entry align="center">-2.6563</entry></row><row><entry>64/32</entry><entry>N_biti</entry><entry align="center">640</entry><entry align="center">336 (16 bit MAC)</entry><entry align="center">TB set size</entry></row><row><entry /><entry>N_codi</entry><entry align="center">2160</entry><entry align="center">1008</entry><entry morerows="1" align="center">N_S=5120, SF=8 DRF = 1.2639</entry></row><row><entry /><entry>N_sixteenths</entry><entry align="center">3846</entry><entry align="center">1274</entry></row><row><entry /><entry>E_S/N₀</entry><entry align="center">-2.7883</entry><entry align="center">-2.7883</entry><entry morerows="1" align="center">ΔE_S/N₀= 0.4208 dB</entry></row><row><entry /><entry>E_S/N₀ w/o RM</entry><entry align="center">-2.3675</entry><entry align="center">-3.8560</entry></row></tbody></tgroup></table></tables>

0045The number of data bits for the LDD service is simply given by its data rate i.e. N_bitLDD = R_bitLDD 10ms. For the UDD service an additional MAC header for e.g. adressing purposes of 16 bits is assumed, i.e. N_bitUDD = R_bitUDD · 10ms + 16 bit. In the example, this bit numbers are equal to the transport block set size. Because the uplink is considered, the number of available symbol bits is simply given by N_S = 40960/SF. The value of ΔE_S/N₀ describes the improvement in E_S/N₀ by use of the proposed rate matching method compared to missing it. The description of the specified service combinations is as follows: <ul id="ul0007" list-style="bullet" compact="compact"><li><u style="single">LDD144/UDD32</u>: The sum of all coded data bits including semi static rate matching is <maths id="math0015"><math display="inline"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover><msub><mi>N</mi><mi mathvariant="italic">codi</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub><mo>=</mo><mn>7855</mn><mo>,</mo><msub><mi>DRF</mi><mi>min</mi></msub><mo>=</mo><mn>0.8</mn><mfenced><msub><mi>SRF</mi><mi>min</mi></msub><mo>=</mo><mn>1</mn></mfenced><mn>.</mn></math><img file="EP1069798B1_D0016.tif" /></maths> The next lowest symbol rate would be N_S = 5120, i.e. SF = 8. According to eq. 6 a DRF &lt; 0.8 would be necessary. Thus, the next highest SF = 4 is chosen with N_S = 10240. The other results can be found in table 4. Using this constellation an improvement in E_S/N₀ of ΔE_S/N₀ = 0.222 dB can be achieved. Since the transmission power is directly related to E_S/N₀ this leads to a decrease of the transmission power of 0.222 dB, too.</li><li><u style="single">LDD64/UDD64</u>: Here, <maths id="math0016"><math display="inline"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover><msub><mi>N</mi><mi mathvariant="italic">codi</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub><mo>=</mo><mn>5059</mn><mn>.</mn></math><img file="EP1069798B1_D0017.tif" /></maths> Thus, SF = 8 is chosen with N_S = 5120. For this service combination, an improvement of ΔE_S/N₀ = 0.6556 dB can be achieved.</li><li><u style="single">LDD64/UDD32</u>: For the last examined service combination <maths id="math0017"><math display="inline"><munderover><mo>∑</mo><mrow><mi>i</mi><mo>=</mo><mn>1</mn></mrow><mi>n</mi></munderover><msub><mi>N</mi><mi mathvariant="italic">coch</mi></msub><mo>⋅</mo><msub><mi mathvariant="italic">SRF</mi><mi>i</mi></msub><mo>=</mo><mn>4051</mn><mn>.</mn></math><img file="EP1069798B1_D0018.tif" /></maths> As in the previous case SF = 8 is chosen with N_s = 5120. Now the improvement is ΔE_S/N₀ = 0.4208 dB.</li></ul>

0046The second example has many of the advantages of the first method according to the invention.