Patent Application
20060239457
1. Field of the Invention
The present invention relates to interpretation of data before a complete data block is received and a cyclic redundancy check (CRC) is available. More specifically, it relates to detection and selection of a scrambled control channel from among a plurality of control channels in a transmission before the complete data block is received and error checked.
2. Background of Related Art
Terminal equipment (handsets) designed to be compatible with the HSDPA extensions in release 5 of the Third Generation Partnership Project (“3GPP”) UMTS standards must be able to receive up to four High-Speed Downlink Packet Access (HSDPA) control channels (called High-Speed Shared Control Channels (HS-SCCHs)) and select zero or one for further processing. (The 3GPP standard is available at www.3gpp.org) The selection of one of the received HS-SCCHs is not trivial because no embedded identification (“ID”) or cyclic redundancy check (“CRC”) is available at the time the selection must take place. The invention provides a method for making this selection.
This is a relatively new problem since it appears only since release 5 of the 3GPP standard.
During the course of standardization, numerous submissions were made to standards body (i.e., 3GPP) meetings. Most notably, the performance of different metrics was discussed and compared in submissions from Motorola [1] and Lucent [2][3] (references cited below). Further analysis was provided in the academic literature [4][5]. The methods proposed included the Yamamoto-Itoh metric and the squared Euclidean distance metric directly from the Viterbi decoder.
This problem also shares similarities with blind transport format detection (BTFD) for which there is some prior art [6][7][8].
[1] 3GPP TSG-RAN WG1#24, R1-02-0610, “Performance of the HS-SCCH”, Motorola, April 2002.
[2] 3GPP TSG-RAN WG1#25, R1-02-0553, “Way forward on HS-SCCH coding”, Lucent Technologies, April 2002.
[3] 3GPP TSG-RAN WG1#25, R1-02-0649, “Performance of the HS-SCCH”, Lucent Technologies, April 2002.
[4] Ghosh, Ratasuk, Frank, Love, Stewart, Buckley [Motorola], “Control Channel Design for High Speed Downlink Shared Channel for 3GPP W-CDMA, Rel-5”, Vehicular Technology Conference 2003 (VTC '03), April 2003, Jeju Korea.
[5] Das, Gopalakrishnan, Hu, Khan, Rudrapatna, Sampath, Su, Tatesh, Zhang [Lucent], “Evolution of UMTS Toward High-Speed Downlink Packet Access”, Bell Labs Technical Journal, Vol. 7, Issue 3, pp 47-68, 2003.
[6] Berns, F.; Kreiselmaier, G.; When, N., “Channel decoder architecture for 3G mobile wireless terminals”, Design, Automation and Test in Europe Conference and Exhibition, 2004. Proceedings, pp 192-197, 16-20 February 2004.
[7] Ahmed, W. K. M.; Balachandran, K., “Methods for estimation of the uncoded symbol error rate at the receiver”, Global Telecommunications Conference, 2002. GLOBECOM '02. IEEE, pp 1334-1338, 17-21 November 2002.
[8] Raghavan, A. R.; Baum, C. W., “A reliability output Viterbi algorithm with applications to hybrid ARQ”, IEEE Trans. Information Theory, Vol. 44, No. 3, pp 1214-1216, May 1998
The recommendation of the standards submissions [1]-[5] concluded that the Yamamoto-Itoh or Squared Euclidean distance metrics should be used to select the HS-SCCH for further processing. The Yamamoto-ltoh metric involves substantial changes to the internals of the Viterbi decoder. Consequently, a generic Viterbi decoder design must be modified and the complexity increased. Using the squared Euclidean metric of the maximum likelihood path through the trellis (the final zero state metric at the conclusion of Viterbi decoding for a properly terminated codeblock), was found to give inferior performance.
Much of the prior art for BTFD is limited by the inability to independently discriminate between correctly and incorrectly decoded data, relying upon analysis of a CRC in the data block. The CRC in the HS-SCCH is not available early enough to be useful [6] [7]. Some methods [6][8] also need accurate channel knowledge, which is unlikely to be available in this case.
There is a need for methods and apparatus for better detecting and selecting one (or no) HS-SCCH from a transmission containing a plurality of HS-SCCHs, before the data block is completely received.
In accordance with the principles of the present invention, apparatus to select one of a plurality of scrambled data channels before receipt of an entire data block comprises a decode module to decode an input symbol from one of the plurality of scrambled data channels. A re-encode module re-encodes decoded symbols output from the decode module. A comparison module compares aspects of a received symbol with a corresponding aspect of the decoded/re-encoded symbol and accumulates those aspects. A decision module selects a best one of the plurality of scrambled data channels based on an accumulated output of the comparison module.
In accordance with another aspect of the present invention, a method of computing a metric to determine selection of a scrambled data channel before receipt of an entire data block comprises descrambling and decoding an initial portion of a block of data received in one of a plurality of data channels. The initial portion of the block of data is re-encoded. A value is computed by accumulating the magnitudes of mismatched data samples based on a comparison of decoded/re-encoded data and the corresponding received and descrambled data. A best one of the plurality of data channels is selected based on the computed value being better than the other channels and possibly also beyond a given threshold value.
Features and advantages of the present invention will become apparent to those skilled in the art from the following description with reference to the drawings, in which:
The inventive technique allows similar or superior performance to methods described previously, with lower complexity, and without the need for accurate channel parameters or separate CRC checks.
Terminal equipment (handsets) designed to be compatible with the HSDPA extensions in release 5 of the 3GPP UMTS standards must be able to receive up to 4 HSDPA control channels (called HS-SCCHs) and select zero or one for further processing. The selection of one of the received HS-SCCHs is not trivial because no embedded ID or CRC is available at the time the selection must take place. The invention provides a method for making this selection.
According to release 5 of the 3GPP UMTS standards, user equipment (UE) must be able to receive up to 4 HSDPA control channels (called HS-SCCHs). Unless significant buffering is permitted, the UE must be able to select zero or one HS-SCCH channel for further processing by observing only Part 1 of the HS-SCCH transmission. The selection of one of the received HS-SCCHs is not trivial because no embedded ID or CRC is available at the time the selection must take place. (A CRC is available at a later time to verify correct receipt of the HS-SCCH, however this is not relevant information in the current instance because the HS-SCCH contains information necessary to configure the receiver, which starts receiving the main HSDPA data channel(s) before the CRC is transmitted. Therefore the correct HS-SCCH must be selected before the CRC is available. Thus the CRC cannot be used as a criterion in the selection process.)
HS-SCCH data is scrambled before transmission, the scrambling mask being dependant upon the destination handset's ID number, and is broadcast to multiple handsets within the cell. A handset will receive up to 4 scrambled HS-SCCHs in a transmission time interval (TTI), however at most one (and possibly zero) are intended for any one handset. The handset descrambles the received data (according to its own ID) and then decodes the (up to 4) blocks. Only correctly descrambled data can be correctly received, however due to noise in the transmission channel it is difficult to determine which block, if any, is intended for the handset in question.
A new metric is proposed and used to facilitate the HS-SCCH selection process. It is important that the metric is reliable for short block lengths (for the 3GPP HSDPA application, the block length is just 8 bits).
In accordance with the principles of the present invention, for part 1 of each received HS_SCCH channel:
Step 1.1: Descramble received data (according to own ID)
Step 1.2: Decode received data
Step 1.3: Re-encode decoder output
Step 1.4: Compute the following metric:
where:
yi=0 when sign(received_sample)=sign(re-encoded_bit);
yj=absolute_value(received_sample) otherwise.
Step 2: Select the best HS-SCCH channel, i.e., that with the minimum metric from Step 1.4.
Once an HS-SCCH is selected, it is necessary to decide whether to continue processing the channel (it corresponds to the UE in question), or to discard the transmission entirely (no HS-SCCH intended for the UE in question). A threshold can be applied to the metric to aid this decision. In a given embodiment, such threshold may be determined empirically.
In particular, as shown in
The decode block 102 shown in the disclosed embodiment shows an additional function of de-puncturing. Puncturing is a technique that allows the encoding and decoding of higher rate codes using standard fixed rate (e.g. rate 1/3) encoders and decoders. A Puncture block removes bits from an encoded bit stream, thereby increasing the code rate. While shown with de-puncturing, the principles of the present invention are equally applicable to embodiments not requiring and/or including puncturing/de-puncturing.
The decoded bits output from the decode block 102 may be utilized by the relevant equipment (e.g., mobile phone) in any otherwise conventional manner. However, in accordance with the principles of the present invention, the output of the decode block 102 is also fed into a Re-encode block 103. The re-encode block 103 in the disclosed embodiments includes a puncture function complementary to the de-puncture function present in the decode block 102.
A time delayed version of the received samples is fed into the metric calculation blocks, 104, 106, 110, 108, 112. Blocks 104, 106 are intended to determine if the decoded and re-encoded samples output from the re-encode block 103 is of the same sign as the original symbol input to the decode block 102. Blocks 108, 110, 112 are intended to compute the value of the metric, the accumulation step 112 only taking place when the sign of the re-encoded sample does not match the sign of the received sample.
If the decoded data channel (e.g., a relevant one of the HS-SCCHs) is intended for the particular mobile device, the received samples will be largely the same as the decoded and re-encoded samples, and thus the sign of both samples will be the same. In such case, the accumulator 112 will accumulate a small value. If the decoded data channel is not intended for the particular mobile device, most likely more input samples will be of different sign, accumulating a larger value in the accumulator 112.
In the disclosed embodiment, the metric calculation logic block includes a hard decision generator 104, an XNOR gate 106, an absolute value block 108, a mulitplexer 110, and an accumulator 112, though other particular logic devices may be implemented within the principles of the present invention.
There are numerous other ways the inventive metric can be computed. For example, through generation of the code words and metric accumulation during the trace-back phase of the Viterbi decoding. This alternate means of calculation removes the need for a separate re-encode block, and may lead to lower latency; an important design consideration due to the need to immediately decode Part 2 of the HS-SCCH once the correct HS-SCCH channel has been selected (if any).
In particular, as shown in step 202 of
In step 204, input symbols are decoded per the decryption relevant to the particular mobile device. If puncture techniques are utilized, input symbols may additionally be de-punctured in step 204.
In step 206, the decoder output is re-encoded per the encryption associated with the relevant mobile device. If puncture techniques are utilized, coded symbols may additionally be punctured in step 206.
In step 208, the decoded and re-encoded symbol is compared to the originally descrambled received data. In the disclosed embodiments, the sign of the symbols are compared. If they are different, it is an indication that the relevant control channel may not be the proper channel. But, due to a noisy environment, a single symbol comparison may not be dispositive. The more symbols utilized for comparison, the more the environmental conditions are removed from the final selection decision.
In step 210, if the signs of the compared symbols are different, their magnitudes are accumulated.
In step 212, a decision is made as to which HS-SCCH is intended to be received by the relevant mobile device. Block 212 selects the HS-SCCH with the minimum metric, but no HS-SCCH is selected in the case that all metrics exceed or otherwise go beyond a given threshold value (step 214) and all control channels are ignored by the mobile device for the current time period. The value of the threshold may be determined empirically.
Ideally, all available control channels are tested simultaneously, to allow selection of a control channel in the shortest possible time. In this way, for a system including 4 HS-SCCHs, for example, the circuit of
Performance Results
In particular, as shown in
To determine metrics which provide the most reliable information, five (5) different decoder metrics were implemented. These six decoder metrics are defined as follows:
Metric 1: Count number of sign differences between re-encoded output and received data.
Metric 2: Sum magnitude of those received samples that have a different sign to the re-encoded data.
Metric 3: Sum squared magnitudes of those received samples that have a different sign to the re-encoded data.
Metric 4: Sum received samples multiplied by sign of the re-encoded data.
Metric 5: Squared Euclidean distance metric (Raw value of zero state metric at end of trellis).
Performance of each of the above metrics was examined to determine how reliable a possible discrimination would be between a correct HS-SCCH and an HS-SCCH belonging to another user in noisy conditions.
In particular, the inventors found, metric 2 to be the most reliable at differentiating between the correct HS-SCCH and an HS-SCCH belonging to another user.
In particular, as shown in
The effect on overall system performance of making a wrong selection can be seen in
Furthermore, metric 2 provides a mechanism to estimate when no HS-SCCH intended for the current UE, despite there being no CRC available. A threshold can be used that allows the determination of cases where no received HS-SCCH is correctly decoded or intended for the current UE and therefore that no HS-SCCH should be selected.
Using a combination of the two above, selection is made for the HS-SCCH block with the minimum metric, provided that metric is below a pre-determined threshold (for example, a threshold of 8 would be useful in the channel conditions simulated (Refer to
The present invention is very simple to implement in a small digital circuit, providing great advantages over prior art systems. The present invention does not require advance knowledge about the channel (e.g., it doesn't require an estimate of the noise variance). Rather, selection of the block with the minimum metric corresponds to the best block with high reliability. Use of a metric threshold may assist deciding if any HS-SCCH is for the intended UE.
The present invention is particularly applicable to any 3GPP release 5 handset, and may be embedded in an ASIC or firmware, e.g., in a handset's firmware.
While the invention has been described with reference to the exemplary embodiments thereof, those skilled in the art will be able to make various modifications to the described embodiments of the invention without departing from the true spirit and scope of the invention.
