Compression based on channel characteristics
View Patent ↗In a packet based data transmission including automatic repeat request (ARQ) protocols, the memory consumption of the ARQ protocol is reduced by compression/decompression of failed data block prior to combination with the retransmitted data blocks using low complexity compression/decompression algorithms. The compression algorithm includes of two parts: calculating and storing a scale factor that estimates the soft values in the data block, and storing the each soft values' sign in local memory instead of the complete soft value. Since one bit is sufficient to store the sign of each soft value, the memory consumption of a data block containing N soft values is N bits. Note, the scale factor increases the memory consumption, in bits, with the word length of the scale factor. However, the relative memory increase becomes smaller the larger the data block.
1. A method of reducing the memory consumption of a retransmission scheme, the method comprising:
transmitting information to a receiver in the form of data blocks using packet-based transmission;
decoding the received data blocks;
compressing failed data blocks, wherein a failed data block is a data block that fails the decoding process;
storing the compressed failed data blocks;
decompressing the failed data blocks;
combining a decompressed failed data block with an associated retransmitted data block; and
decoding the combined data block.
2. The method according to claim 1 , wherein compressing failed data blocks comprises:
calculating a scale factor that represents soft values within the failed data block; and
storing the scale factor and a sign of each soft value within the failed data block in memory.
3. The method according to claim 2 , wherein said scale factor is an average of absolute values of the soft values in the failed data block.
4. The method according to claim 2 , wherein decompressing the failed data block comprises: multiplying the scale factor by the stored soft value signs.
5. The method according to claim 2 , wherein said soft values are logarithmically scaled values.
6. The method according to claim 5 , wherein combining a decompressed failed data block with an associated retransmitted data block comprises:
adding the values of the retransmitted data block and the decompressed values of the failed data block.
7. The method according to claim 2 , further comprising:
dividing the data blocks into a plurality of sub-blocks, wherein each sub-block is characterized by a substantially constant signal-to-interference and noise-ratio;
calculating a plurality of scale factors, wherein each scale factor represents soft values within a corresponding sub-block; and
storing the plurality of scale factors and a sign of each soft value within the failed data block in memory.
8. A receiver for receiving packet based data transmissions, the receiver comprising:
a soft output detector;
a decoder;
local memory;
logic that compresses failed data blocks and stores the compressed failed data blocks in said local memory;
logic that decompresses a compressed failed data block stored in said local memory; and
logic that combines said decompressed data block with an associated retransmitted data block.
9. The receiver according to claim 8 , further comprising:
logic that calculates a scale factor that represents soft values within the failed data block; and
logic that stores the scale factor and a sign of each soft value within the failed data block in memory.
10. The receiver according to claim 9 , wherein said scale factor is an average of the absolute values of the soft values in the failed data block.
11. The receiver according to claim 9 , further comprising:
logic that multiplies the scale factor by the stored soft value signs.
12. The receiver according to claim 9 , wherein said soft values are logarithmically scaled values.
13. The receiver according to claim 12 , wherein the logic that combines the decompressed failed data block with the associated retransmitted data block comprises:
logic that adds the values of the retransmitted data block and the decompressed values of the failed data block.
14. The receiver according to claim 8 , further comprising:
logic that divides the data block into a plurality of sub-blocks, wherein each sub-block contains a constant SINR;
logic that calculates a plurality of scale factors, wherein each scale factor represents soft values within a corresponding sub-block; and
logic that stores the plurality of scale factors and a sign of each soft value within the failed data block in memory.
15. A method of processing data blocks within a receiver of a communications system, comprising:
receiving a data blocks in the receiver;
calculating a scale factor that represents soft values within the received data block;
storing the scale factor and a representation of each soft value within the data block in memory, the representation indicating only a sign of each soft value within the received data block;
generating a decompressed data block by decompressing the scale factor and the representation of each soft value within the data block stored in memory; and
combining the decompressed data block with an associated retransmitted data block.
16. The method according to claim 15 , wherein said soft values are logarithmicaily scaled values.
17. A method of compressing data blocks within a receiver of a communications system, comprising:
receiving a data block in the receiver;
calculating a scale factor that represents soft values within the received data block; and
storing the scale factor and a sign of each soft value within the data block in memory;
dividing the data blocks into a plurality of sub-blocks, wherein each sub-block is characterized by a substantially constant signal-to-interference and noise-ratio;
calculating a plurality of scale factors, wherein each scale factor represents soft values within a corresponding sub-block; and
storing the plurality of scale factors and a sign of each value within the data block in memory.