Systems and methods for rate matching via a heterogeneous kernel when using general polar codes
Systems and methods are disclosed for performing rate matching when using general polar codes. In one embodiment, a method of generating a codeword includes receiving bits at a polar encoder and encoding the bits using polar encoder kernels. The polar encoder kernels include a first kernel and a second kernel. The first kernel receives a set of input q-ary symbols and modifies the set of input q-ary symbols according to a first kernel generator matrix to produce a set of output q-ary symbols. The second kernel receives a set of input l-ary symbols, where l does not equal q, and modifies the set of input l-ary symbols according to a second kernel generator matrix to produce a set of output l-ary symbols. For example, the first kernel may be a binary kernel and the second kernel may be a Reed-Solomon (RS) based kernel.
1. A method of generating a codeword comprising:
receiving a plurality of bits at a polar encoder, the plurality of bits including a first set of bits and a second set of bits that does not overlap with the first set of bits;
encoding the plurality of bits using a plurality of polar encoder kernels, wherein the plurality of polar encoder kernels used is based on at least one of a coding rate R and a transmitted bit length M, and wherein the plurality of polar encoder kernels includes:
a binary kernel that receives a set of input binary symbols represented by the first set of bits, and that modifies the set of input binary symbols according to a first kernel generator matrix to produce a set of output binary symbols represented by first bits; and
a Reed-Solomon (RS) based kernel, different from the binary kernel, and that receives a set of input l-ary symbols represented by the second set of bits, wherein l is an integer greater than 2, and that modifies the set of input l-ary symbols according to a second kernel generator matrix to produce a set of output l-ary symbols represented by second bits.
2. The method of claim 1 , wherein l=4 and the second kernel is an RS(4) kernel.
3. The method of claim 2 wherein the plurality of polar encoder kernels includes two encoding layers of binary kernels and two encoding layers of RS(4) kernels.
4. The method of claim 3 , wherein the codeword has a bit length N b equal to 128 bits.
5. The method of claim 1 , wherein l=8 and the second kernel is an RS(8) kernel.
6. The method of claim 1 , wherein the codeword has bit length N b , and N b =M.
7. The method of claim 1 , wherein the codeword has bit length N b , and N b >M, and the method further comprising:
puncturing the codeword to result in the transmitted bit length M.
8. The method of claim 7 , wherein puncturing the codeword comprises shortening the codeword.
9. The method of claim 1 , wherein the codeword is output from the polar encoder and is a first codeword that has a first bit length N b1 , and wherein the method further comprises, subsequent to generating the first codeword:
generating a second codeword output from the polar encoder, the second codeword having a second bit length N b2 that is different from the first bit length N b1 , by encoding another plurality of bits using a different plurality of polar encoder kernels.
10. A transmitter comprising:
a polar encoder having a plurality of polar encoder kernels to encode a plurality of bits received at the polar encoder and generate a codeword, the plurality of bits including a first set of bits and a second set of bits that does not overlap with the first set of bits;
wherein the plurality of polar encoder kernels to be used is based on at least one of a coding rate R and a transmitted bit length M, and wherein the plurality of polar encoder kernels includes:
a binary kernel to receive a set of input binary symbols represented by the first set of bits, to modify the set of input binary symbols according to a first kernel generator matrix to produce a set of output binary symbols represented by first bits; and
a Reed-Solomon (RS) based kernel different from the binary kernel, the Reed-Solomon (RS) based kernel to receive a set of input l-ary symbols represented by the second set of bits, wherein l is an integer greater than 2, to modify the set of input l-ary symbols according to a second kernel generator matrix to produce a set of output l-ary symbols represented by second bits.
11. The transmitter of claim 10 , wherein l=4 and the second kernel is an RS(4) kernel.
12. The transmitter of claim 11 wherein the plurality of polar encoder kernels includes two encoding layers of binary kernels and two encoding layers of RS(4) kernels.
13. The transmitter of claim 12 , wherein the codeword has a bit length N b equal to 128 bits.
14. The transmitter of claim 10 , wherein l=8 and the second kernel is an RS(8) kernel.
15. The transmitter of claim 10 , wherein the codeword has bit length N b and N b =M.
16. The transmitter of claim 10 , wherein the codeword has bit length N b and N b >M, and the transmitter further includes a puncturer to puncture the codeword to result in the transmitted bit length M.
17. The transmitter of claim 16 , wherein the puncturer is to puncture the codeword by shortening the codeword.
18. The transmitter of claim 10 , wherein the codeword is a first codeword output from the polar encoder that has a first bit length N b1 , and wherein the polar encoder is further to, subsequent to generating the first codeword:
generate a second codeword as an output of the polar encoder, the second codeword having a second bit length N b2 that is different from the first bit length N b1 , by encoding another plurality of bits using a different plurality of polar encoder kernels.