Sequentially decoded low density parity coding (LDPC) forward error correction (FEC) in orthogonal frequency division modulation (OFDM) systems
View Patent ↗A method is provided for implementing a coding and adaptive modulation scheme for application to a point-to-point orthogonal frequency division multiplexed radio communications link. The method determines estimates of the likelihood of the two least significant bits of the digital representation of an input amplitude being a predetermined logical level. The likelihood estimates are input to a Forward Error Correction decoder, which produces a decision as to the state of the bits taking into account previous samples. Once the decision has been made, it is known what the ideal input amplitude would have been in the absence of noise and distortion, on the assumption that the decision was correct, and hence the contribution due to noise and distortion can be estimated. This knowledge can be used to assist the decoding of the most significant bits.
1. A method of multilevel coding, for use in communicating at least one signal over at least one communications channel, the method comprising the steps of:
generating a digital representation of a first amplitude of said signal, said digital representation of said first amplitude of said signal including a plurality of least significant bits and a plurality of most significant bits;
determining at least one likelihood estimate of at least two of said plurality of least significant bits being a predetermined logical level;
decoding said at least one likelihood estimate to determine a state of said at least two least significant bits; and
determining the plurality of most significant bits using the state of the at least two least significant bits and the first amplitude of said signal.
2. The method of claim 1 wherein the plurality of least significant bits are coded as a Gray code.
3. The method of claim 1 wherein determining at least one likelihood estimate comprises passing the digital representation of the first amplitude to a look up table to determine the at least one likelihood estimate of at least two of said plurality of least significant bits being a predetermined logical level.
4. The method of claim 3 wherein two of said plurality of least significant bits are bit 0 and bit 1 and the at least one likelihood estimate represents a likelihood of bit 0 and bit 1 being a logical 1.
5. The method of claim 4 wherein the likelihood estimate is represented by a logarithm of a ratio of a probability of a logical 1 divided by a probability of a logical 0.
6. The method of claim 3 wherein determining the plurality of most significant bits comprises using the at least two least significant bits and the first amplitude of said signal as indices into a look up table containing the plurality of most significant bits for each amplitude two least significant bits combination.
7. A method of multilevel coding, for use in communicating at least one signal over at least one communications channel, the method comprising the steps of:
generating a digital representation of a first amplitude of said signal, said digital representation of said first amplitude of said signal including a plurality of least significant bits and a plurality of most significant bits;
determining at least one likelihood estimate of at least two of said plurality of least significant bits being a predetermined logical level;
decoding said at least one likelihood estimate to determine a state of said at least two least significant bits;
determining one of the plurality of most significant bits using the at least two least significant bits and the first amplitude of said signal; and
determining whether a pulse amplitude modulation mode can be changed.
8. The method of claim 7 wherein determining whether a pulse amplitude modulation can be changed comprises:
determining whether a flag is set, the flag being set indicating that a vector error for reliable operation in a different pulse amplitude modulation mode has been exceeded; and
when the flag is not set for a predetermined amount of samples of the input signal, determining that the pulse amplitude modulation mode can be changed.