Decision directed multi-modulus searching algorithm

A method comprising:

• receiving, from a transmitter, an optical signal comprising a plurality of data symbols;
• iteratively determining a set of centroids for demodulating the plurality of data symbols according to a modulation constellation associated with the plurality of data symbols, wherein the iteratively determining comprises, for each data symbol of at least a subset of the plurality of data symbols:
• assigning a centroid of the set of centroids to the each data symbol of at least the subset of the plurality of data symbols;
• and
• updating one or more of the set of centroids based at least in part on the assigning the centroid to the each data symbol of at least the subset of the plurality of data symbols;
• demodulating the plurality of data symbols based at least in part on the updated set of centroids;
• and
• determining initial values for the set of centroids prior to the iteratively determining, wherein determining the initial values is based at least in part on a value of a randomly selected data symbol of the plurality of data symbols.

A method comprising:

• receiving, from a transmitter, an optical signal comprising a plurality of data symbols;
• iteratively determining a set of centroids for demodulating the plurality of data symbols according to a modulation constellation associated with the plurality of data symbols, wherein the iteratively determining comprises, for each data symbol of at least a subset of the plurality of data symbols:
• assigning a centroid of the set of centroids to the each data symbol of at least the subset of the plurality of data symbols;
• and
• updating one or more of the set of centroids based at least in part on the assigning the centroid to the each data symbol of at least the subset of the plurality of data symbols;
• demodulating the plurality of data symbols based at least in part on the updated set of centroids;
• and
• determining distance metrics between a radius of the each data symbol and at least a subset of the set of centroids, wherein the assigning the each data symbol of the at least the subset of the plurality of data symbols to the centroid of the set of centroids is based at least in part on the determining the distance metrics.

A method comprising:

• receiving, from a transmitter, an optical signal comprising a plurality of data symbols;
• iteratively determining a set of centroids for demodulating the plurality of data symbols according to a modulation constellation associated with the plurality of data symbols, wherein the iteratively determining comprises, for each data symbol of at least a subset of the plurality of data symbols:
• assigning a centroid of the set of centroids to the each data symbol of at least the subset of the plurality of data symbols;
• and
• updating one or more of the set of centroids based at least in part on the assigning the centroid to the each data symbol of at least the subset of the plurality of data symbols;
• demodulating the plurality of data symbols based at least in part on the updated set of centroids;
• and
• determining an updated centroid for the one or more of the set of centroids based at least in part on a function of radii of data symbols assigned to the one or more of the set of centroids.

A method comprising:

• receiving, from a transmitter, an optical signal comprising a plurality of data symbols;
• iteratively determining a set of centroids for demodulating the plurality of data symbols according to a modulation constellation associated with the plurality of data symbols, wherein the iteratively determining comprises, for each data symbol of at least a subset of the plurality of data symbols:
• assigning a centroid of the set of centroids to the each data symbol of at least the subset of the plurality of data symbols;
• and
• updating one or more of the set of centroids based at least in part on the assigning the centroid to the each data symbol of at least the subset of the plurality of data symbols;
• demodulating the plurality of data symbols based at least in part on the updated set of centroids;
• and
• determining that a difference between the set of centroids and the updated set of centroids exceeds a threshold;
• and
• transmitting an indication to the transmitter that the difference between the set of centroids and the updated set of centroids exceeds the threshold.

A method comprising:

• receiving, from a transmitter, an optical signal comprising a plurality of data symbols;
• iteratively determining a set of centroids for demodulating the plurality of data symbols according to a modulation constellation associated with the plurality of data symbols, wherein the iteratively determining comprises, for each data symbol of at least a subset of the plurality of data symbols:
• assigning a centroid of the set of centroids to the each data symbol of at least the subset of the plurality of data symbols;
• and
• updating one or more of the set of centroids based at least in part on the assigning the centroid to the each data symbol of at least the subset of the plurality of data symbols;
• demodulating the plurality of data symbols based at least in part on the updated set of centroids;
• and
• assigning one centroid of the updated set of centroids to each data symbol of a second subset of the plurality of data symbols after iteratively determining the set of centroids, wherein the second subset of the plurality of data symbols is different than the subset of the plurality of data symbols.

A method comprising:

• receiving an optical signal comprising a plurality of data symbols;
• performing a training procedure to update an initial set of centroids until a convergence metric associated with an updated set of centroids satisfies a threshold, wherein performing the training procedure comprises, for each data symbol of at least a subset of the plurality of data symbols:
• assigning a centroid of the initial set of centroids to the each data symbol of the at least the subset of the plurality of data symbols;
• updating one or more of the initial set of centroids to generate one or more updated centroids based at least in part on the assigning the centroid to the each data symbol of at least the subset of the plurality of data symbols;
• and
• monitoring the convergence metric associated with the updated set of centroids, the convergence metric based at least in part on a difference between the one or more updated centroids for a first iteration and for a second iteration;
• classifying, after performing the training procedure, the each of the plurality of data symbols according to the updated set of centroids;
• and
• demodulating the plurality of data symbols based at least in part on the classifying the each of the plurality of data symbols.