Probabilistic shaping QAM dynamic equalization and digital signal processing method

A probabilistic shaping quadrature amplitude modulation (QAM) dynamic equalization apparatus, comprising:

• an inner-ring interception module, configured to intercept n inner rings after clock recovery from a received QAM transmission signal;
• a polar coordinate conversion module, configured to convert the n inner rings from a two-dimensional rectangular coordinate system into a polar coordinate system, to obtain a polar coordinate constellation diagram;
• a local density calculation module, configured to calculate a local density ρi=Σj≠ie−(disti,j/Dc)2 of each data point X=[X1, X2, . . . , XN] in the polar coordinate constellation diagram,
• wherein N represents a total quantity of data symbols, i, j ∈N, disti,j is a distance between a point Xi and a point Xj and Dc is a cutoff distance;
• a minimum distance calculation module, configured to calculate relative distances between each data point and a plurality of data points with a local density greater than the local density of the each data point, and acquire a minimum distance corresponding to each data point:
• arranging the local densities of the data points in descending order;
• calculating the minimum distance δi=minj≠1:i−1(disti,j) between the point Xi and a plurality of data points with a local density greater than the local density of the point Xi;
• and
• setting the minimum distance of a point X1 with a highest local density to a maximum distance δl=maxj≠2:N(disti,j) between the data point X1 and other data points;
• a cluster center determination module, configured to generate a decision diagram according to the minimum distance corresponding to each data point, and screen outliers according to the decision diagram, to obtain n cluster centers;
• a clustering module, configured to classify the data points according to the n cluster centers by using a K-means algorithm, to obtain a K-means cluster graph;
• a decision area determination module, configured to average a maximum value of local densities of an nth cluster of clustering data points and a minimum value of local densities of an (n−1)th cluster of clustering data points, to obtain a decision radius and a decision area;
• and
• a dynamic equalization module, configured to introduce the decision radius and the decision area into a cascaded multi-mode algorithm equalizer and a radius-directed equalizer to perform dynamic equalization.