Modulation mode recognition method based on constellation diagram regression clustering

The invention discloses a modulation mode recognition method based on constellation diagram regression clustering. The method comprises the following steps: step 1, sequencing constellation signal data points to determine a modulation order of a cluster structure: calculating the modulation order by using an OPTICS clustering algorithm; and step 2, identification of a modulation domain: the modulation type of the signal is identified by using linear regression and k-means clustering, specifically, a square error e is calculated through a linear regression algorithm, and when the value of e is less than 1, the signal is an ASK (Amplitude Shift Keying) signal. The mass center is calculated through improved k-means, k is equal to the estimated modulation order, the maximum distance dmax and the minimum distance dmin are obtained through the mass center, the signals with the difference value smaller than the threshold value are PSK signals, and the signals with the difference value larger than the threshold value