Multireceiver distributed collaborative identification process
Step 1: Signal reception: the data received by each receiver is a baseband modulated complex sequence of length ,;
Step 2: Signal preprocessing: extract the real and imaginary parts of the complex baseband signal sequence as and channels:
;
Will be stored as a two-dimensional matrix: ;
Step 3: Feature extraction: input the processed data of each node into the feature extraction network for training, map the data to the high-dimensional feature space, and extract the feature vector after training;
Step 4: Feature fusion: fusion of the feature vectors of different receivers in step 3, dimensionality reduction processing of high-dimensional features, retaining the main features with differences, removing redundant features, and obtaining the fused feature vector;
Step 5: Classification output: send the fused feature vector to the classifier for classification and recognition, and output the recognition result.