ELM-Based Frame Synchronization in Nonlinear Distortion Scenario Using Superimposed Training

The requirement of high spectrum efficiency puts forward higher requirements on frame synchronization (FS) in wireless communication systems. Meanwhile, a large number nonlinear devices or blocks will inevitably cause distortion. To avoid the occupation bandwidth resources and overcome difficulty distortion, an extreme learning machine (ELM)-based network is introduced into superimposed training-based FS with Firstly, preprocessing procedure utilized to reap features metric (SM). Then, based rough SM, ELM constructed estimate offset boundary. analysis experiment results show that, compared existing methods, proposed method can improve error probability bit rate (BER) symbol detection (SD). In addition, this improvement has its robustness against impacts parameter variations.