Blind High-Order Modulation Recognition for beyond 5G OSTBC-OFDM Systems via Projected Constellation Vector Learning Network

The blind modulation recognition (BMR) of high-order modulation types is a pressing task and needs to be raised in the calendar for the Beyond 5G (B5G) OSTBC-OFDM (Orthogonal Space-Time Block Coded-Orthogonal Frequency Division Multiplexing) system. In this letter, a BMR algorithm based on a project constellation vector which employs a temporal convolutional network (PCV-TCNet) is proposed to recognize 13 modulation formats, such as high-order 1024QAM and 2048QAM. Without any prior information, a zero-forcing blind equalization algorithm is leveraged to reconstruct the impaired signal. Furthermore, the learning content of PCV-TCNet is PCV features, which are transformed by the constellation diagram of the reconstructed signal. In addition, PCV-TCNet utilizes causal and dilated convolutions to accelerate the BMR process. The simulation results verify the proposed PCV-TCNet for recognizing the high-order modulation types in the B5G OSTBC-OFDM system and demonstrate its preferable recognition performance with the lowest complexity to existing methods.