A Novel PAPR Reduction Scheme for OFDM Systems Based on Neural Networks

Algorithm 1

Training of the PAPR reduction model.

Definition:

  1. Define the structure of the two modules;

  2. Obtain the OFDM signal and input data from Equation (8);

  3. Define the cost function from Equation (18).

Initialization:

  1. Initialize coefficient vector of model;

  2. Set the weight of the objective function ;

  3. Initialize parameters and of the SCF scheme.

Acquisition of Label Data:

  1. Calculate the clipped signal from Equation (4);

  2. Calculate frequency-domain clipping noise .

  3. Calculate the filtered frequency-domain clipping noise from Equation (6);

  4. Obtain PAPR reduction signal ;

  5. Get the label data from Equation (12).

Model Training:

Loop:

  1. Compute the output data and of the modules from Equation (11) and Equation (15);

  2. Compute objective functions , from Equation (16) and Equation (17);

  3. Compute the cost function ;

  4. Update coefficients according to Adam algorithm.

End