Hyperbolic Functions Impact Evaluation on Channel Identification Based on Recursive Kernel Algorithm

Over the last years, the subject of non-linear system identification has attracted considerable interest due to the numerous applications that could be used and the broad multidisciplinary scope of the field. In this paper, we exploit a non-linear system with a linear finite impulse response (FIR) sub-element under the existence of Gaussian noise, while using an algorithm based on positive defined kernels to identify the channel model parameters. Firstly, we have used an algorithm based on the theory of positive definite kernels to estimate the parameters of the selective channel. Secondly, we have studied the influence of the nonlinearity function of modeled single-input single-output (SISO) communication systems with binary-valued output observations on the identification performance of the channel impulse responses. To show which nonlinear function can achieve the most efficient result for channel parameter identification, some examples of simulation results are provided in this works.