Abstract
The total least squares (TLS) algorithm has shown good convergence performance in the errors-in-variables (EIV) model, where the input and output signals are polluted by noise at the same time. However, the convergence performance will deteriorate when impulse noise is included. In order to solve this problem, a hyperbolic secant total least squares (HSTLS) algorithm is proposed in this paper based on the hyperbolic secant function as the cost function. This algorithm significantly improves the ability to resist impulse noise without significantly increasing the computational complexity. The simulation results of the algorithm demonstrate that the proposed algorithm has better convergence performance.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
