The base station antenna array system output SNIR determination for different spatial channel models

Abstract

Absir-mi -In this paper we consider the application of some adaptive algorithms for the antenna array beam steering. Three spatial channel models ( Lee's Model, Gaussian Angle of Arrival Model and Discrete Uniform Distribution Angle of Arrival Model) and 12 -element antenna array are simulated. The antenna array beam steering is performed by using the adaptive algorithms: Least Mean Sgziare -LMS, Least Mean Fourth -LMF, Recursive Least Square -RLS and Sample Matrix Inversion -SMZ. The output antenna array SNIR is analyzed and plotted relating to the angular width and the iteration number the adaptive algorithms.

Adaptive Antenna Array Control in the Presence of Interfering Signals With Stochastic Arrivals: Assessment of a GA-Based Procedure

“…The convergence rate of GA-based array control strategies only depends on the population dimension P, [17], being independent of the specific parameters of the optimization problem to be faced. This is not true for LMS algorithm, whose convergence rate directly depends on the eigenvalue spread of the covariance matrix [3]. Better performances can be achieved by RLS, as clearly stated in [3].…”

“…Generally, this is not a trivial task [2]. Consequently, many alternative solutions based on dynamic programming in order to avoid the matrix inversion [2] [3] have been proposed. Least-Mean-Square (LMS) and Recursive-Least-Square (RLS) algorithms are well-known examples of mathematical solutions to array optimization (see [3] for a thorough overview).…”

“…Consequently, many alternative solutions based on dynamic programming in order to avoid the matrix inversion [2] [3] have been proposed. Least-Mean-Square (LMS) and Recursive-Least-Square (RLS) algorithms are well-known examples of mathematical solutions to array optimization (see [3] for a thorough overview). Despite their mathematical elegance, such methods present some drawbacks, hindering their practical implementation [4] [5] [6].…”

“…The analysis presented in [7] points out that a choice of 8 bits for weight quantization could be satisfactory in several practical applications. On the other hand, the convergence of conventional approaches for array optimization strongly depends upon the eigenvalue spread [3] and on the external noise environment. Moreover, these techniques result very slow in severe jamming situations [8] and do not prevent the solution be trapped in local minima [4] [6].…”

“…In this framework, the use of Genetic Algorithms (GAs) can be regarded as a valuable solution for the array optimization problem [4] [5] [9]. In [9], it has been shown that GA-based approaches outperform conventional solutions based on LMS strategy (being RLS an efficient variant of LMS providing faster convergence [3]). …”