Real- and Complex-Valued Artificial Intelligence Weight Optimization Algorithms for Smart Antennas in 5/6G Wireless Systems: Linear and Nonlinear Arrays

• Published in: Smart Antennas and Electromagnetic Signal Processing for Advanced Wireless Technology pp. 263 - 302
• Main Authors: Senthilkumar, K.S., Pirapaharan, K., Kunsei, H., Hoole, S.R.H., Hoole, P.R.P.
• Format: Book Chapter
• Language: English
• Published: Denmark Routledge 2020; River Publishers
• Edition: 1
• Subjects: Aerospace & Radar Technology; Electronic Devices; Electronics & Semiconductors; Radar & Navigation Systems; artificial neural network; uniform linear array; Optimized Radiation Patterns; Hyperbolic Tangent Function; MLP; Antenna Beam; Adaptive Array; SLP; Adaptive Array Antenna; Radiation Patterns; Sigmoid Activation Function; Complex Weights; activation function; Perceptron Ann; least mean square (LMS); AF; Squash Functions; Space Time Block Coding; smart antenna; Antenna Elements; Hidden Layer; 5G; Hyperbolic Tangent Activation Function; Neural Network; Adaptive Beamforming; LMS Method; End Fire Directions; beamforming; single neuron weight optimization model (SNWOM)
• ISBN: 9788770222068; 8770222061
• DOI: 10.1201/9781003339564-9

Artificial neural network (ANN) has been applied in many fields including wireless systems. ANN has been used in the design and optimization of different antenna types. The dynamic ability of the ANN to adapt a system variable in any given application is beneficial in smart antenna design for 5/6G wireless systems. This chapter presents a novel ANN algorithm, the single neuron weight optimization model (SNWOM), that optimizes the radiation patterns of uniform linear and nonlinear array (ULA) smart antenna in a desired direction. The robustness of the algorithm was compared against the least mean square (LMS) algorithm for three different functions, namely, the hyperbolic tangent, bipolar, and squash or Elliot functions, for varying number of antenna array elements. SNWOM showed excellent performance as a smart antenna beamformer. The benefit of SNWOM includes fast convergences and demands less hardware resources to offer the favorable performance.