Anisotropic Channel Equalization using optimized clustering and Kalman-HMM trained ADALINE

Artificial feature extraction models evolving around adaptive computational decisions which involve spectrum analysis and channel estimation for dedicated cognitive antenna allocation including smart pipelining applications has become a domain of interest for communication protocols which are specifically designed to meet the upcoming architectures and frameworks of 5G and 6G based WLAN/SCADA/DWWAN networks. A novel anisotropic clustering approach has been designed to provide channel equalization in an orthogonally multiplexed, multi-access transmission channel using constellation estimation and entropy interpolation, by using Calinski-Harabasz tuning condition over affinity propagation clusters while employing a novel Kalman optimized and Markov trained Adaptive Linear Neuron (ADALINE) based artificial perceptron layer (ApNN). HuangHilbert spectrum margins have been provided to interpretate signal improvement after equalization for a Rayleigh Fading channel undergoing Active White Gaussian Noise (AWGN). MQAM Modulation is considered over an Orthogonal Frequency Division Multiple Access (OF DMA) channel for experimental analysis in a virtual software defined Wireless Local Area Network (Vi-WLAN) environment.