Code division multiple access (CDMA) is based on the spread-spectrum technology and is a dominant air interface for 2.5G, 3G, and future wireless networks. For the CDMA downlink, the transmitted CDMA signals from the base station (BS) propagate through a noisy multipath fading communication channel before arriving at the receiver of the user equipment/mobile station (UE/MS). Classical CDMA single-user detection (SUD) algorithms implemented in the UE/MS receiver do not provide the required performance for modern high data-rate applications. In contrast, multi-user detection (MUD) approaches require a lot of a priori information not available to the UE/MS. In this paper, three promising adaptive Riemannian contra-variant (or natural) gradient based user detection approaches, capable of handling the highly dynamic wireless environments, are proposed. The first approach, blind multiuser detection (BMUD), is the process of simultaneously estimating multiple symbol sequences associated with all the users in the downlink of a CDMA communication system using only the received wireless data and without any knowledge of the user spreading codes. This approach is applicable to CDMA systems with relatively short spreading codes but becomes impractical for systems using long spreading codes. We also propose two other adaptive approaches, namely, RAKE -blind source recovery (RAKE-BSR) and RAKE-principal component analysis (RAKE-PCA) that fuse an adaptive stage into a standard RAKE receiver. This adaptation results in robust user detection algorithms with performance exceeding the linear minimum mean squared error (LMMSE) detectors for both Direct Sequence CDMA (DS-CDMA) and wide-band CDMA (WCDMA) systems under conditions of congestion, imprecise channel estimation and unmodeled multiple access interference (MAI).

Recurrent neural networks with various types of hidden units have been used to solve a diverse range of problems involving sequence data. Two of the most recent proposals, gated recurrent units (GRU) and minimal gated units (MGU), have shown comparable promising results on example public datasets. In this paper, we introduce three model variants of the minimal gated unit (MGU) which further simplify that design by reducing the number of parameters in the forget-gate dynamic equation. These three model variants, referred to simply as MGU1, MGU2, and MGU3, were tested on sequences generated from the MNIST dataset and from the Reuters Newswire Topics (RNT) dataset. The new models have shown similar accuracy to the MGU model while using fewer parameters and thus lowering training expense. One model variant, namely MGU2, performed better than MGU on the datasets considered, and thus may be used as an alternate to MGU or GRU in recurrent neural networks. Keywords-recurrent neural networks (RNN), gated recurrent units (GRU), minimal gated units (MGU).

The standard LSTM recurrent neural networks while very powerful in long-range dependency sequence applications have highly complex structure and relatively large (adaptive) parameters. In this work, we present empirical comparison between the standard LSTM recurrent neural network architecture and three new parameter-reduced variants obtained by eliminating combinations of the input signal, bias, and hidden unit signals from individual gating signals. The experiments on two sequence datasets show that the three new variants, called simply as LSTM1, LSTM2, and LSTM3, can achieve comparable performance to the standard LSTM model with less (adaptive) parameters.

Code Division Multiple Access (CDMA) is a channel access method adopted by various radio technologies world-wide. In particular, CDMA is used as an access method in many mobile standards such as CDMA2000 and WCDMA. We address the problem of blind multiuser equalization in the wideband CDMA systems in the noisy multipath propagation environment. Herein, we propose three new blind receiver schemes based on variations of Independent Component Analysis (ICA) within several filtering structures. These adaptive blind CDMA (ABC) receivers do not require knowledge of the propagation parameters or spreading code sequences of the users-they primarily exploit the natural assumption of statistical independence among the symbol signals. We also develop three semi-blind adaptive detectors by incorporating new adaptive methods into the standard Rake receiver structure. Extensive comparative case-studies, based on Bit error rate (BER) performance are carried out for as a function of (i) the number of users, (ii) the number of symbols per user, and (iii) the signal to noise ratio (SNR).The conventional detectors include the baseline Linear Minimum mean squared error (LMMSE) detector. The results show that the proposed methods outperform other detectors in estimating the symbol signals from the received mixed CDMA signals. Moreover, the new blind detectors mitigate the multi access interference (MAI) in CDMA.Keywords Direct Sequence Code Division Multiple Access (DS-CDMA) systems · Wide-band CDMA (W-CDMA) · Independent Component Analysis (ICA) · Robust ICA · Linear Minimum mean squared error (LMMSE) · Multi Access Interference (MAI) · Rake detector · Principle Component Anaylisis (PCA) · FAST ICA · Bit error rate (BER) · signal to noise ratio (SNR).

The Long Short-Term Memory (LSTM) layer is an important advancement in the field of neural networks and machine learning, allowing for effective training and impressive inference performance. LSTM-based neural networks have been successfully employed in various applications such as speech processing and language translation. The LSTM layer can be simplified by removing certain components, potentially speeding up training and runtime with limited change in performance. In particular, the recently introduced variants, called SLIM LSTMs, have shown success in initial experiments to support this view. Here, we perform computational analysis of the validation accuracy of a convolutional plus recurrent neural network architecture using comparatively the standard LSTM and three SLIM LSTM layers. We have found that some realizations of the SLIM LSTM layers can potentially perform as well as the standard LSTM layer for our considered architecture.