Younes Jabrane

High Power Amplifier Pre-Distorter Based on Neural-Fuzzy Systems for OFDM Signals

In this paper, a novel High Power Amplifier (HPA) pre-distorter based on Adaptive Networks - Fuzzy Inference Systems (ANFIS) for Orthogonal Frequency Division Multiplexing (OFDM) signals is proposed and analyzed. Models of Traveling Wave Tube Amplifiers (TWTA) and Solid State Power Amplifiers (SSPA), both memoryless and with memory, have been used for evaluation of the proposed technique. After training, the ANFIS linearizes the HPA response and thus, the obtained signal is extremely similar to the original. An average Error Vector Magnitude (EVM) of 10-6 can be easily obtained with our proposal. As a consequence, the Bit Error Rate (BER) degradation is negligible showing a better performance than what can be achieved with other methods available in the literature. Moreover, the complexity of the proposed scheme is reduced.

Reduction of Power Envelope Fluctuations in OFDM Signals by using Neural Networks

One of the main drawbacks of Orthogonal Frequency Division Multiplexing (OFDM) are the large fluctuations of its power envelope. In this letter, a novel and efficient scheme based on Multilayer Perceptron (MLP) Neural Networks (NN) is proposed. The NN synthesizes the Active Constellation Expansion - (ACE) technique which is able to drastically reduce envelope fluctuations. This is achieved with much lower complexity, faster convergence, and better performance compared to previously available methods.

Reduction of the Envelope Fluctuations of Multi-Carrier Modulations using Adaptive Neural Fuzzy Inference Systems

In this paper, a novel scheme for reducing the envelope fluctuations in multi-carrier signals applying Adaptive Neural Fuzzy Inference Systems (ANFIS) is proposed and analyzed. Once trained with signals with very low envelope fluctuations, such as those obtained by the Active Constellation Expansion - Approximate Gradient Project (ACE-AGP) algorithm, ANFIS approximately reaches a similar reduction as with ACE-AGP for multi-carrier signals without the complexity and the large convergence time of conventional ACE-AGP. We show that our approach is less complex than other previous schemes and with better performance.

Evaluation of the effects of pilots on the envelope fluctuations reduction based on neural fuzzy systems

In this paper, a novel scheme for reducing the envelope fluctuations in OFDM (Orthogonal Frequency division Multiplexing) based on ANFIS (Adaptive Neural Fuzzy Inference Systems) is evaluated when some data can not be modified such as pilot sub-carriers. The loss on the envelope fluctuation reduction is shown to be smaller than 0.3 dB when the proportion of pilots lies into the common values in standards.

Reduction of power envelope fluctuations in ECMA-368 ultra wideband communication system

ECMA-368 is a recent standard, describing the UWB physical layer (PHY-UWB) for wireless personal area network (WPAN) by using the Multiband Orthogonal Frequency Division Multiplexing (MB-OFDM) approach. However, the large power envelope fluctuations of MB-OFDM UWB signal, limits the power efficiency of the High Power Amplifier (HPA) due to nonlinear distortion. In this paper, we will use the Active Constellation Extension Approximate Gradient-Project technique (ACE-AGP) to reduce the power envelope fluctuations in real ECMA-368 system, the efficiency of ACE-AGP is compared with partial transmit sequences (PTS) algorithm. Simulation results show that the reduction of CM by ACE-AGP is better (5dB) than PTS (3dB) and with less complexity, also the BER achieved by both methods does not undergo a significant degradation.