Linearization of Bipolar Amplifiers Based on Neural-Network Training Algorithm

Abstract

Traditional bipolar differential amplifiers have only a ±5mV operational range with nonlinear distortion below 0.1dB. In this paper, a linearization technique based on neural network training algorithm is proposed to expand this 0.1dB linear region to a much wider ±200mV range. Compared with the traditional and recent state-of-the-art techniques for linearization, where gain or noise performance is always being tradeoff for high linearity, the proposed technique leads to the increase of the amplifier's gain with low noise. Another advantage of the proposed approach is that, it is implemented using a simple, paralleled structure that leads to a smaller area and less power consumption than other linearization schemes. The design procedure also enables a completely customizable solution based on required linear range, ripples, and the number of available transistors. The experimental results show that the proposed linearized amplifier has a very good linearity and is capable to wide bandwidth, good gain, and low noise performance.

RF Power Amplifier Linearization in Professional Mobile Radio Communications Using Artificial Neural Networks

“…However, neither is relevant, since the initial training is only carried out during the manufacturing process of the terminal using external instrumentation, and the need for an additional analog output is easily obtainable with the DSP. Unlike previous works that apply complex ANNs to linearize RF power amplifiers [7]- [13] requiring additional computing hardware as FPGAs or powerful DSPs (thus orienting this solutions to static infrastructures), this work performs the practical implementation of a simple ANN into the DSP of a radio terminal, and the compensation of possible changes in working conditions such as frequency, temperature and supply voltage, therefore allowing to simplify and eliminate parts of hardware that until now were necessary. This will have repercussions in a reduction of cost and size of the terminals.…”

“…In order to increase efficiency and to meet the telecommunications standards, many different techniques have been proposed and applied to extend the linear range of the power amplifier response: Cartesian Feedback [1], Feedforward [2], Volterra series [3], Memory Polynomials [4], Wiener and Hammerstein models [5], Lookup Tables (LUT) [6], Artificial Neural Networks (ANN) [7]- [13], Neural-Fuzzy systems [14], Genetic Algorithms [15], etc.…”

CMOS Implementation of ANNs Based on Analog Optimization of N-Dimensional Objective Functions

“…The dimensions of the main semiconductor N-channel and P-channel transistors are listed in Table 4. The characteristic values of the integrator are described in Table 5 [18,19].…”

“…Hence, when many input voltages are connected to the inverting input terminal, the resulting output is the sum of all the input voltages applied, although inverted; this output, combined with the feedback resistor, generates the multiplication by a weight. The circuit for implementing a neuron is shown in Figure 7, where the function Σ computes multiplications and the activation function is a Sigmoid [19].…”

Computer English Speech Independent Evaluation System of the Fusion Discrimination Training Algorithm

“…Computer English speech refers to the independent evaluation rules have been set on the basis of good, all kinds of interactive [1] and other subjects of environmental development, through the initiative to achieve the interaction of information feedback, in order to fully assess the results of computer English pronunciation [2], the need for independent assessments of the information obtained, to improve their English pronunciation the level of [3][4]. If there is a strong adaptability in a certain level of English pronunciation assessment, this adaptability can be transferred to other levels of voice evaluation, and ultimately make the whole English pronunciation assessment innovation ability can be improved [5].…”