Janusz Pochmara

Efficient algorithm for adjustment of adaptive predistorter in OFDM transmitter

Adaptive predistortion is a technique which is aimed at compensation of nonlinear distortion introduced by a high power amplifier (HPA). This paper presents a simple method of acceleration of the adaptive algorithm of the HPA predistorter. The values of the predistorter complex gain stored in RAM and depending on the input signal amplitude are efficiently modified nor only in the address indicated by the current signal amplitude but also in the whole range of locations around it. The simulation results of the predistorters adaptation show a visible improvement in the convergence rate as compared with the standard gradient algorithm.

Improving compensation of nonlinear distortions in OFDM system using recurrent neural network with conjugate gradient algorithm

The paper presents a neural network predistortion technique compensating for nonlinear distortions caused by an HPA (high power amplifier) cascaded with a filter in an OFDM (orthogonal frequency division multiplexing) system. It is confirmed by computer simulation that the proposed approach produces a faster convergence speed than the conventional backpropagation algorithm. The predistortion technique based on a neural network is very attractive from the implementation point of view, because of the low amount of RAM required and rapid convergence from a blind start.

Multisensor data fusion using Elman neural networks

The paper presents a navigation system based on Elman Artificial Neural Network (ANN). The task of data fusion from different sensors is realized by trained ANN. Determining position in space is an issue of nonlinear hence. Not every type of ANN is used for such a task. Choice of Elman ANN was dictated by its construction and successfully applications to nonlinear problems requiring prediction. Elman network is composed of three layers. Comprises a layer of hidden layer units context which is connected to the hidden layer. Context-sensitive layer allows for store the values of previous hidden units. With this layer prediction is possible in sequential order. This is the effect of contextual memory where information is stored about what it was before. This kind of functionality is not able to provide any other standard neural network unidirectional. The system consists of MEMS (Micro Electro-Mechanical Systems) sensors, which are based on IMU (Inertial Measurement Unit). IMU is composed from gyroscopes, accelerometers and magnetometers which provide three dimensional linear accelerations and angular rates. This is a classic set of sensors for determining the position in space. The study presents the results of the implementation of algorithms for determining the position in space using trained Elman ANN. The data samples to train ANN were collected during the test flight of Quadrocopter. Paper presents the performance for different configurations of Elman ANN. Presented system provides easy addition of other sensors e.g. GPS/GLONASS receiver.

A set of dynamic artificial neural networks for robot sensor failure detection

Paper presents a novel idea of failure detection mechanism for complex control environments. The mechanism is composed of several dynamic artificial neural networks that work in parallel in order to detect a failing signal from one of the on-board robot sensors. The simulation results show that the system is capable of detecting a failing control system quickly and efficiently.

Modeling Power Amplifier Nonlinearities with Artifical Neural Network

This paper describes a method for modeling nonlinear power amplifier for RF applications. Presented model is based on the neural network architecture and can be applied to characterize memoryless behaviour of power amplifiers. For simulation we use feed-forward neural network to make a normalized input-output conversion for nonlinear characteristic of power amplifier. The results show that neural network can be a good tool in modeling process of nonlinear components used in RF circuits. The numerical comparison between existing methods (Saleh model) is computed in order to evaluate performance of the proposed model of interpolation of power amplifier nonlinearities.

Using neural network for reduction distrotion introduced by power amplifier in digital communication systems

We proposed and improved an adaptive neural predistorter, which can automatically compensate for amplifier nonlinearity and thus makes it possible to transmit OFDM signals without incurring intolerable distortions. The neural predistorter utilizes gradient algorithms for its adaptation. Our results indicate clear improvements in performance for neural networks networks incorporating memory into their structure