Fernando Morgado-Dias

A survey of artificial neural network training tools

Artificial neural networks (ANN) are currently an additional tool which the engineer can use for a variety of purposes. Classification and regression are the most common tasks; however, control, modeling, prediction and forecasting are common tasks as well. For over three decades, the field of ANN has been the center of intense research. As a result, one of the outcomes has been the development of a large set of software tools used to train these kinds of networks, making the selection of an adequate tool difficult for a new user. This paper aims to help the ANN user choose the most appropriate tool for its application by providing a large survey of the solutions available, as well as listing and explaining their characteristics and terms of use. The paper limits itself to focusing on the tools which were developed for ANN and the relevant characteristics of these tools, such as the operating systems, hardware requirements, license types, architectures and algorithms available.

Low-resource hardware implementation of the hyperbolic tangent for artificial neural networks

Artificial neural networks are a widespread tool with application in a variety of areas ranging from the social sciences to engineering. Many of these applications have reached a hardware implementation phase and have been documented in scientific papers. Unfortunately, most of the implementations have a simplified hyperbolic tangent replacement which has been the most common problem, as well as the most resource-consuming block in terms of hardware. This paper proposes a low-resource hardware implementation of the hyperbolic tangent, by using the simplest solution in order to obtain the lowest error possible thus far with a set of 25 polynomials of third order, obtained with Chebyshev interpolations. The results obtained show that the solution proposed holds a low error while simultaneously promising the use of low resources, as only third-order polynomials are used.

A new solution to the hyperbolic tangent implementation in hardware: polynomial modeling of the fractional exponential part

The most difficult part of an artificial neural network to implement in hardware is the nonlinear activation function. For most implementations, the function used is the hyperbolic tangent. This function has received much attention in relation to hardware implementation. Nevertheless, there is no consensus regarding the best solution. In this paper, we propose a new approach by implementing the hyperbolic tangent in hardware with a polynomial modeling of the fractional exponential part. The results in the paper then demonstrate, through the use of an example, that this solution is faster than the CORDIC algorithm, but slower than the piecewise linear solution with the same error. The advantage over the piecewise linear approach is that it uses less memory.

Automatic general-purpose neural hardware generator

Abstract Artificial neural networks have a wide range of applications. In some applications, specific hardware is necessary when a PC cannot be connected or due to other factors such as speed, price and fault tolerance. The difficulty in producing hardware for neural networks is associated with price, accuracy and development time. Most users also prefer the network trained with a high-level tool without reducing resolution and simplifying the activation function for hardware implementation. This paper proposes an automatic general-purpose neural hardware generator, simple to use, with adjustable accuracy that provides direct hardware implementation for neural networks with FPGAs without further development.

Hardware implementation of an artificial neural network model to predict the energy production of a photovoltaic system

An artificial neural network trained using only the data of solar radiation presents a good solution to predict, in real time, the power produced by a photovoltaic system. Even though the neural network can run on a Personal Computer, it is expensive to have a control room with a Personal Computer for small photovoltaic installations. A FPGA running the neural network hardware will be faster and less expensive. In this work, to assist the hardware implementation of an artificial neural network with a FPGA, a specific tool was used: an Automatic General Purpose Neural Hardware Generator. This tool allows for an automatic configuration system that enables the user to configure the artificial neural network, releasing the user from the details of the physical implementation. The results show that it is possible to accurately model the photovoltaic installation based on data from a nearby meteorological installation and the hardware implementation produces low cost and precise results.

A portable wireless device based on oximetry for sleep apnea detection

Obstructive sleep apnea is a highly prevalent sleep related breathing disorder and polysomnography is the gold standard exam for diagnosis. Despite providing results with high accuracy this multi-parametric test is expensive, time consuming and does not fit with the new tendency in health care that is changing the focus to prevention and wellness. Home health care is seen as a possible way to address this problematic by using minimal invasive devices, providing low cost of diagnosis and higher accessibility. To address this, a portable and automated sleep apnea detector was designed and evaluated. The device uses one SpO2 sensor and the analysis is based on the connection between oxygen saturation and apnea events. The measured signals are received in a field-programmable gate array that checks for errors and implements the communication protocols of two wireless transmitters. Two solutions were implemented for processing the data: one based on a smartphone (due to availability and low cost) and another based on a personal computer (for a higher computation capability). The algorithms were implemented in Java, for the smartphone, and in Python, for the computer. Both implementations have a graphical user interface to simplify the device operation. The algorithms were tested using a database consisting of 70 patients with the SpO2 signal collected in a Hospital. The algorithm performance achieved an average accuracy, sensitivity and specificity of 87.5, 79.5 and 90.8% respectively.

Performance comparison of ANN training algorithms for classification

The Artificial Neural Network research community has been actively working since the beginning of the 80s. Since then many existing algorithm were adapted, many new algorithms were created and many times the set of algorithms was revisited and reinvented. As a result an enormous set of algorithms exists and, even for the experienced user it is not easy to choose the best algorithm for a given task or dataset, even though many of the algorithms are available in implementations of existing tools. In this work we have chosen a set of algorithms which are tested with a few datasets and tested several times for different initial sets of weights and different numbers of hidden neurons while keeping one hidden layer for all the Feedforward Artificial Neural Networks.

Automatic Detection of a Phases for CAP Classification.

The aim of this study is to develop an automatic detector of the cyclic alternating pattern by first detecting the activation phases (A phases) of this pattern, analysing the electroencephalogram during sleep, and then applying a finite state machine to implement the final classification. A public database was used to test the algorithms and a total of eleven features were analysed. Sequential feature selection was employed to select the most relevant features and a post processing procedure was used for further improvement of the classification. The classification of the A phases was produced using linear discriminant analysis and the average accuracy, sensitivity and specificity was, respectively, 75%, 78% and 74%. The cyclic alternating pattern detection accuracy was 75%. When comparing with the state of the art, the proposed method achieved the highest sensitivity but a lower accuracy since the fallowed approach was to keep the REM periods, contrary to the method that is used in the majority of the state of the art publications which leads to an increase in the overall performance. However, the approach of this work is more suitable for automatic system implementation since no alteration of the EEG data is needed.

NanEye-An Endoscopy Sensor With 3-D Image Synchronization

In this paper, an innovative camera synchronization technique is presented, enabling the combination of two 1 mm

A Method for Designing Emg Integrator using an FPGA

\times 1