Error Back Propagation Learing of Neural Networks Including Preorganized Structure and its Appplication to Inverse Dynamics Learning of Robot Manipulators

The chromatic aberration estimations of touch panel (TP) film by using neural networks are presented in this paper. The neural networks with error back-propagation (BP) learning algorithm were used to catch the complex relationship between the chromatic aberration, i.e., L.A.B. values, and the relative parameters of TP decoration film. An artificial intelligent (AI) estimator based on neural model for the estimation of physical property of TP film is expected to be developed. From the simulation results shown, the estimations of chromatic aberration of TP film are very accurate. In other words, such an AI estimator is quite promising and potential in commercial using.

Download Full-text

End effector target position learning using feedforward with error back-propagation and recurrent neural networks

Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN'94) ◽

10.1109/icnn.1994.374637 ◽

2002 ◽

Cited By ~ 1

Author(s):

J. Zurada

Keyword(s):

Neural Networks ◽

Recurrent Neural Networks ◽

Back Propagation ◽

Target Position ◽

End Effector ◽

Error Back Propagation ◽

Position Learning ◽

Effector Target

Download Full-text

Building robust neural networks using different loss functions

Analysis and data processing systems ◽

10.17212/2782-2001-2021-2-67-82 ◽

2021 ◽

pp. 67-82

Author(s):

Maria Sivak ◽

◽

Vladimir Timofeev ◽

Keyword(s):

Neural Network ◽

Neural Networks ◽

Loss Function ◽

Back Propagation ◽

Loss Functions ◽

Back Propagation Algorithm ◽

Error Back Propagation ◽

Propagation Algorithm ◽

Error Back Propagation Algorithm ◽

Parameter Values

The paper considers the problem of building robust neural networks using different robust loss functions. Applying such neural networks is reasonably when working with noisy data, and it can serve as an alternative to data preprocessing and to making neural network architecture more complex. In order to work adequately, the error back-propagation algorithm requires a loss function to be continuously or two-times differentiable. According to this requirement, two five robust loss functions were chosen (Andrews, Welsch, Huber, Ramsey and Fair). Using the above-mentioned functions in the error back-propagation algorithm instead of the quadratic one allows obtaining an entirely new class of neural networks. For investigating the properties of the built networks a number of computational experiments were carried out. Different values of outliers’ fraction and various numbers of epochs were considered. The first step included adjusting the obtained neural networks, which lead to choosing such values of internal loss function parameters that resulted in achieving the highest accuracy of a neural network. To determine the ranges of parameter values, a preliminary study was pursued. The results of the first stage allowed giving recommendations on choosing the best parameter values for each of the loss functions under study. The second stage dealt with comparing the investigated robust networks with each other and with the classical one. The analysis of the results shows that using the robust technique leads to a significant increase in neural network accuracy and in a learning rate.

Download Full-text

Stellar Image Interpretation System Using Artificial Neural Networks:

Acta Polytechnica ◽

10.14311/506 ◽

2004 ◽

Vol 44 (1) ◽

Author(s):

A. El-Bassuny Alawy ◽

F. I. Y. Elnagahy ◽

A. A. Haroon ◽

Y. A. Azzam ◽

B. Šimák

Keyword(s):

Neural Networks ◽

Artificial Neural Networks ◽

Image Interpretation ◽

Back Propagation ◽

Present Approach ◽

Error Back Propagation ◽

Polar Function ◽

Open Star ◽

Artificial Neural ◽

Artificial Neural Network Ann

A supervised Artificial Neural Network (ANN) based system is being developed employing the Bi-polar function for identifying stellar images in CCD frames. It is based on feed-forward artificial neural networks with error back-propagation learning. It has been coded in C language. The learning process was performed on a 341 input pattern set, while a similar set was used for testing. The present approach has been applied on a CCD frame of the open star cluster M67. The results obtained have been discussed and compared with those derived in our previous work employing the Uni-polar function and by a package known in the astronomical community (DAOPHOT-II). Full agreement was found between the present approach, that of Elnagahy et al, and the standard astronomical data for the cluster. It has been shown that the developed technique resembles that of the Uni-Polar function, possessing a simple, much faster yet reliable approach. Moreover, neither prior knowledge on, nor initial data from, the frame to be analysed is required, as it is for DAOPHOT-II.

Download Full-text

Gravity Compensation for Manipulator Control by Neural Networks with Partially Preorganized Structure

Journal of Robotics and Mechatronics ◽

10.20965/jrm.1990.p0282 ◽

1990 ◽

Vol 2 (4) ◽

pp. 282-287

Author(s):

Toshio Tsuji ◽

◽

Masataka Nishida ◽

Toshiaki Takahashi ◽

Koji Ito

Keyword(s):

Neural Networks ◽

Adaptive Learning ◽

Center Of Mass ◽

Back Propagation ◽

Equation Of Motion ◽

Point Load ◽

Gravity Compensation ◽

Error Back Propagation ◽

Industrial Manipulators ◽

Parameter Values

The gravity torque of a manipulator can be compensated if the equation of motion can be correctly introduced, but in general industrial manipulators, there are many cases when the parameter values such as the position of center of mass are not clear, and these values largely change by the exchange of hand portions and the grasping of substances. Furthermore, in addition to unclear parameters, there are factors which occur by structural gravity compensation (spring and counter-balance) and which in many cases are difficult to express with the equation of motion. In this paper, compensation of the gravity torque of the manipulator is studied by, the use of neural networks. For this purpose, a model which makes the structure known to be contained in mapping as a unit with preorganized characteristics prepared in parallel with hidden unit of error back propagation-type neural network is proposed, by which the characteristics of the link system which is the object for learning can be imbedded into the network as preorganized knowledge beforehand. Finally, the results of experiments done with the use of industrial manipulators are given, and it is made clear that the compensation of gravity torque of manipulator and adaptive learning for end-point load are possible by the use of this model.

Download Full-text

KINK FINDING IN THE ALEPH TPC USING NEURAL NETWORKS

Modern Physics Letters A ◽

10.1142/s021773239300310x ◽

1993 ◽

Vol 08 (29) ◽

pp. 2715-2727 ◽

Cited By ~ 1

Author(s):

GEORG STIMPFL-ABELE

Keyword(s):

Neural Networks ◽

Back Propagation ◽

Simulated Data ◽

Radial Position ◽

Neural Net ◽

Decay Vertex ◽

Feed Forward ◽

Error Back Propagation ◽

Conventional Methods ◽

Tracking Device

The task of finding the decays of charged tracks inside a tracking device is divided into two parts. First a neural net is used to recognize kinks in well-reconstructed tracks. If a kink has been found, a second net determines the radial position of the decay vertex. Both algorithms use feed-forward nets with error back-propagation. Very good performance is obtained in comparison with conventional methods using simulated data from the ALEPH TPC. The behavior of the nets is analyzed by studying the correlations between the inputs and the output.

Download Full-text

Neural Network Circuits for Embedded Sensors Applications

System and Circuit Design for Biologically-Inspired Intelligent Learning ◽

10.4018/978-1-60960-018-1.ch012 ◽

2011 ◽

pp. 276-294

Author(s):

N. Medrano ◽

G. Zatorre ◽

M. T. Sanz ◽

B. Calvo ◽

S. Celma

Keyword(s):

Neural Network ◽

Neural Networks ◽

Ad Hoc ◽

Low Cost ◽

Back Propagation ◽

Cmos Technology ◽

Error Back Propagation ◽

Digital Electronics ◽

Low Sensitivity ◽

Selection Of

This chapter presents the suitability, development and implementation of programmable analogue artificial neural networks for sensor conditioning in embedded systems. Comments on the use of analogue instead of digital electronics due to the size and power constraints of these applications are included. Performance of an ad-hoc analogue architecture is evaluated, and its characteristics are analyzed. We will verify its low sensitivity to undesired effects, such as component mismatching, due to the capability of selecting and programming the proper weights for a given task. In addition, a brief discussion is offered on the selection of perturbative algorithms instead of classical error back-propagation techniques for weight tuning. At the end of the chapter, we will show the main characteristics of the proposed arithmetic cells implemented in a low-cost CMOS technology.

Download Full-text

Flower image classification with basket of features and multi layered artificial neural networks

International Journal of Engineering & Technology ◽

10.14419/ijet.v7i1.1.10795 ◽

2017 ◽

Vol 7 (1.1) ◽

pp. 642

Author(s):

Syed Inthiyaz ◽

B T.P.Madhav ◽

Ch Raghava Prasad

Keyword(s):

Neural Networks ◽

Active Contours ◽

Back Propagation ◽

Back Propagation Algorithm ◽

Fusion Model ◽

Multiple Sources ◽

Error Back Propagation ◽

Multiple Feature ◽

Error Back Propagation Algorithm ◽

Segmentation Models

Artificial intelligence is penetrating most of the classification and recognition tasks performed by a computer. This work proposes to classify flower images based on features extracted during segmentation and after segmentation using multiple layered neural networks. The segmentation models used are watershed, wavelet, wavelet fusion model, supervised active contours based on shape, color and Local binary pattern textures and color, fused textures based active contours. Multi-dimension feature vectors are constructed from these segmented results for each indexed flower image labelled with their name. Each feature becomes input to a neuron in various feature layers and error back propagation algorithm with convex optimization structure trains these multiple feature layers. Testing with different flower images sets from multiple sources resulted in average classification accuracy of 92% for shape, color and texture supervised active contour segmented flower images.

Download Full-text

Artificial neural networks employment in the prediction of evapotranspiration of greenhouse-grown sweet pepper

Revista Brasileira de Engenharia Agrícola e Ambiental ◽

10.1590/1807-1929/agriambi.v20n6p507-512 ◽

2016 ◽

Vol 20 (6) ◽

pp. 507-512 ◽

Cited By ~ 5

Author(s):

Héliton Pandorfi ◽

Alan C. Bezerra ◽

Roberto T. Atarassi ◽

Frederico M. C. Vieira ◽

José A. D. Barbosa Filho ◽

...

Keyword(s):

Neural Networks ◽

Artificial Neural Networks ◽

Network Architecture ◽

Learning Algorithm ◽

Meteorological Data ◽

Back Propagation ◽

Sweet Pepper ◽

Coefficient Of Determination ◽

Error Back Propagation ◽

Artificial Neural

ABSTRACT This study aimed to investigate the applicability of artificial neural networks (ANNs) in the prediction of evapotranspiration of sweet pepper cultivated in a greenhouse. The used data encompass the second crop cycle, from September 2013 to February 2014, constituting 135 days of daily meteorological data, referring to the following variables: temperature and relative air humidity, wind speed and solar radiation (input variables), as well as evapotranspiration (output variable), determined using data obtained by load-cell weighing lysimeter. The recorded data were divided into three sets for training, testing and validation. The ANN learning model recognized the evapotranspiration patterns with acceptable accuracy, with mean square error of 0.005, in comparison to the data recorded in the lysimeter, with coefficient of determination of 0.87, demonstrating the best approximation for the 4-21-1 network architecture, with multilayers, error back-propagation learning algorithm and learning rate of 0.01.

Download Full-text