Fevzullah Temurtas

Fuzzy Logic and Neural Network Applications on the Gas Sensor Data: Concentration Estimation

In this study, a fuzzy logic based algorithm is presented for the concentration estimation of the CCl4 and CHCl3 gases by using the steady state sensor response and an artificial neural network (ANN) structure is proposed for the concentration estimation of the same gases inside the sensor response time by using the transient sensor response. The Quartz Crystal Microbalance (QCM) type sensors were used as gas sensors. A computer controlled measurement and automation system with IEEE 488 card was used to control the gas concentration values and to collect the sensor responses. Acceptable performance was obtained for the concentration estimation with fuzzy inference. The appropriateness of the artificial neural network for the gas concentration determination inside the sensor response time is observed.

Application of neural generalized predictive control to robotic manipulators with a cubic trajectory and random disturbances

Abstract In this study, a single-input single-output (SISO) neural generalized predictive control (NGPC) was applied to a three-joint robotic manipulator with a cubic trajectory and random disturbances. The SISO generalized predictive control (GPC) was also used for comparison. Modelling of the dynamics of the robotic manipulator was carried out by using the Lagrange–Euler equations. The frictional effects, random disturbance, carrying and falling load effects were added to the dynamics model. The cubic trajectory principle is used for position reference and velocity reference trajectories. A simulation program was prepared by using Delphi 5.0. All computations for the manipulator dynamics model, GPC_SISO, and NGPC_SISO were done on a PC with 733 MHz CPUs using this program. The parameter estimation algorithm used in the GPC_SISO is Recursive Least Squares. The minimization algorithm used in the NGPC_SISO is Newton–Raphson. According to the simulation outcome, the results from the NGPC_SISO algorithm were better than those from the GPC_SISO algorithm. And these results showed also that the NGPC_SISO reduced the influence of the load changes and disturbances. This means that the NGPC_SISO algorithm combines the advantages of predictive control and the neural network.

Fast detection of hazardous organic gases in the ambient air using adaptive neuro-fuzzy inference systems

In this study, an adaptive neuro-fuzzy inference system (ANFIS) is proposed for the concentration estimation of volatile organic gases before the sensor response time by using the transient sensor response. A neural network (NN) structure with tapped time delays and Mamdanis fuzzy inference system (FIS) are also used for comparison. The estimation results of ANFIS are better than those of the Mamdanis FIS and much closer to those of the NN. Acceptable performance is obtained for all systems, and the appropriateness of ANFIS for the gas-concentration determination before the sensor response time is observed.

A neural network implemented microcontroller system for quantitative classification of hazardous organic gases in the ambient air

In this study, a microcontroller-based gas mixture classification system is proposed to use real-time analyses of the trichloroethylene and acetone binary mixture. A Feed Forward Neural Network (FFNN) structure is performed for quantitative identification of individual gas concentrations (trichloroethylene and acetone) in their gas mixtures. The phthalocyaninecoated Quartz Crystal Microbalance (QCM) type sensors were used as gas sensors. A calibrated Mass Flow Controller (MFC) was used to control the flow rates of carrier gas and trichloroethylene and acetone gas mixtures streams. The components in the binary mixture were quantified by applying the sensor responses from the QCMs sensor array as inputs to the FFNN. The microcontroller-based gas mixture classification system performs Neural Network (NN)-based estimation, the data acquisition and user interface tasks. This system can estimate the gas concentrations of trichloroethylene and acetone with the average errors of 0.08% and 0.97%, respectively.

A Study on Neural Networks with Tapped Time Delays: Gas Concentration Estimation

In this study, an artificial neural network (ANN) structure with tapped time delays is used for the concentration estimation of Toluene gas inside the sensor response time by using the transient sensor response. The Quartz Crystal Microbalance (QCM) type sensors were used as gas sensors. A computer controlled measurement and automation system with IEEE 488 card was used to control the gas concentration values and to collect the sensor responses. The determination of Toluene gas concentrations from the trend of the transient sensor responses achieved with acceptable good performances, and the appropriateness of the artificial neural network for the gas concentration determination inside the sensor response time is observed with these training methods.

Effects of the Trajectory Planning on the Model Based Predictive Robotic Manipulator Control

In this study, the application of the single input single output (SISO) neural generalized predictive control (NGPC) and SISO generalized predictive control (GPC) of a three joint robotic manipulator are presented. The sinusoidal and cubic trajectory principles were used for position reference and velocity reference trajectories. NGPC-SISO algorithm performs better than GPC-SISO algorithm for both trajectories. The GPC-SISO robotic manipulator control results have better values in the case of the sinusoidal trajectory, but the NGPC-SISO robotic manipulator control results for both the cubic and sinusoidal trajectory are almost similar.

Elman's recurrent neural networks using resilient back propagation for harmonic detection

In this study, the method to apply the Elmans recurrent neural networks using resilient back propagation for harmonic detection is described. The feed forward neural networks are also used for comparison. The distorted wave including 5th, 7th, 11th, 13th harmonics were simulated and used for training of the neural networks. The distorted wave including up to 25th harmonics were prepared for testing of the neural networks. Elmans recurrent and feed forward neural networks were used to recognize each harmonic. The results obtained using Elmans recurrent neural networks are better than the results values obtained using the feed forward neural networks for resilient back propagation.

A Study on Neural Networks Using Taylor Series Expansion of Sigmoid Activation Function

The use of microcontroller in neural network realizations is cheaper than those specific neural chips. However, realization of complicated mathematical operations such as sigmoid activation function is difficult via general microcontrollers. On the other hand, it is possible to make approximation to the sigmoid activation function. In this study, Taylor series expansions up to nine terms are used to realize sigmoid activation function. The neural network (NN) structures with Taylor series expansions of sigmoid activation function are used for the concentration estimation of Toluene gas from the trend of the transient sensor responses. The Quartz Crystal Microbalance (QCM) type sensors were used as gas sensors. The appropriateness of the NNs for the gas concentration determination inside the sensor response time is observed with five different terms of Taylor series expansion.

An application of Elman's recurrent neural networks to harmonic detection

One of the difficulties of using Artificial Neural Networks (ANNs) to estimate atmospheric temperature is the large number of potential input variables available. In this study, four different feature extraction methods were used to reduce the input vector to train four networks to estimate temperature at different atmospheric levels. The four techniques used were: genetic algorithms (GA), coefficient of determination (CoD), mutual information (MI) and simple neural analysis (SNA). The results demonstrate that of the four methods used for this data set, mutual information and simple neural analysis can generate networks that have a smaller input parameter set, while still maintaining a high degree of accuracy.

A Study on Neural Networks and Fuzzy Inference Systems for Transient Data

In this study, a Neural Network (NN) structure with tapped time delays and Mamdani’s fuzzy inference system (FIS) are used for the concentration estimation of the Acetone and Chloroform gases inside the sensor response time by using the transient sensor response. The Coated Quartz Crystal Microbalance (QCM) type sensor is used as gas sensors. The estimation results of Mamdani’s FIS are very closer to estimation results of ANN for the transient sensor response. Acceptable good performances are obtained for both systems and the appropriateness of the NN and FIS for the gas concentration determination inside the sensor response time is observed.