Time-delay Artificial Neural Networks Research Articles

An Experimental Study on Sensorless Determination of the Projectile Position by Artificial Neural Network in Magnetic Launcher Systems

Electromagnetic launchers are systems in which energy is converted from one form to another. The energy applied to the launcher is converted into magnetic energy in the coil and then into motion energy. The object to be launched follows the magnetic field created in one or more fixed coils. The coil can be fed by direct current (dc) or alternative current (ac) source. In dc systems, the capacitor charging time delays the serial launch. In ac systems, the absence of capacitors enables serial launching. In this study, the application of an ac electromagnetic launch system, which provides a serial launch, is proposed. The maximum force acting on the projectile can be achieved by timely cutting off the energy applied to the coil. Therefore, in electromagnetic launchers, it is critical to know the position of the projectile in the coil. In launcher systems fed with alternating current, the projectile position can be predicted from current changes without the need for an additional setup. In the study, TMS320F28335 digital signal processor (DSP) microprocessor with high processing capacity was used to detect the trigger moment from coil current. Time-delayed artificial neural network was used to estimate the position of the projectile within the coil by using the obtained current differences. The normalized values of the current differences were applied to the time-delayed multilayer neural network (MLNN) algorithm and the projectile position was estimated with 94.67% accuracy. The proposed system increases efficiency compared to systems using conventional sensors, as it prevents or reduces the retraction force.

Soft Computing of a Medically Important Arthropod Vector with Autoregressive Recurrent and Focused Time Delay Artificial Neural Networks.

Simple SummaryArthropod vectors are responsible for transmitting a large number of diseases, and for most, there are still not available effective vaccines. Vector disease control is mostly achieved by a sustained prediction of vector populations to maintain support for surveillance and control activities. Mathematical models may assist in predicting arthropod population dynamics. However, arthropod dynamics, and mosquitoes particularly, due their complex life cycle, often exhibit an abrupt and non-linear occurrence. Therefore, there is a growing interest in describing mosquito population dynamics using new methodologies. In this work, we made an effort to gain insights into the non-linear population dynamics of Culex sp. adults, aiming to introduce straightforward soft-computing techniques based on artificial neural networks (ANNs). We propose two kind of models, one autoregressive, handling temperature as an exogenous driver and population as an endogenous one, and a second based only on the exogenous factor. To the best of our knowledge, this is the first study using recurrent neural networks and the most influential environmental variable for prediction of the WNv vector Culex sp. population dynamics, providing a new framework to be used in arthropod decision-support systems.A central issue of public health strategies is the availability of decision tools to be used in the preventive management of the transmission cycle of vector-borne diseases. In this work, we present, for the first time, a soft system computing modeling approach using two dynamic artificial neural network (ANNs) models to describe and predict the non-linear incidence and time evolution of a medically important mosquito species, Culex sp., in Northern Greece. The first model is an exogenous non-linear autoregressive recurrent neural network (NARX), which is designed to take as inputs the temperature as an exogenous variable and mosquito abundance as endogenous variable. The second model is a focused time-delay neural network (FTD), which takes into account only the temperature variable as input to provide forecasts of the mosquito abundance as the target variable. Both models behaved well considering the non-linear nature of the adult mosquito abundance data. Although, the NARX model predicted slightly better (R = 0.623) compared to the FTD model (R = 0.534), the advantage of the FTD over the NARX neural network model is that it can be applied in the case where past values of the population system, here mosquito abundance, are not available for their forecasting.

Data-driven order reduction and velocity field reconstruction using neural networks: The case of a turbulent boundary layer

We present a data-driven methodology to achieve the identification of coherent structure dynamics and system order reduction of an experimental turbulent boundary layer flow. The flow is characterized using time-resolved optical flow particle image velocimetry, leading to dense velocity fields that can be used both to monitor the overall dynamics of the flow and to define as many local visual sensors as needed. A Proper Orthogonal Decomposition (POD) is first applied to define a reduced-order system. A non-linear mapping between the local upstream sensors (inputs sensors) and the full-field dynamics (POD coefficients) as outputs is sought using an optimal focused time-delay Artificial Neural Network (ANN). The choices of sensors, ANN architecture, and training parameters are shown to play a critical role. It is verified that a shallow ANN, with the proper sensor memory size, can lead to a satisfying full-field dynamics identification, coherent structure reconstruction, and system order reduction of this turbulent flow.

Ankle torque forecasting using time-delayed neural networks.

A method for ankle torque prediction ahead of the current time is proposed in this paper. The mean average value of EMG signals from four muscles, alongside the joint angle and angular velocity of the right ankle, were used as input parameters to train a time-delayed artificial neural network. Data collected from five healthy subjects were used to generate the dataset to train and test the model. The model predicted ankle torque for five different future times from zero to 2 seconds. Model predictions were compared to torque calculated from inverse dynamics for each subject. The model predicted ankle torque up to 1 second ahead of time with normalized root mean squared error of less than 15 percent while the coefficient of determination was over 0.85.Clinical Relevance- the potential of the model for predicting joint torque ahead of time is helpful to establish an intuitive interaction between human and assistive robots. This model has application to assist patients with neurological disorders.

An Intelligent Calculation Method of Volterra Time-Domain Kernel Based on Time-Delay Artificial Neural Network

To solve the problems of high complexity and low accuracy in Volterra time-domain kernel calculation of a nonlinear system, this paper proposes an intelligent calculation method of Volterra time-domain kernel by time-delay artificial neural networks (TDANNs) and also designs a root mean square error (RMSE) index to choose the neuron number of the network input layer. Firstly, a three-layer TDANN is designed according to the characteristics of the Volterra model. Secondly, the relationship between parameters of TDANN and Volterra time-domain kernel is analyzed, and then three-order expressions of Volterra time-domain kernel are derived. The calculation of Volterra time-domain kernel is completed by network training. Finally, it is verified by a nonlinear system. Simulation results indicate that compared with traditional methods, the new method has higher accuracy, and it can realize the batch calculation of Volterra kernel, which not only improves the calculation efficiency but also provides accurate data for fault diagnosis based on Volterra kernel in further research work.

Ankle Joint Torque Prediction Based on Surface Electromyographic and Angular Velocity Signals

Joint torque prediction plays an important role in quantitative limb rehabilitation assessment and exoskeleton robot, and it is essential to acquire feedback or feedforward signal for adaptive functional electrical stimulation (FES) control. The Surface electromyography (sEMG) signal is one of the basic processing techniques to detect muscle activity, and also one favorable technique to estimate joint torque. In order to predict joint torque in a wide range of real time convenient applications, it is necessary to fuse sEMG signals with other convenient physical sensors such as accelerometers and gyroscopes, herein, we use a time delay artificial neural network to predict human joint force of ankle eversion and inversion based on sEMG and angular velocity signals. We testify our method on the data recorded from 8 subjects (5 healthy subjects and 3 patients) who are on isokinetic ankle eversion and inversion. The results show that the mean Cross-correlation coefficients ( $\rho$ ) and the mean normalized root-mean-square deviation (NRMSE %) calculated between the prediction and the real value for isokinetic contraction is 0.966± 0.019 and 7.9% ±0.026. Compared with artificial neural network (ANN) and support vector regression (SVR), the proposed method can predict the joint torque effectively. For the future application, the method has the potential to be employed to predict the ankle moments in real-time application for quantitative lower limb rehabilitation assessment and exoskeleton robot control.

A machine learning approach to reconstruction of heart surface potentials from body surface potentials.

Invasive cardiac catheterisation is a precursor to ablation therapy for ventricular tachycardia. Invasive cardiac diagnostics are fraught with risks. Decades of research has been conducted on the inverse problem of electrocardiography, which can be used to reconstruct Heart Surface Potentials (HSPs) from Body Surface Potentials (BSPs), for non-invasive cardiac diagnostics. State of the art solutions to the inverse problem are unsatisfactory, since the inverse problem is known to be ill-posed. In this paper we propose a novel approach to reconstructing HSPs from BSPs using a Time-Delay Artificial Neural Network (TDANN). We first design the TDANN architecture, and then develop an iterative search space algorithm to find the parameters of the TDANN, which results in the best overall HSP prediction. We use recorded BSPs and HSPs from individuals suffering from serious cardiac conditions to validate our TDANN. The results are encouraging, in that the predicted and recorded HSPs have an average correlation coefficient of 0.7 under diseased conditions.

The Virtual Trackpad: An Electromyography-Based, Wireless, Real-Time, Low-Power, Embedded Hand-Gesture-Recognition System Using an Event-Driven Artificial Neural Network

This brief presents a wireless, low-power embedded system that recognizes hand gestures by decoding surface electromyography (EMG) signals. Ten hand gestures used on commercial trackpads, including pinch, stretch, swipe left, swipe right, scroll up, scroll down, single click, double click, pat, and ok, can be recognized in real time. Features from four differential EMG channels are extracted in multiple time windows. Unlike traditional data segmentation methods, an event-driven method is proposed, with the gesture event detected in the hardware. Feature extraction is triggered only when an event is detected, minimizing computation, memory, and system power. A time-delayed artificial neural network (ANN) is used to predict the gesture from the transient EMG features instead of traditional steady-state features. The ANN is implemented in the microcontroller with a processing time less than 0.2 ms. The detection results are sent wirelessly to a computer. The device weights 15.2 g. A 4.6 g battery supports up to 40 h continuous operation. To our knowledge, this brief shows the first real-time, embedded hand-gesture-recognition system using only transient EMG signals. Experiments with four subjects show that the device can achieve a recognition of ten gestures with an average accuracy of 94%.

Gramophone Noise Detection and Reconstruction Using Time Delay Artificial Neural Networks

Gramophone records were the main recording medium for more than seven decades and regained widespread popularity over the past several years. Being an analog storage medium, gramophone records are subject to distortions caused by scratches, dust particles, degradation, and other means of improper handling. The observed noise often leads to an unpleasant listening experience and requires a filtering process to remove the unwanted disruptions and improve the audio quality. This paper proposes a novel approach that employs various feed forward time delay artificial neural networks to detect and reconstruct noise in musical sound waves. A set of 800 songs from eight different genres were used to validate the performance of the neural networks. The performance was analyzed according to the outlier detection and interpolation accuracy, the computational time and the tradeoff between the accuracy and the time. The empirical results of both detection and reconstruction neural networks were compared to a number of other algorithms, including various statistical measurements, duplication approaches, trigonometric processes, polynomials, and time series models. It was found that the neural networks’ outlier detection accuracy was slightly lower than some of the other noise identification algorithms, but achieved a more efficient tradeoff by detecting most of the noise in real time. The reconstruction process favored neural networks with an increase in the interpolation accuracy compared to other widely used time series models. It was also found that certain genres such as classical, country, and jazz music were interpolated more accurately. Volatile signals, such as electronic, metal, and pop music were more challenging to reconstruct and were substantially better interpolated using neural networks than the other examined algorithms.

Biologically inspired time-delay soft sensors for online monitoring of automotive coldstart operations: a comparative analysis

In this paper, an attempt is made to develop a bio-inspired soft sensor as a real-time identifier for online estimations of the engine-out hydrocarbon emission and exhaust gas temperature over the coldstart operation of an automotive engine. To date, a number of studies have been done to design soft models as offline identifiers for capturing the knowledge of a set of collected coldstart databases. The results indicated that bio-inspired computation can be of great use for the identification of coldstart phenomenon. It also can be a good alternative to classic approaches. However, bio-inspired soft sensors have not been used as online estimators for monitoring the behavior of coldstart phenomenon. This is while it is desirable to have a flexible, inexpensive online identifier which can be employed to detect the real-time behavior of the engine over coldstart periods. Here, an exhaustive numerical study is conducted to demonstrate the potentials of bio-inspired approaches for online measuring of engine-out hydrocarbon emission and exhaust gas temperature during the coldstart operation. To do so, a set of well-known bio-inspired optimizers are combined with a time-delayed artificial neural network to design an evolvable soft sensor. The results clearly demonstrate the high potential of bio-inspired computation for handling the difficulties associated with the real-time identification of engine-out hydrocarbon emission and exhaust gas temperature.

Feasibility of using combined EMG and kinematic signals for prosthesis control: A simulation study using a virtual reality environment

Transhumeral amputation has a significant effect on a person’s independence and quality of life. Myoelectric prostheses have the potential to restore upper limb function, however their use is currently limited due to lack of intuitive and natural control of multiple degrees of freedom. The goal of this study was to evaluate a novel transhumeral prosthesis controller that uses a combination of kinematic and electromyographic (EMG) signals recorded from the person’s proximal humerus. Specifically, we trained a time-delayed artificial neural network to predict elbow flexion/extension and forearm pronation/supination from six proximal EMG signals, and humeral angular velocity and linear acceleration. We evaluated this scheme with ten able-bodied subjects offline, as well as in a target-reaching task presented in an immersive virtual reality environment. The offline training had a target of 4° for flexion/extension and 8° for pronation/supination, which it easily exceeded (2.7° and 5.5° respectively). During online testing, all subjects completed the target-reaching task with path efficiency of 78% and minimal overshoot (1.5%). Thus, combining kinematic and muscle activity signals from the proximal humerus can provide adequate prosthesis control, and testing in a virtual reality environment can provide meaningful data on controller performance.

SEMG-based prediction of masticatory kinematics in rhythmic clenching movements

This paper investigated the ability of a hybrid time-delayed artificial neural network (TDANN)/autoregressive TDANN (AR-TDANN) to predict clenching movements during mastication from surface electromyography (SEMG) signals. Actual jaw motions and SEMG signals from the masticatory muscles were recorded and used as output and input, respectively. Three separate TDANNs/AR-TDANNs were used to predict displacement (in terms of position/orientation), velocity, and acceleration. The optimal number of neurons in the hidden layer and total duration of delays were obtained for each TDANN/AR-TDANN and each subject through a genetic algorithm (GA). The kinematic modeling of a human-like masticatory robot, based on a 6-universal-prismatic-spherical parallel robot, is described. The structure and motion variables of the robot were determined. The closed-form solution of the inverse kinematic problem (IKP) of the robot was found by vector analysis. Thereafter, the framework for an EMG-based human mastication robot interface is explained. Predictions by AR-TDANN were superior to those by TDANN. SEMG signals from mastication muscles contained important information about the mandibular kinematic parameters. This information can be employed to develop control systems for rehabilitation robots. Thus, by predicting the subject's movement and solving the IKP, we provide applicable tools for EMG-based masticatory robot control.

Neural network pattern recognition of lingual-palatal pressure for automated detection of swallow.

We describe a novel device and method for real-time measurement of lingual-palatal pressure and automatic identification of the oral transfer phase of deglutition. Clinical measurement of the oral transport phase of swallowing is a complicated process requiring either placement of obstructive sensors or sitting within a fluoroscope or articulograph for recording. Existing detection algorithms distinguish oral events with EMG, sound, and pressure signals from the head and neck, but are imprecise and frequently result in false detection. We placed seven pressure sensors on a molded mouthpiece fitting over the upper teeth and hard palate and recorded pressure during a variety of swallow and non-swallow activities. Pressure measures and swallow times from 12 healthy and 7 Parkinson's subjects provided training data for a time-delay artificial neural network to categorize the recordings as swallow or non-swallow events. User-specific neural networks properly categorized 96% of swallow and non-swallow events, while a generalized population-trained network was able to properly categorize 93% of swallow and non-swallow events across all recordings. Lingual-palatal pressure signals are sufficient to selectively and specifically recognize the initiation of swallowing in healthy and dysphagic patients.

Comparison of neural networks and Auto regression performance in prediction of Dividend and Price Index

This paper present two models for short and long term stock price forecasting using Artificial Neural Networks and Vector Auto regression analyses as complementary methods. Presented models also will be able to forecast the stock price of future day. The models forecast the stock price of future day with using of previous stock price information. In presented models, it is possible to use the calendar parameters, like the day of week, special days and etc to forecast the stock price. A major contribution of this work is the resulting time-delayed artificial neural network model that allows stock return predictions and is particularly useful as an investment decision support system for hedge funds and other investors, whose portfolios are at risk of losing market value. After designing mentioned models, and specifying construction of multi-layer perception neural network, designed network is trained using available data. Presented method for training the perception neural network is, Error back propagation algorithm, which has variable learning rate and momentum factor, and also is fast. Finally, the models have been applied on Tehran-stock Exchange Price Index (TEPIX) and Tehran- stock Exchange-Dividend and Price Index (TEDPIX) and result are compared together and advantages and disadvantages of these models are described. The results show validity of presented models.

TÜRKİYE’YE YÖNELİK DIŞ TURİZM TALEBİNİN MLP, RBF VE TDNN YAPAY SİNİR AĞI MİMARİLERİ İLE MODELLENMESİ VE TAHMİNİ: KARŞILAŞTIRMALI BİR ANALİZ

Tourism demand forecasting plays important role in public and private sector officials’ future planning activities. The aim of this study is modeling inbound tourism demand to Turkey by Feed Forward-Back Propagation (MLP), Radial Basis Function (RBF) and Time Delay (TDNN) artificial neural network architectures and forecasting monthly tourism demand for 2013 via the model providing the highest accuracy.In this study, it is used the total number of foreign tourist arrivals as a measure of inbound tourism demand and monthly foreign tourist arrivals to Turkey in the period of January 1987 – December 2012 were utilized. In the process of modeling the data by ANN’s, it is analyzed the forecasting performance of different network architectures constituted by both unpreprocessed raw data and seasonally adjusted data taking into account of the distinct viewpoints in related studies. As a consequence of several attempts, it has been observed that 12 lagged MLP model which has [4-5-1-] architechture has presented best forecasting performance. By this model it has been produced monthly inbound tourism demand forecasts to Turkey for year 2013.

Surface Electrocardiogram Reconstruction From Intracardiac Electrograms Using a Dynamic Time Delay Artificial Neural Network

This study proposes a method to facilitate the remote follow up of patients suffering from cardiac pathologies and treated with an implantable device, by synthesizing a 12-lead surface ECG from the intracardiac electrograms (EGM) recorded by the device. Two methods (direct and indirect), based on dynamic time-delay artificial neural networks (TDNNs) are proposed and compared with classical linear approaches. The direct method aims to estimate 12 different transfer functions between the EGM and each surface ECG signal. The indirect method is based on a preliminary orthogonalization phase of the available EGM and ECG signals, and the application of the TDNN between these orthogonalized signals, using only three transfer functions. These methods are evaluated on a dataset issued from 15 patients. Correlation coefficients calculated between the synthesized and the real ECG show that the proposed TDNN methods represent an efficient way to synthesize 12-lead ECG, from two or four EGM and perform better than the linear ones. We also evaluate the results as a function of the EGM configuration. Results are also supported by the comparison of extracted features and a qualitative analysis performed by a cardiologist.

Time–delay single layer artificial neural network models for estimating shelf life of burfi

Time-delay single layer artificial neural network models were developed for estimating the shelf life of burfi stored at 30oC. Input variables for the models were moisture, titratable acidity, free fatty acids, tyrosine, and peroxide value, while the overall acceptability score was taken as output variable. Mean square error, root mean square error, coefficient of determination and Nash - sutcliffo coefficient were applied in order to compare the prediction ability of the Time-delay single layer models. The combination of 5-1-1 showed a very high correlation between the training and validation data, establishing that the developed models are effective in predicting the shelf life of burfi.

Shelf Life Estimation of Processed Cheese by Artificial Neural Network Expert Systems

Time–delay artificial neural network (ANN) single layer and multilayer artificial models were developed for predicting the shelf life of processed cheese stored at 7-8o C. Soluble nitrogen, pH; standard plate count, yeast & mould count, and spore count were input variables, and sensory score was output variable. The results showed excellent agreement between training and validation data with high coefficient of determination and nash sutcliffo coefficient, thus suggesting that the developed models are good for predicting the shelf life of processed cheese.

Predicting stock market returns from malicious attacks: A comparative analysis of vector autoregression and time-delayed neural networks

With the growing importance of Internet-based businesses, malicious code attacks on information technology infrastructures have been on the rise. Prior studies have indicated that these malicious attacks are associated with detrimental economic effects on the attacked firms. On the other hand, we conjecture that more intense malicious attacks boost the stock price of information security firms. Furthermore, we use artificial neural networks and vector autoregression analyses as complementary methods to study the relationship between the stock market returns of information security firms and the intensity of malicious attacks, computed as the product of the number of malicious attacks and their severity levels. A major contribution of this work is the resulting time-delayed artificial neural network model that allows stock return predictions and is particularly useful as an investment decision support system for hedge funds and other investors, whose portfolios are at risk of losing market value during malicious attacks.

Stress-Strain Behavior of Geomaterials in Loading Reversal Simulated by Time-Delay Neural Networks

The effectiveness of time-delay artificial neural networks (TDANNs) as highly nonlinear mapping tools for simulation of hysteresis stress-strain (σ-e) response of geomaterials under repeated reversal loading is investigated. A nonlinear recursive simulator containing the developed TDANNs was designed to enable forecasting of complete σ-e curves from the knowledge of only the initial σ-e condition of the tested material. Theoretically-obtained data were used to derive the TDANNs, and simulations were performed to validate the proposed approach. Advantages, limitations, and method of enhancement of the approach are presented.