Search

Nowadays brushless DC motors (BLDCMs) are becoming indispensable components as the electrification revolution in the mobility industry is happening. Electric kick scooters, so-called e-scooters, are among these micro-mobility vehicles which are powered by these motors. Due to the uncertain and nonlinear features, the controller performance developed for these motors degrades. For these reasons, a chattering-reduced cascaded Sliding Mode Control (SMC) scheme to effectively track reference motor speed in the outer loop by eliminating torque ripples in the inner loop current control was designed. Field-oriented Control (FOC) methodology was used to implement the SMC in the BLDCM. An exponential reaching law algorithm was proposed for sliding surfaces of the inner and outer loop controllers. The suitability and performance of electric scooter-hub motors were analyzed in terms of traction control. A cascaded speed and torque controller produced significantly favorable results representing minimized torque and current ripples, and operation over a wide speed range.

Modelling errors and robust stabilization/tracking problems under parameter and model uncertainties complicate the control of the flexible underactuated systems. Chattering-free sliding-mode-based input-output control law realizes robustness against the structured and unstructured uncertainties in the system dynamics and avoids the excitation of unmodeled dynamics. The main purpose of this paper was to propose a robust adaptive solution for stabilizing and tracking direct-drive (DD) flexible robot arms under parameter and model uncertainties, as well as external disturbances. A lightweight robot arm subject to external and internal dynamic effects was taken into consideration. The challenges were compensating actuator dynamics with the inverter switching effects and torque ripples, stabilizing the zero dynamics under parameter/model uncertainties and disturbances while precisely tracking the predefined reference position. The precise control of this kind of system demands an accurate system model and knowledge of all sources that excite unmodeled dynamics. For this purpose, equations of motion for a flexible robot arm were derived and formulated for the large motion via Lagrange’s method. The goals were determined to achieve high-speed, precise position control, and satisfied accuracy by compensating the unwanted torque ripple and friction that degrades performance through an adaptive robust control approach. The actuator dynamics and their effect on the torque output were investigated due to the transmitted torque to the load side. The high-performance goals, precision and robustness issues, and stability concerns were satisfied by using robust-adaptive input-output linearization-based control law combining chattering-free sliding mode control (SMC) while avoiding the excitation of unmodeled dynamics. The following highlights are covered: A 2-DOF flexible robot arm considering actuator dynamics was modelled; the theoretical implication of the chattering-free sliding mode-adaptive linearizing algorithm, which ensures robust stabilization and precise tracking control, was designed based on the full system model including actuator dynamics with computer simulations. Stability analysis of the zero dynamics originated from the Lyapunov theorem was performed. The conceptual design necessity of nonlinear observers for the estimation of immeasurable variables and parameters required for the control algorithms was emphasized.

Human-machine interfaces contribute to the improvement of the life quality of physically disabled users. In this study, a non-invasive brain-machine interface (BMI) design methodology was proposed to control a robot arm through magnetoencephalography (MEG) based on directionally modulated MEG activity that was acquired during the user’s imagined wrist movements in four various directions. The partial directed coherence (PDC) measure derived from functional connectivity between cortical brain regions was utilized in the feature extraction process. The time-varying parameters were estimated based on a time-varying multivariate adaptive autoregressive (AAR) model, that can detect task-dependent features and non-symmetric channel relevance for mental task discrimination. An extreme learning machine (ELM), that utilizes Moore-Penrose (MP) generalized inverse to set its weights and does not necessitate a gradient-based backpropagation algorithm was employed to generate a model with the extracted feature set. The output of the task classification model was embedded into the robotic arm model for realizing control-based tasks. The classification results dictate that the proposed BMI methodology is a feasible solution for rehabilitation or assistance systems that are devised to help motor-impaired people. The proposed methodology provides very satisfactory classification performance at a fast learning speed.

The collector and brushless electronic commutation machines based on the working principle of the direct current machines have been widely used in industrial applications through the help of the developments in power electronics, microelectronics, permanent magnets, microprocessors&control, digital signal processing technologies, etc. Internal permanent magnet synchronous motors (IPMSMs) are used in increasing numbers due to their advantages such as high torque/current and torque/inertia, robust construction, high efficiency, reliability, etc. The problems brought by position sensors, especially in terms of application, performance, mass production, and cost, have made sensorless control a necessity in drive systems and applications.This paper presents a backstepping control method for speed sensorless IPMSM based on an extended Kalman filter (EKF). First, a comprehensive nonlinear dynamical model of the IPMSM in the direct and quadrature ( ) rotor frame is derived and its state-space representation is obtained. Then, the rotor speed and current tracking backstepping controllers are designed to achieve precise tracking and anti-disturbance performance. The designed controllers are embedded into the field-oriented control (FOC) scheme. The asymptotic stability condition for the backstepping controller is guaranteed through the Lyapunov stability theorem. Finally, An EKF is designed for estimating the immeasurable mechanical parameters of IPMSM and tracking the system states in a finite time with high steady-state precision. The effectiveness of the proposed methodology is proved by conducting simulations having various dynamic operating conditions such as sudden torque load change, command speed change, and parameter variation.

Nowadays brushless DC motors (BLDCMs) are becoming indispensable components as the electrification revolution in the mobility industry is happening. Electric kick scooters, so-called e-scooters, are among these micro-mobility vehicles which are powered by these motors. Due to the uncertain and nonlinear features, the controller performance developed for these motors degrades. For these reasons, a chattering-reduced cascaded Sliding Mode Control (SMC) scheme to effectively track reference motor speed in the outer loop by eliminating torque ripples in the inner loop current control was designed. Field-oriented Control (FOC) methodology was used to implement the SMC in the BLDCM. An exponential reaching law algorithm was proposed for sliding surfaces of the inner and outer loop controllers. The suitability and performance of electric scooter-hub motors were analyzed in terms of traction control. A cascaded speed and torque controller produced significantly favorable results representing minimized torque and current ripples, and operation over a wide speed range.

Automatic detection of Attention Deficit Hyperactivity Disorder (ADHD) based on the functional Magnetic Resonance Imaging (fMRI) through Deep Learning (DL) is becoming a quite useful methodology due to the curse of-dimensionality problem of the data is solved. Also, this method proposes an invasive and robust solution to the variances in data acquisition and class distribution imbalances. In this paper, a transfer learning approach, specifically ResNet-50 type pre-trained 2D-Convolutional Neural Network (CNN) was used to automatically classify ADHD and healthy children. The results demonstrated that ResNet-50 architecture with 10-k cross-validation (CV) achieves an overall classification accuracy of 93.45%. The interpretation of the results was done via the Class Activation Map (CAM) analysis which showed that children with ADHD differed from controls in a wide range of brain areas including frontal, parietal and temporal lobes.

The human brain is characterized by complex structural, functional connections that integrate unique cognitive characteristics. There is a fundamental hurdle for the evaluation of both structural and functional connections of the brain and the effects in the diagnosis and treatment of neurodegenerative diseases. Currently, there is no clinically specific diagnostic biomarker capable of confirming the diagnosis of major depressive disorder (MDD). Therefore, exploring translational biomarkers of mood disorders based on deep learning (DL) has valuable potential with its recently underlined promising outcomes. In this article, an electroencephalography (EEG)-based diagnosis model for MDD is built through advanced computational neuroscience methodology coupled with a deep convolutional neural network (CNN) approach. EEG recordings are analyzed by modeling 3 different deep CNN structure, namely, ResNet-50, MobileNet, Inception-v3, in order to dichotomize MDD patients and healthy controls. EEG data are collected for 4 main frequency bands (Δ, θ, α, and β, accompanying spatial resolution with location information by collecting data from 19 electrodes. Following the pre-processing step, different DL architectures were employed to underline discrimination performance by comparing classification accuracies. The classification performance of models based on location data, MobileNet architecture generated 89.33% and 92.66% classification accuracy. As to the frequency bands, delta frequency band outperformed compared to other bands with 90.22% predictive accuracy and area under curve (AUC) value of 0.9 for ResNet-50 architecture. The main contribution of the study is the delineation of distinctive spatial and temporal features using various DL architectures to dichotomize 46 MDD subjects from 46 healthy subjects. Exploring translational biomarkers of mood disorders based on DL perspective is the main focus of this study and, though it is challenging, with its promising potential to improve our understanding of the psychiatric disorders, computational methods are highly worthy for the diagnosis process and valuable in terms of both speed and accuracy compared with classical approaches.

Electroencephalography (EEG) signals are known to be nonstationary and often multicomponential signals containing information about the condition of the brain. Since the EEG signal has complex, nonlinear, nonstationary, and highly random behaviour, numerous linear feature extraction methods related to the short-time windowing technique do not satisfy higher classification accuracy. Since biosignals are highly subjective, the symptoms may appear at random in the time scale and very small variations in EEG signals may depict a definite type of brain abnormality it is valuable and vital to extract and analyze the EEG signal parameters using computers. The challenge is to design and develop signal processing algorithms that extract this subtle information and use it for diagnosis, monitoring, and treatment of subjects suffering from psychiatric disorders. For this purpose, finite impulse response–based filtering process was employed rather than traditional time and frequency domain methods. Finite impulse response subbands were analyzed further to obtain feature vectors of different entropy markers and these features were fed into a classifier namely multilayer perceptron. The performances of the classifiers were finally compared considering overall classification accuracies, area under receiver operating characteristic curve scores. Our results underline the potential benefit of the introduced methodology is promising and is to be treated as a clinical interface in dichotomizing substance use disorders subjects and for other medical data analysis studies. The results also indicate that entropy estimators can distinguish normal and opioid use disorder subjects. EEG data and theta frequency band have distinctive capability for almost all types of entropies while nonextensive Tsallis entropy outperforms compared with other types of entropies.

Train safety and operational efficiency can be improved by investigating the dynamics of the train under varying conditions. Longitudinal train dynamics (LTD) simulations performed for such purposes, usually by utilising a nonlinear time-domain model. This paper covers two modes of LTD results corresponding to the time domain and frequency domain analysis. Time-domain solutions are essential to evaluate the full response used for parameter optimisation and controller design studies while frequency domain solutions can provide significant but straightforward clues regarding system dynamics. An advanced draft gear model, which works under a four-stage process is constructed considering all structural components, geometric relationships, friction modelling and dynamic characteristics such as hysteresis, stiffening, state transition, locked unloading, softening. Then, this model is parametrically reduced and implemented into an LTD simulation. The simulation in the time domain is conducted assuming a locomotive connected with a nine wagon via “ode3” fixed-step solver. The transfer function among the first wagon acceleration (output) and the locomotive force (input) estimated through system identification methodology. Then, the identification results interpreted by investigating step-response characteristic and best response giving parameter set is selected. Next, the modal and spectral analysis performed to reveal the behaviour of the in-train forces and the effects of vibration. This paper discusses a reliable methodology for the longitudinal dynamic analysis of the multi-bodied train in time and frequency domain and clarifies in-train vibration behaviour under the existence of sophisticated draft gear.

Event-related mental task information collected from electroencephalography (EEG) signals, which are functionally related to different brain areas, possesses complex and non-stationary signal features. It is essential to be able to classify mental task information through the use in brain-computer interface (BCI) applications. This paper proposes a wavelet packet transform (WPT) technique merged with a specific entropy biomarker as a feature extraction tool to classify six mental tasks. First, the data were collected from a healthy control group and the multi-signal information comprised six mental tasks which were decomposed into a number of subspaces spread over a wide frequency spectrum by projecting six different wavelet basis functions. Later, the decomposed subspaces were subjected to three entropy-type statistical measure functions to extract the feature vectors for each mental task to be fed into a backpropagation time-recurrent neural network (BPTT-RNN) model. Cross-validated classification results demonstrated that the model could classify with 85% accuracy through a discrete Meyer basis function coupled with a Renyi entropy biomarker. The classifier model was finally tested in the Simulink platform to demonstrate the Fourier series representation of periodic signals by tracking the harmonic pattern. In order to boost the model performance, ant colony optimization (ACO)-based feature selection method was employed. The overall accuracy increased to 88.98%. The results underlined that the WPT combined with an entropy uncertainty measure methodology is both effective and versatile to discriminate the features of the signal localized in a time-frequency domain.

Electroencephalogram (EEG)-based automated depression diagnosis systems have been suggested for early and accurate detection of mood disorders. EEG signals are highly irregular, nonlinear, and nonstationary in nature and are traditionally studied from a linear viewpoint by means of statistical and frequency features. Since, linear metrics present certain limitations and nonlinear methods have proven to be an efficient tool in understanding the complexities of the brain in the identification of underlying behavior of biological signals, such as electrocardiogram, EEG and magnetoencephalogram and thus, can be applied to all nonstationary signals. Various nonlinear algorithms can be used in the analysis of EEG signals. In this research paper, we aim to develop a novel methodology for EEG-based depression diagnosis utilizing 2 advanced computational techniques: frequency-domain extended multivariate autoregressive (eMVAR) and deep learning (DL). We proposed a hybrid method comprising a pretrained ResNet-50 and long-short term memory (LSTM) to capture depression-specific information and compared with a strong conventional machine learning (ML) framework having eMVAR connectivity features. The following 8 causality measures, which interpret the interaction mechanisms among spectrally decomposed oscillations, were used to extract features from multivariate EEG time series: directed coherence (DC), directed transfer function (DTF), partial DC (PDC), generalized PDC (gPDC), extended DC (eDC), delayed DC (dDC), extended PDC (ePDC), and delayed PDC (dPDC). The classification accuracies were 84% with DC, 85% with DTF, 95.3% with PDC, 95.1% with gPDC, 84.8% with eDC, 84.6% with dDC, 84.2% with ePDC, and 95.9% with dPDC for the eMVAR framework. Through a DL framework (ResNet-50 + LSTM), the classification accuracy was achieved as 90.22%. The results demonstrate that our DL methodology is a competitive alternative to the strong feature extraction-based ML methods in depression classification.

Recent studies underline the contribution of brain-computer interface (BCI) applications to the enhancement process of the life quality of physically impaired subjects. In this context, to design an effective stroke rehabilitation or assistance system, the classification of motor imagery (MI) tasks are performed through deep learning (DL) algorithms. Although the utilization of DL in the BCI field remains relatively premature as compared to the fields related to natural language processing, object detection, etc., DL has proven its effectiveness in carrying out this task. In this paper, a hybrid method, which fuses the one-dimensional convolutional neural network (1D CNN) with the long short-term memory (LSTM), was performed for classifying four different MI tasks, i.e. left hand, right hand, tongue, and feet movements. The time representation of MI tasks is extracted through the hybrid deep learning model training after principal component analysis (PCA)-based artefact removal process. The performance criteria given in the BCI Competition IV dataset A are estimated. 10-folded Cross-validation (CV) results show that the proposed method outperforms in classifying electroencephalogram (EEG)-electrooculogram (EOG) combined motor imagery tasks compared to the state of art methods and is robust against data variations. The CNN-LSTM classification model reached 95.62 % (±1.2290742) accuracy and 0.9462 (±0.01216265) kappa value for datasets with four MI-based class validated using 10-fold CV. Also, the receiver operator characteristic (ROC) curve, the area under the ROC curve (AUC) score, and confusion matrix are evaluated for further interpretations.

WOS: 000437469500001, The derailment model of a 51-DOF railway vehicle including coupling effects of the longitudinal and lateral modes was theoretically built by investigating the geometrical and dynamical effects of the lateral acceleration, gyro factors and mechanical factors such as flange angle, friction coefficient, effective radius of the wheel and track gauge. The lateral dynamic of the railway vehicle comprising lateral, vertical displacement and roll, yaw, pitch angular displacements of each six wheelset, three bogie frames, and vehicle body was modeled in detail. Depending on this model, the initiation of different kind of derailments such as wheel lifting, wheel climbing, roll-over and their synthesis can be predicted. In addition, the effects of vehicle speed on derailment quotient (DQ)'s were investigated under various suspension parameters and curved track radius. Main objective of the development of such a numerical model is to analyze various dynamical and geometrical influences on the wheelset, which is not considered in conventional derailments models such as those based on Nadal and Weinstock criteria.

Adhesion coefficient and the resultant normal force occurred at the wheel-rail contact determine braking and traction forces in railway applications. Due to the limits on controlling the resultant normal force, maximization of the adhesion coefficient is the only way to obtain larger braking and tractive works. There are various advantages of utilization of adhesion in an efficient way, such as reducing operating costs, minimizing trip time, preventing wheel-rail wear. On the other hand, the adhesion mechanism at the wheel-rail contact has a highly non-linear complex nature, whose dynamics are changed as a function of parameters like environmental conditions, vehicle speed, slip ratio etc. There is not any satisfactory accurate and trustworthy way of estimating these parameters yet. In this paper, an event based adaptive control scheme has been introduced to maximize the adhesion coefficient without requiring the exact value of those parameters. The efficient adhesion utilization can be obtained by using the proposed method while maintaining the stability. The continuous excitement of traction system and slow recuperation detection time difficulties in the former research has been overcome. The dynamics of phase shift were analyzed and an adaptive structure were built. Results acquired by using the proposed adaptive method were compared with the conventional control scheme in "Matlab&Simulink" software under various driving scenarios and wheel-rail contact conditions.

Background: Although attention-deficit/hyperactivity disorder (ADHD) symptoms decrease with the factors such as age, many individuals keep suffering from the disorder in adolescence and adulthood. Objective: In this paper, a deep learning algorithm was built to forecast the prognosis of ADHD, using the patient's clinical features, the measurement of their response to treatment, and the degree of improvement seen after six years of follow-up. Participants and Settings: The clinical findings such as socio-demographic data of 433 patients followed by the child and adolescent psychiatry department for an average of 6 years with diagnoses of ADHD, and ADHD sub-type, accompanying oppositional/conduct disorders, other psychiatric and organic disorders, the effectiveness of psychotherapy and medication on attention, academic status, and behavioral problems were used to help predict prognosis. Methods: A deep neural network (DNN) learning-based regressor was used to make a prediction model. Results: The results obtained from the DNN regression model achieved accurate predictions for all outputs. The mean error for all outputs was evaluated as mean-squared error (mse) and 0.0068 mean-absolute error (mae), respectively. The R-value was evaluated as 0.91316. It was proven that the model prediction power was adequate as tested with these metrics. Conclusions: This methodology improves the prediction of ADHD prognosis and provides a robust predictive model. Studies with larger samples may support the results.

Excessive driving force applied to the trains leads to an inadequate utilization of the adhesion phenomenon occurred at the wheel–rail contact, and an unnecessary power consumption, while inadequate driving force causes the train to run inefficiently. For this reason, the necessity of re-adhesion control in the safe and reliable operation, in the balance of energy consumption, is indisputable. A comparison of the two re-adhesion control strategies, one of which is robust adaptive and the other of which is the modified super-twisting sliding mode, has been presented in this article. These control algorithms developed suppress the wheel slip on time and maintain optimal traction performance after re-adhesion under the nonlinear properties of the traction system and the uncertainties of the adhesion level at the wheel–rail interface. Due to the complex nonlinear relationship between the adhesion force and the slip angular velocity, such a problem becomes a hard problem to overcome as long as the optimal slip ratio is not known. An optimal search strategy has also been developed to estimate and to track the desired slip angular velocity. By means of the proposed strategies, the traction motor control torque is automatically adjusted so as to ensure that the train operates away from the unstable slip zone but adjacent to the optimal adhesion region, and the desired traction capability is attainable once adhesion is regained. Mathematical analyzes are also provided to ensure the ultimate boundedness of the algorithms developed. The effectiveness of the proposed re-adhesion strategies is validated through the theoretical analysis and numerical simulations conducted in MATLAB and Simulink. As a result of consecutive simulations, modified super-twisting algorithm has shown better performance as compared to the robust adaptive one in tracking the optimal slip velocity as wheel–rail contact conditions switch suddenly.

Increasing the traction force is a complex problem in the design of railway vehicles; therefore, effective traction systems and algorithms have to be developed. During the traction process, the verification of traction algorithms and control strategies are based on simulations covering all locomotive dynamics. In this article, traction model of a railway vehicle and re-adhesion control method based on simulation approach are investigated to obtain more effective results. The longitudinal dynamic of a railway vehicle having traction system, which comprises two parallel motor groups, each of which has two field-oriented induction motor connected in series, is simulated to examine time-dependent changes in motor stator currents, traction torque, adhesion and resistance forces according to a given speed reference. The interaction between the adhesion force and the slip ratio is established according to the Burckhardt adhesion model, and a modified super-twisting sliding mode slip control is implemented in a computer simulation under various contact conditions so that simulation results approve the presented control method works under the maximum adhesion force. The comparison between the classical and modified version of the proposed control strategy was made to better evaluate the performance of the control system and to better optimize the traction system.

It is a critical issue to maintain stability in high-speed railway vehicles and to ensure comfortable and safe driving. Multi-body models of railway vehicles have non-linear properties originated from the wheel–rail contact and characteristics of the suspension systems. The critical speed values at which the unstable oscillations and the amplitudes of the limit cycle-type vibrations take place vary by adjusting the design parameters; therefore, these effects on non-linear railway dynamics must be evaluated with a higher precision by using numerical and/or analytical methods to determine the bifurcation behavior. The main objective of this paper is to examine the non-linear phenomena in a railway bogie from a broad perspective, concentrating on non-linear analysis methods. Thus, non-linear equations of motion of a 12-degrees of freedom railway bogie involving dual wheelsets, non-linear wheel flange contact, heuristic non-linear creep model, and suspension system are solved in the time domain with small time steps by using ode23s (stiff/Mod.Rosenbrock) method. The critical speeds were calculated with respect to the effects of various lateral stiffness and damping coefficients. The bifurcation diagrams of the maximum lateral displacement of the leading wheelset were depicted within a wide speed range. In the case of the suspension parameter set where the subcritical/supercritical Hopf bifurcation takes place, the phase portraits and the symmetric/asymmetric oscillations of the leading wheelset at the critical speed were represented. The type of the Hopf bifurcation can be transformed from the subcritical state to the supercritical state by increasing the given suspension ratio. The Lyapunov exponents of the lateral displacement, lateral velocity, yaw angular displacement, and yaw angular velocity of the leading wheelset were evaluated above the critical speeds to examine chaotic motion. The effect of the suspension parameters on the non-linear dynamical behavior of the railway bogie at the stability limit and on the bifurcation type has been proved.

Nonlinear dynamics connect the neurons that form the brain, and thus, the complex information is produced and transported. The function of the neurons and the problem of understanding the dynamics of the brain has been the research area of mathematical neuroscience. In this study, the modelling and simulation of the propagation of the electric field based Action Potential (AP) on the Two Dimensional (2-D) field of axon network, whose matrix consists of 128×128 electrically coupled neurons were done using nonlinear Spatial FitzHugh Nagumo (SFN) equations. SFN equations are a particular class of Partial Differential Equation’s (PDE’s) exhibiting travelling wave behaviour occurred in neuron systems. The motivation of this paper is to evaluate the SFN equation, which is a special kind of the time-dependent nonlinear reaction-diffusion problem governing neuron dynamics numerically in 2-D space addressed by investigating the Polynomial-based Differential Quadrature Method (PDQM) having Chebyshev-Gauss-Lobatto quadrature points. The solution occurs as elliptical spiral waves induced by electrical stimulation. Thus, the neuronal system behaviour and the interaction with the specific type of Boundary Conditions (BC’s) are predicted. The space derivatives are discretised through PDQM. In this way, the problem is reduced into a system of first-order non-linear differential equations. Hereafter the time derivatives of the SFN equation are solved through the Finite Difference Method (FDM). The various dynamical behaviour that governs the travelling wave pattern regarding the Initial Condition’s (IC’s), BC’s and way of stimulation of the neuron model examined in details. Numerical results indicated that the proposed PDQM provide reliable, fast, and efficient solutions.

Many endogenous and external components may affect the physiological, mental and behavioral states in humans. Monitoring tools are required to evaluate biomarkers, identify biological events, and predict their outcomes. Being one of the valuable indicators, brain biomarkers derived from temporal or spectral electroencephalography (EEG) signals processing, allow for the classification of mental disorders and mental tasks. An EEG signal has a nonstationary nature and individual frequency feature, hence it can be concluded that each subject has peculiar timing and data to extract unique features. In order to classify data, which are collected by performing four mental task (reciting the alphabet backwards, imagination of rotation of a cube, imagination of right hand movements (open/close) and performing mathematical operations), discriminative features were extracted using four competitive time-frequency techniques; Wavelet Packet Decomposition (WPD), Morlet Wavelet Transform (MWT), Short Time Fourier Transform (STFT) and Wavelet Filter Bank (WFB), respectively. The extracted features using both time and frequency domain information were then reduced using a principal component analysis for subset reduction. Finally, the reduced subsets were fed into a multi-layer perceptron neural network (MP-NN) trained with back propagation (BP) algorithm to generate a predictive model. This study mainly focuses on comparing the relative performance of time-frequency feature extraction methods that are used to classify mental tasks. The real-time (RT) conducted experimental results underlined that the WPD feature extraction method outperforms with 92% classification accuracy compared to three other aforementioned methods for four different mental tasks.

URL: http://sujest.selcuk.edu.tr/sumbtd/article/view/436 DOI: 10.15317/Scitech.2017.103, Monitoring the adhesion force between a railway wheel and a rail surface is very essential in maintaining high acceleration and braking performance of railway vehicles. Due to the difficulties encountered in direct measurement of friction coefficient, creepage and adhesion force; state observers are used as indirect estimation methods. This paper proposes an effective estimation method, which exploits railway traction motor behaviour to give an assistance for realizing wheel slip and adhesion control in order to be used in railway applications. This method plays an active role in optimizing the use of the existing adhesion and reducing wheel wear by decreasing high creep values. With this method, adhesion force can be indirectly estimated by measuring stator currents, and angular speed of the AC traction motor and using dynamic relationships based on the extended Kalman filter (EKF) simulation model. The re-adhesion controller can be designed to regulate the motor torque command according to the maximum available adhesion depending on the estimated results. To test the proposed method, simulations were performed under different friction coefficients., Bir demiryolu tekerleği ile rayı arasında meydana gelen tutunma kuvvetinin izlenmesi, demiryolu araçlarının yüksek hızlanma ve frenleme performansının korunmasında oldukça önemlidir. Sürtünme katsayısı, kayma ve tutunma kuvvetinin doğrudan ölçülmesinde karşılaşılan zorluklardan dolayı, durum gözetleyicilerine dayalı dolaylı tahmin yöntemleri kullanılır. Bu makale, demiryolu uygulamalarında kullanılmak üzere tekerlek kayma ve yeniden tutunma kontrolünü gerçekleştirmek için demiryolu cer motor davranışını kullanan etkili bir tahmin yöntemi önermektedir. Bu yöntem, mevcut tutunmanın kullanımını iyileştirmede ve yüksek kayma değerlerini düşürerek tekerlek aşınmasının azaltılmasında etkin bir rol oynamaktadır. Bu yöntemle, stator akımları ve asenkron cer motorun açısal hızı ölçülerek, genişletilmiş Kalman filtresi (GKF) simülasyon modeline dayanan dinamik ilişkiler kullanılarak tutunma kuvveti dolaylı olarak tahmin edilebilir. Yeniden tutunma kontrolörü, tahmin sonuçlarına bağlı olan maksimum erişilebilir tutunma özelliklerine göre motor moment komutu düzenlenerek tasarlanabilir. Önerilen yöntemi test etmek için, farklı tekerlek-ray sürtünme katsayıları altında simülasyonlar gerçekleştirilmiştir

In order to make effective works with the mobile robot and maximize its working performance, it is necessary to estimate and track the current pose of the mobile robot. In this paper, under the assumption that the initial pose, kinematics and environmental model of a mobile robot are known, the localization and tracking of the mobile robot's position and orientation have been carried out. The odometry model with the problem of accumulation of unlimited errors is used for tracking the pose, and sensor fusion algorithms are applied to solve this problem. By using the odometry and laser range finder model, the Extended Kalman Filter (EKF), Unscented Kalman Filter (UKF), Unscented Information Filter (UIF), Extended Information Filter (EIF) algorithms were tested on a graphical user interface (GUI) based on occupancy grid maps as an environment model, respectively. In this context, the pose tracking and estimation performances of the non-linear model based estimators are compared to each other. Since occupancy grid maps are utilized, only the laser range finder measurement uncertainty should be considered unlike feature based maps. In this way, the computational complexity can be reduced. When the simulation results are evaluated, it is determined that the Extended Information Filter algorithm has expressed more stable performance in terms of the mobile robot pose estimation.

The nonlinear dynamic characteristics of a railway vehicle are checked into thoroughly by applying two different wheel-rail contact model: a heuristic nonlinear friction creepage model derived by using Kalker ’s theory and Polach model including dead-zone clearance. This two models are matched with the quasi-static form of the LuGre model to obtain more realistic wheel-rail contact model. LuGre model parameters are determined using nonlinear optimization method, which it’s objective is to minimize the error between the output of the Polach and Kalker model and quasi-static LuGre model for specific operating conditions. The symmetric/asymmetric bifurcation attitude and stable/unstable motion of the railway vehicle in the presence of nonlinearities which are yaw damping forces in the longitudinal suspension system are analyzed in great detail by changing the vehicle speed. Phase portraits of the lateral displacement of the leading wheelset of the railway vehicle are drawn below and on the critical speeds, where sub-critical Hopf bifurcation take place, for two wheel-rail contact model. Asymmetric periodic motions have been observed during the simulation in the lateral displacement of the wheelset under different vehicle speed range. The coexistence of multiple steady states cause bounces in the amplitude of vibrations, resulting instability problems of the railway vehicle. By using Lyapunov’s indirect method, the critical hunting speeds are calculated with respect to the radius of the curved track parameter changes. Hunting, which is defined as the oscillation of the lateral displacement of wheelset with a large domain, is described by a limit cycle-type oscillation nature. The evaluated accuracy of the LuGre model adopted from Kalker’s model results for prediction of critical speed is higher than the results of the LuGre model adopted from Polach’s model. From the results of the analysis, the critical hunting speed must be resolved by investigating the track tests under various kind of excitations.

The main purpose of this paper is to analyze and compare the Hopf bifurcation behavior of a two-axle railway bogie and a dual wheelset in the presence of nonlinearities, which are yaw damping forces in the longitudinal suspension system and heuristic creep model of the wheel-rail contact including dead-zone clearance, while running on a curved track. Two-axle railway bogie and dual wheelset were modeled using 12-DOF and 8-DOF system with considering lateral, vertical, roll, and yaw motions. By utilizing Lyapunov’s indirect method, the critical hunting speeds related to these models are evaluated as track radius changes. Hunting defined as the lateral vibration of the wheelset with a large domain was characterized by a limit cycle-type oscillation behavior. Influence of the curved track radius on the lateral displacement of the leading wheelset was also investigated through 2D bifurcation diagram, which is employed in the design of a stable model. Frequency power spectra at critical speeds, which are related to the subcritical and supercritical bifurcations, were represented by comparing the two-axle bogie and dual wheelset model. The evaluated accuracy to predict the critical hunting speed is higher and the hunting frequency in unstable region is lower compared to the dual wheelset model.

An advanced train model, which examines hunting instability and derailment in one integrated model implicitly is presented. In terms of a control subject, proposed model is compatible with nonlinear controllers to stabilise hunting oscillations and perform real-time derailment avoidance. The dynamical model, which consists of a vehicle body, two bogie frames, and two wheelsets in each bogie frame was modelled with 35-DOF. Heuristic nonlinear creep model and flange-rail contact model were used to reveal the effects of the creep forces and moments. The eigenvalues at the hunting speed were calculated by the assistance of the Gershgorin disc theorem. The vehicle speed influence on evaluated derailment quotient was investigated at a sharp radius of the curved track. Safe speeds were also estimated via active derailment criteria. The main superiority of the proposed model is that one can both predict incipient derailment actively and also determine nonlinear critical hunting speeds with higher precision.

Wavelet theory is a widely used feature extraction method for raw electroencephalogram (EEG) signal processing. The nature of the EEG signal is non-stationary, therefore applying wavelet transform on EEG signals is a valuable process for extraction promising features. On the other hand, determining the proper wavelet family is a challenging step to get the best fitted features for high classification accuracy. In this paper, therefore, we focused on a comparative study of different Discrete Wavelet Transform (DWT) methods to find the most convenient wavelet function of wavelet families for a non-stationary EEG signal analysis to be used to classify mental tasks. For the classification process, four different mental tasks were selected to and we grouped each with another one to set dual tasked sets including all possible combinations. Feature extraction steps are performed using wavelet functions haar, coiflets (order 1), biorthogonal (order 6.8), reverse biorthogonal (order 6.8), daubechies (order 2) and, daubechies (order 4). Later, a specific feature reduction formula is applied to the extracted feature vector. Generated feature vector is then split into train and test data before the classification. Artificial neural network was used for classification of the extracted feature sets. From the result of the repeated analysis for each DWT methods, Coiflets performed relatively better compared to other wavelet families.