Road profile estimation using wavelet neural network and 7-DOF vehicle dynamic systems

Size: px
Start display at page:

Download "Road profile estimation using wavelet neural network and 7-DOF vehicle dynamic systems"

Transcription

1 Journal of Mechanical Science and Technology 26 (10) (2012) 3029~ DOI /s x Road profile estimation using wavelet neural network and 7-DOF vehicle dynamic systems Ali Solhmirzaei 1,*, Shahram Azadi 2 and Reza Kazemi 2 1 Technical and Engineering Department, Mapna Locomotive Company, Mapna Group, Tehran, Iran 2 Department of Mechanical Engineering, K. N. Toosi University, Tehran, Iran (Manuscript Received October 1, 2011; Revised March 18, 2012; Accepted May 2, 2012) Abstract Road roughness is a broad term that incorporates everything from potholes and cracks to the random deviations that exist in a profile. To build a roughness index, road irregularities need to be measured first. Existing methods of gauging the roughness are based either on visual inspections or using one of a limited number of instrumented vehicles that can take physical measurements of the road irregularities. This paper more specifically focuses on the estimation of a road profile (i.e., along the "wheel track"). This paper proposes a solution to the road profile estimation using a wavelet neural network (WNN) approach. The method incorporates a WNN which is trained using the data obtained from a 7-DOF vehicle dynamic model in the MATLAB Simulink software to approximate road profiles via the accelerations picked up from the vehicle. In this paper, a novel WNN, multi-input and multi-output feed forward wavelet neural network is constructed. In the hidden layer, wavelet basis functions are used as activate function instead of the sigmoid function of feed forward network. The training formulas based on BP algorithm are mathematically derived and a training algorithm is presented. The study investigates the estimation capability of wavelet neural networks through comparison between some estimated and real road profiles in the form of actual road roughness. Keywords: Road profile; Simulation; Wavelet neural network; BP algorithm; Estimation Introduction A road profile is one of the most effective vehicle environmental conditions that influences ride, handling, fatigue, fuel consumption, tire wear, maintenance costs, and vehicle delay costs. Therefore, establishment of methods for road profile measurement is completely essential. Currently, many routines are available for road profile measurement. Most of them measure vertical deviations of the road surface along the traveling wheel path. The American Society of Testing and Materials (ASTM) standard E867 [1] defines road roughness as the deviations of a pavement surface from a true planar surface with characteristic dimensions that affect vehicle dynamics, ride quality, dynamic loads, and drainage. About some of the road profile measuring methods and tools, their accuracy is affected by inaccurate vehicle manufacturer's data and insufficient degrees of freedom. Furthermore, both of these approaches demand formulating the inverse of a dynamic model. To avoid these problems, Ngwangwa et al. [2] developed an artificial neural network * Corresponding author. Tel.: , Fax.: address: Solhmirzaei.ali@gmail.com Recommended by Editor Yeon June Kang KSME & Springer 2012 (ANN) based technique to reconstruct the road profile. They used displacement responses of a quarter car model as inputs to a two-layer Narx network. They concluded that the technique is capable of reconstructing the road profile within a margin of error of 45%. They also indicated that with other considerations, the error may decrease to 20%. The applications of ANN based methods are rapidly increasing in various fields of science. They are able to approximate complicated systems. As for vehicle technology, neural network has contributed many solutions to areas such as control and dynamic simulations. The following is a brief summary of some of the neural network contributions to the vehicular field. In 1993, Kageyama [3] used a three-layer feed forward neural network to transform a group including 17 state variables of a vehicle model to four state variables of force. The outputs of the network were properly in agreement with the values resulting from the simulation. In 1994, Palkovics and his team [4] examined the ability of neural networks and also compared the feedforward and feedback neural network accuracy in simulation of a tire under vertical dynamic load. Due to lack of experimental data for training the network, they used results from simulation of a magic formula (MF)-tire model, which was proposed by Pace-

2 3030 A. Solhmirzaei et al. / Journal of Mechanical Science and Technology 26 (10) (2012) 3029~3036 jka and Takahashi in Their research showed feedforward methods are more accurate in estimation but not as robust as feedback methods [5]. In conclusion, they showed that the majority of complicated models can be replaced with the tire model created by the neural network [5]. In 1994, Wurtenberger and Iserman [5] used a feedforward neural network for obtaining a tire and vehicle model to study the lateral vehicle dynamics. In 1996, Ghazizadeh and Fahim [6] utilized a two-layer feedforward neural network to obtain a quasi-static roll model to study vehicle roll-over behavior. They used a vehicle model in which inputs were vehicle velocity and steering angle, and outputs were lateral acceleration, yaw rate, and quasi-static load transfer in the rear and front axles. They used network outputs with the number of time delays as feedback to the input. In 1997, Pasterkamp and Pacjeka [7] compared feedforward and radial basis networks to estimate the slip angle and the friction between the tire and road considering self-aligning torque and tire loads. They showed that although both networks perform reasonable estimation, feedforward networks are preferred because they have a smaller structure and are more robust. In addition, research pointed out that for determining the suspension system behavior, implementation of a suitably learned feedforward neural network is more appropriate than methods in which, to save on cost of measuring instruments, real-time complicated kinematic calculation has to be done. In 2009 Yousefzadeh and Azadi and Soltani [8] proposed a solution to the road profile estimation using an artificial neural network (ANN) approach. The method incorporates an ANN which is trained using the data obtained from a validated vehicle model in the ADAMS software to approximate road profiles via the accelerations picked up from the vehicle. The study investigates the estimation capability of neural networks through comparison between some estimated and real road profiles in the form of actual road roughness and power spectral density. They showed that all of the combinations indicate road profile PSDs have higher correlation in comparison to the real road profiles (in time domain). The models of natural phenomenon and physical system which include a nonlinear feature have been linearized via various linearizing techniques, because of their convenience of analysis. However, the nonlinear models have been driven by the improvement of the processor and the development of new mathematical theories. One of them is to utilize the neural networks as identification technique. The performance of identification technique depends on the type and learning algorithm of neural networks. The most popular neural networks are multi-layer perceptron network (MLPN). However, the MLPN has large structures. It induces the increase of calculation effort. Therefore, the wavelet neural network (WNN), which is a powerful tool as an estimator, was introduced by Zhang and Benveniste recently [10]. The WNN with a simple structure has excellent performance compared with the MLPN. But conventional back-propagation neural networks (BPNN) most frequently used in practical applications have low learning speed, difficulty in choosing the proper size of network, and easy to fall into local minima. WNN combines the time-frequency characteristic of wavelet transformation with the self-learning of conventional neural network. The basis of WNN is using a wavelet function as the activation function of neurons and combining wavelet with neural network directly [10]. As wavelet analysis employs mainly the expansion and contraction of basis function to detect simultaneously the characteristics of global and local of the measured signal [11], WNN inherits these characters from wavelet analysis, and has stronger approximating, tolerance and classification capacity than a conventional neural network, which makes it have strong advantages in dealing with nonlinear mapping and on-line estimates [12]. This paper selects the Mexican hat wavelet as the basis function, using the error BP algorithm to train the network. The data used in training and testing the WNN were obtained from the simulation of the MATLAB Simulink model, which was excited by road profiles generated via MATLAB. 2. Dynamic modeling of the vehicle 2.1 Full car vibrating model (a) (b) Fig. 1. (a) Full car vibrating model of a vehicle; (b) Full car vibrating model of a vehicle. A general vibrating model of a vehicle is called the full car model. Such a model that is shown in Fig. 1 includes the body bounce x, body roll φ, body pitch θ, wheels hop x 1, x 2, x 3, x 4 and independent road excitations y 1, y 2, y 3, y 4. A full car vibrating model has 7-DOF with the following equations of motion.

3 A. Solhmirzaei et al. / Journal of Mechanical Science and Technology 26 (10) (2012) 3029~ Table 1. Full vehicle model parameter.. (1) (2) (3) (4) (5) (6) Parameter Front suspension stiffness Rear suspension average stiffness Front tire vertical stiffness Rear tire vertical stiffness Anti roll bar stiffness Front vertical damping Rear vertical damping Body vehicle mass Front wheel mass Rear wheel mass Value k r = (N/m) k r = (N/m) k tf = (N/m) k tr = (N/m) k R = k Rf = k Rr = N m/rad c f = 2305 (N.s/m) c r = 1226 (N.s/m) m = 930 kg m f = m 1 = m 2 = 31.5 kg m r = m 3 = m 4 = 29 kg Pitch moment of inertia 1243 (kg.m 2 ) Roll moment of inertia 298 (kg.m 2 ) Wheel base Body vertical motion coordinate Front right wheel vertical motion coordinate Front left wheel vertical motion coordinate Rear right wheel vertical motion coordinate Rear left wheel vertical motion coordinate Body pitch motion coordinate Body roll motion coordinate Road excitation at the front right wheel Road excitation at the front left wheel Road excitation at the rear right wheel Road excitation at the rear left wheel 3.45 m x (m) x 1 (m) x 2 (m) x 3 (m) x 4 (m) θ (rad/s) φ (rad/s) y 1 (m) y 2 (m) y 3 (m) y 4 (m) Body longitudinal mass moment of inertia I x = 298 (kg.m 2 ) Body lateral mass moment of inertia I y = 1243 (kg.m 2 ) Distance of C from front axle Distance of C from rear axle Distance of C from right side Distance of C from left side a 1 = 1.7 (m) a 2 = 1.75 (m) b 1 = 0.7 (m) b 2 = (m) The body of the vehicle is assumed to be a rigid slab of mass m, which is the total body mass, a longitudinal mass moment of inertia I x and a lateral mass moment of inertia I y. The moments of inertia are only the body mass moments of inertia not the vehicle s mass moments of inertia. The wheels have a mass m 1, m 2, m 3, and m 4, respectively. The front and rear tires stiffness is indicated by k tf and k tr, respectively. Because the damping of tires is much smaller than the damping of shock absorbers, we may ignore the tires damping for simpler calculation. The suspension of the car has stiffness k f and damping c f in the front and stiffness k r and damping c r in the rear. It is common to make the suspension of the left and right wheels mirror. So, their stiffness and damping are equal. The (7) vehicle may also have an antiroll bar in front and in the back, with a torsional stiffness k Rf and k Rr. Using a simple model, the antiroll bar provides a torque m r proportional to the roll angle φ. Definitions of the employed parameters are indicated in Table Wavelet neural network and training algorithm The wavelet theory was proposed in multi-resolution analysis in the early 1980s for improving the defect of the Fourier series by Mallet. The WNN, which has a wavelet function, is one type of neural network [9]. Combining the wavelet transform theory with the basic concept of neural networks, a new mapping network called adaptive wavelet neural network (WNN) is proposed as an alternative to feedforward neural networks for approximating arbitrary nonlinear functions [13]. A wavelet neural network generally consists of a feedfor-

4 3032 A. Solhmirzaei et al. / Journal of Mechanical Science and Technology 26 (10) (2012) 3029~3036 Fig. 2. Mexican hat wavelet function [15]. ward neural network, with one hidden layer, whose activation functions are drawn from an orthonormal wavelet family. The WNN algorithms consist of two processes: the selfconstruction of networks and the minimization of error. In the first process, the network structures applied for representation are determined by using wavelet analysis [14]. The network gradually recruits hidden units to effectively and sufficiently cover the time-frequency region occupied by a given target. Simultaneously, the network parameters are updated to preserve the network topology and take advantage of the later process. Each hidden unit has a square window in the timefrequency plane. The optimization rule is only applied to the hidden units where the selected point falls into their windows. Therefore, the learning cost can be reduced. 3.1 Wavelet neural network Wavelet networks use three-layer structure (input layer, hidden layer, output layer) and wavelet activation function. The wavelet function is a waveform that has limited duration and average value of zero. The WNN architecture shown in Fig. 3 approximates any desired signal y(t) by generalizing a linear combination of a set of mother wavelets φ a,b (t). The mother wavelet is composed of the translation factor and the dilation factor a i, where the subscript i indicates the ith wavelet and n indicates the nth input signal [13]: ψ 1 un bi ( un) =. a ϕ a i i Note that the dilation factor a > 0. In this work, the Mexican hat wavelet is used for the wavelet neural network. Compared with other wavelet functions, the Mexican hat wavelet function has several characteristics that are advantageous in this work: (1) it has an analytical expression and therefore can be used conveniently for decomposing multidimensional time series, (2) it can be differentiated analytically, (3) it is a non-compactly supported but rapidly vanishing function (Jiang and Adeli 2004b), and (4) it is computationally efficient. The Mexican hat wavelet function is expressed as follows (Fig. 2) [15]: 2 2 t ϕ( t) = (1 t ) ex p 2 (8) (9) Fig. 3. Structure of a wavelet neural network. where X b t=. (10) a The structure of a wavelet neural network is very similar to that of a (1+1/2) layer neural network. That is, a feedforward neural network, taking one or more inputs, with one hidden layer and whose output layer consists of one or more linear combiners or summers (see Fig. 3). The hidden layer consists of neurons, whose activation functions are drawn from a wavelet basis. These wavelet neurons are usually referred to as wavelons. There are two main approaches to creating wavelet neural networks [13]. In the first, the wavelet and the neural network processing are performed separately. The input signal is first decomposed using some wavelet basis by the neurons in the hidden layer. The wavelet coefficients are then output to one or more summers whose input weights are modified in accordance with some learning algorithm. The second type combines the two theories. In this case the translation and dilation of the wavelets along with the summer weights are modified in accordance with some learning algorithm. In general, when the first approach is used, only dyadic dilations and translations of the mother wavelet form the wavelet basis. This type of wavelet neural network is usually referred to as a wavenet. We will refer to the second type as a wavelet network. The input in this case is a multidimensional vector and the wavelons consist of multidimensional wavelet activation functions. They will produce a non-zero output when the input vector lies within a small area of the multidimensional input space. The output of the wavelet neural network is one or more linear combinations of these multidimensional wavelets. Fig. 4 shows the form of a wavelon. The output is defined as:. (11) This wavelon is in effect equivalent to a multidimensional wavelet. The architecture of a multidimensional wavelet neural network is shown in Fig. 3. The hidden layer consists of M wavelons. The output layer consists of K summers. The output

5 A. Solhmirzaei et al. / Journal of Mechanical Science and Technology 26 (10) (2012) 3029~ (14) where N y indicates the number of the past outputs and N u describes the past inputs. And also, d k (n) is the nonlinear system output and u 1 (n) is the identification input. In this research N y and N u are Wavelet neural network training algorithm Fig. 4. A wavelet neuron with a multidimensional wavelet activation function. A nonlinear optimization algorithm, such as gradient descent, conjugate gradients or Byden-Fletcher-Goldfarb Shanno (BFGS), could be applied to training a wavelet neural network. However, the advantage of the wavelet neural network architecture is that it can be trained in stages using linear optimization algorithms, which allows for faster training and improved convergence compared with nonlinear alternatives. One method often used to vary the weights and biases is known as the backpropagation algorithm, in which the weights and biases are modified so as to minimize an average quadratic error function of the form: 1 E= d n y n 2 N N 2 k ( ) k ( ) (15) n= 1 k= 1 Fig. 5. Identification structure using the WNN. of the network is defined as where d k (n) is the expected output of WNN. The backpropagation algorithm actually adopts gradient descent to minimize E and the corresponding iterative formulas are presented as the following [17]:. (12) (16) Therefore, the input-output mapping of the network is defined as: (13) where M is the number of windowing wavelets, W i is the weight coefficients, K is a number of outputs, and N is a number of input. 3.2 Identification method for nonlinear systems In this paper, we employ the serial-parallel method for identifying model of the nonlinear system. Fig. 5 represents the identification structure. The inputs of the WNN for the identifying model consist of the current input, the past inputs, and the past outputs of the nonlinear system. The current output of the WNN represents as follows: (17) (18) (19) (20) (21) where η refers to w ik, a i and b i learning rate parameter; µ refers to momentum their own factor 0 < µ < Profile estimation using wavelet neural network In this section, we apply the proposed algorithm to vehicle dynamic systems. Choosing an appropriate architecture of WNN is dependent

6 3034 A. Solhmirzaei et al. / Journal of Mechanical Science and Technology 26 (10) (2012) 3029~3036 Table 2. Simulation parameters and the results for WNN. Simulation condition Model Number of wavelet node 12 Number of past inputs 2 Number of past output of plant 2 Sampling rate 0.01 Table 3. Road profiles used for training and testing the network. Road profile No. Road type Application 1 C Training 2 D Training 3 C Test 4 D Test on the type of system being modeled. In this work, the MATLAB software was used to model the intended WNN. The inverse of a vehicle model was used to construct the WNN model, where the inputs were accelerations of a vehicle moving along a road and the outputs were the road profiles. The network is a dynamic WNN. In addition to the vehicle accelerations, the delayed version of the vehicle accelerations and the road profiles were also input to the WNN block. The number of hidden layers and their nodes is an arbitrary parameter that cannot be determined according to a specified rule. In other words, an optimized network size can be achieved using trial and error and by considering the accuracy of the results and the training convergence speed. In this work, a network including one hidden layer with fifteen Mexican hat wavelet function nodes and an output layer with linear transfer function nodes was created. The outputs of the network were four road profiles related to each of the vertical displacements of the wheels during the vehicle trip. The network input consisted of two groups. The first group, called independent input, consisted of seven vehicle accelerations. The second group, dependent input, was the feedback of the network output and independent input, both with one and two delay elements. A dependent input was used because state variables in a dynamic system depend not only on the current inputs of the system but also on the state variables in previous time. In this model, several different delays and their combinations were analyzed, and, finally, the combination of one and two delays was found to be suitable. Neural networks adjust the values of weights and biases through a process that is referred to as a learning rule or training algorithm. 5. Training data collection Network training data were gathered using the vehicle dynamic systems model in the MATLAB Simulink software. Two road profiles, generated in MATLAB, excited the front wheels. The other two road profiles for exciting the rear wheels were mostly similar to those of the front wheels but with some delays caused by the distance between the front and rear axles of the vehicle. These four road profiles with a specified vehicle velocity were applied to the vehicle model to derive seven accelerations, including three accelerations of the body (roll, bounce and pitch) and four accelerations of the wheels. The generated road profiles were considered as network output training data. Input training data consisted of independent and dependent Table 4. Combination sets used for evaluating the networks. Combination No. Training set Vehicle velocity (m/s) data, which were introduced in the previous section. 6. Training and testing the network Testing set 1 Road profile No.1 30 Road profile No. 3 2 Road profile No.1 30 Road profile No. 4 3 Road profile No.2 30 Road profile No. 3 4 Road profile No.2 30 Road profile No. 4 For training and testing the network, road profiles similar to the types C and D of ISO 8608 standard [16] were used. As shown in Table 3, four groups of road profiles were created. Each of the road profiles consisted of four series of data for exciting the wheels of the vehicle in the MATLAB Simulink software. The road profiles were originally 1000 m long, but with considering 3.45 m distance between the rear and front axles, m of road profiles were used. Furthermore, the frequency content of the road profiles was around cycles/m, as was intended from the beginning. After applying the roads to the vehicle at different speed and deriving the accelerations, the transient span of the accelerations should be removed to have proper network training. As shown in Table 4, four combinations of road data from Table 3 with their corresponding accelerations were used to train and test four different networks. The transient parts of the data were removed before training and testing the network. The criterion is a minimum value for the normalized root of the sum of the square of errors (RSSE) for all training data points as follows [15]: (22) where y p and y m represent the measured and predicted outputs, and y P is the mean of the measured outputs and Na is the total number of training samples. RSSE is also an indicator of the performance of the WNN model. In this research, the value of RSSE < ε = is specified. Training usually starts

7 A. Solhmirzaei et al. / Journal of Mechanical Science and Technology 26 (10) (2012) 3029~ Fig. 6. Combination 1 front left wheel. Fig. 9. Combination 2 front right wheel. Fig. 7. Combination 1 front right wheel. Fig. 10. Combination 3 front right wheel. Fig. 8. Combination 2 front left wheel. Fig. 11. Combination 3 front right wheel. from a random set of weights and proceeds until a specified value of the RSSE is met or a maximum number of iterations are reached. In most cases, the cost function displays many local minima, and the training result depends on the initial weight values. Figs indicate the networks in predicting the target profile. 7. Conclusion We have proposed a BP based training algorithm for the WNN. To verify the effectiveness of the proposed algorithm, we applied it to train the parameters of the WNN. And then using the WNN, we executed the identification for the vehicle system dynamic. In addition, the WNN training by the pro- Fig. 12. Combination 4 front right wheel.

8 3036 A. Solhmirzaei et al. / Journal of Mechanical Science and Technology 26 (10) (2012) 3029~3036 Fig. 13. Combination 4 front right wheel. posed algorithm adapts well to the abrupt change and the high nonlinearity of the chaotic systems, because the proposed theorem concerns the learning rates of each parameters of the WNN, respectively. In this paper, the idea of road profile estimation using neural network algorithm is presented. Due to lack of equipment, such as a four-post laboratory and a vehicle provided with accelerometers and accurate distance measuring system, a full ride model in the MATLAB Simulink software was used for simulations. In the next stage of this research, using road data including frequency contents with much longer wavelengths is planned to identify the effect of profiles frequency contents and vehicle speed on estimation accuracy. References [1] American Society of Testing and Materials, Standard Terminology Relating to Vehicle Pavement Sections, ASTM E867, Annual Book of ASTM Standards, 4.03 (2000). [2] H. M. Ngwangwa, P. Stephan Heyns, K. F. J. Labuschange and G. K. Kululanga, An overview of the neural network based technique for monitoring of road condition via reconstructed road profiles, Proc. of the 27th Southern African Transport Conference, ISBN No.: (2008). [3] I. Kageyama, On a control of tire force coefficient for vehicle handling with neural network system estimation for load rate of tire forces, 13th IAVSD Symposium on Dynamics of Vehicles on Roads and Tracks, Chengdu, Sichuan, China (1992) [4] H. B. Pacejka and T. Takahashi, Pure slip characteristics of tires on flat and undulated road surfaces, Proc. Int. Symp. Advanced Vehicle Control, SAE, Japan, No (1992) [5] L. Palkoviks, M. El-Gindy and H. B. Pacejka, Modeling of the cornering characteristics of tires on an uneven road surface, Dynamic Version of the "Neuro-Tire, Int. Journal of Vehicle Design, 15 (1/2) (1994) [6] A. Ghazizadeh and A. Fahim, Neural network representation of a vehicle model: "Neuro-Vehicle, Int. J. of Vehicle Design, 17 (1) (1996) [7] W. R. Pasterkamp and H. B. Pacejka, Application of neural networks in the estimation of tire/road friction using the tire as sensor, SAE, No (1997). [8] M. Yousefzadeh, Sh. Azadi and A. Soltani, Road profile estimation using neural network algorithm, Journal of Mechanical Science and Technology, 24 (3) (2010) [9] K. J. Kim, J. B. Park and Y. H. Choi, The adaptive learning rates of extended kalman filter based training algorithm for wavelet neural networks, Springer-Verlag Berlin Heidelberg, LNAI 4293 ( 2006) [10] Q. Zhang and A.T. Benveniste, Wavelet networks, IEEE Trans. On Neural Networks, 3 (6) (1992) [11] Q. Xu, X. Meng, X. Han and S. Meng, Gas turbine fault diagnosis based on wavelet neural network, In: Proceedings of the 2007 International Conference on Wavelet Analysis and Pattern Recognition, Beijing, China, 2-4, November (2007) [12] Z. Rui, L. Xu and R. Feng, Gear faults diagnosis based on wavelet neural network, Journal of Mechanical Transmission 01 (2008) [13] D. Veitch, Wavelet neural networks and their application in the study of dynamical system, Department of Mathematics university of York UK, August [14] G. Lekutai, Adaptive self-tuning neuro wavelet network controllers, Doctor of philosophy in The Electrical Engineering Department Virginia Polytechnic Institute and State University, March 31, [15] X. Jiang, Dynamic fazzy neural network for system identification, damage detection and active control of Highrise building, Degree Doctor of Philosophy in Ohio State University, [16] Mechanical Vibration-Road Surface Profiles-Reporting of Measured Data, International Organization for Standardization, ISO 8608 (1995). [17] Q. Huang, D. Jiang, L. Hong and Y. Ding, Application of wavelet neural networks on vibration fault diagnosis for wind turbine gearbox, Springer-Verlag Berlin Heidelberg, LNCS 5264 (2008) Ali Solhmirzaei received his BSc in Railway Engineering (Rolling Stock) from Iran University of Science and Technology in 2008, and his MSc in Mechanical Engineering from K.N.T University of Technology, Iran, in His research is mainly focused on vehicle dynamics, railway vehicle dynamics, finite elements and fatigue analysis of railway structures. Shahram Azadi received his B.S. and M.S. in Mechanical Engineering from Sharif University of Technology, Iran, in 1988 and 1992, respectively. He then received his Ph.D from Amirkabir University of Technology, Iran, in Dr. Azadi is currently an assistant professor in the faculty of Mechanical Engineering at K.N.Toosi University of Technology in Tehran, Iran.

EXPERIMENTALLY OBTAINED BIFURCATION PHENOMENON IN CHAOTIC TRACTOR VIBRATING IN TIME AND FREQUENCY DOMAIN

EXPERIMENTALLY OBTAINED BIFURCATION PHENOMENON IN CHAOTIC TRACTOR VIBRATING IN TIME AND FREQUENCY DOMAIN International Journal of Bifurcation and Chaos, Vol. 15, No. 1 (2005) 225 231 c World Scientific Publishing Company EXPERIMENTALLY OBTAINED BIFURCATION PHENOMENON IN CHAOTIC TRACTOR VIBRATING IN TIME AND

More information

NONLINEAR PLANT IDENTIFICATION BY WAVELETS

NONLINEAR PLANT IDENTIFICATION BY WAVELETS NONLINEAR PLANT IDENTIFICATION BY WAVELETS Edison Righeto UNESP Ilha Solteira, Department of Mathematics, Av. Brasil 56, 5385000, Ilha Solteira, SP, Brazil righeto@fqm.feis.unesp.br Luiz Henrique M. Grassi

More information

Road Vehicle Dynamics

Road Vehicle Dynamics Road Vehicle Dynamics Table of Contents: Foreword Preface Chapter 1 Introduction 1.1 General 1.2 Vehicle System Classification 1.3 Dynamic System 1.4 Classification of Dynamic System Models 1.5 Constraints,

More information

Frequency Response of 10 Degrees of Freedom Full-Car Model For Ride Comfort

Frequency Response of 10 Degrees of Freedom Full-Car Model For Ride Comfort International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 882 Volume 4, Issue 1, January 215 43 Frequency Response of 1 Degrees of Freedom Full-Car Model For Ride Comfort

More information

Estimation of Tire-Road Friction by Tire Rotational Vibration Model

Estimation of Tire-Road Friction by Tire Rotational Vibration Model 53 Research Report Estimation of Tire-Road Friction by Tire Rotational Vibration Model Takaji Umeno Abstract Tire-road friction is the most important piece of information used by active safety systems.

More information

Hybrid neural network bushing model for vehicle dynamics simulation

Hybrid neural network bushing model for vehicle dynamics simulation Journal of Mechanical Science and Technology 22 (2008) 2365~2374 Journal of Mechanical Science and Technology www.springerlink.com/content/1738-494x DOI 10.1007/s12206-008-0712-2 Hybrid neural network

More information

STUDY OF EFFECTS OF VIBRATIONS CAUSED BY RAILWAY TRAFFIC TO BUILDINGS

STUDY OF EFFECTS OF VIBRATIONS CAUSED BY RAILWAY TRAFFIC TO BUILDINGS Bulletin of the Transilvania University of Braşov CIBv 2014 Vol. 7 (56) Special Issue No. 1-2014 STUDY OF EFFECTS OF VIBRATIONS CAUSED BY RAILWAY TRAFFIC TO BUILDINGS R. NERIŞANU 1 D. DRĂGAN 1 M. SUCIU

More information

String tyre model for evaluating steering agility performance using tyre cornering force and lateral static characteristics

String tyre model for evaluating steering agility performance using tyre cornering force and lateral static characteristics Vehicle System Dynamics International Journal of Vehicle Mechanics and Mobility ISSN: 0042-3114 (Print) 1744-5159 (Online) Journal homepage: http://www.tandfonline.com/loi/nvsd20 String tyre model for

More information

Simple Car Dynamics. Outline. Claude Lacoursière HPC2N/VRlab, Umeå Universitet, Sweden, May 18, 2005

Simple Car Dynamics. Outline. Claude Lacoursière HPC2N/VRlab, Umeå Universitet, Sweden, May 18, 2005 Simple Car Dynamics Claude Lacoursière HPC2N/VRlab, Umeå Universitet, Sweden, and CMLabs Simulations, Montréal, Canada May 18, 2005 Typeset by FoilTEX May 16th 2005 Outline basics of vehicle dynamics different

More information

International Journal of Multidisciplinary and Current Research

International Journal of Multidisciplinary and Current Research International Journal of Multidisciplinary and Current Research Research Article ISSN: 2321-3124 Available at: http://ijmcr.com Theoretical and Numerical Analysis of Half Car Vehicle Dynamic Model Subjected

More information

Modeling of Pantograph-Catenary dynamic stability

Modeling of Pantograph-Catenary dynamic stability Technical Journal of Engineering and Applied Sciences Available online at www.tjeas.com 2013 TJEAS Journal-2013-3-14/1486-1491 ISSN 2051-0853 2013 TJEAS Modeling of Pantograph-Catenary dynamic stability

More information

MECH 3140 Final Project

MECH 3140 Final Project MECH 3140 Final Project Final presentation will be held December 7-8. The presentation will be the only deliverable for the final project and should be approximately 20-25 minutes with an additional 10

More information

2621. Road profile estimation for suspension system based on the minimum model error criterion combined with a Kalman filter

2621. Road profile estimation for suspension system based on the minimum model error criterion combined with a Kalman filter 2621. Road profile estimation for suspension system based on the minimum model error criterion combined with a Kalman filter Zhen Feng Wang 1, Ming Ming Dong 2, Ye Chen Qin 3, Liang Gu 4 Noise and Vibration

More information

PSD Analysis and Optimization of 2500hp Shale Gas Fracturing Truck Chassis Frame

PSD Analysis and Optimization of 2500hp Shale Gas Fracturing Truck Chassis Frame Send Orders for Reprints to reprints@benthamscience.ae The Open Mechanical Engineering Journal, 2014, 8, 533-538 533 Open Access PSD Analysis and Optimization of 2500hp Shale Gas Fracturing Truck Chassis

More information

Online Identification And Control of A PV-Supplied DC Motor Using Universal Learning Networks

Online Identification And Control of A PV-Supplied DC Motor Using Universal Learning Networks Online Identification And Control of A PV-Supplied DC Motor Using Universal Learning Networks Ahmed Hussein * Kotaro Hirasawa ** Jinglu Hu ** * Graduate School of Information Science & Electrical Eng.,

More information

UNIVERSITY OF BOLTON SCHOOL OF ENGINEERING

UNIVERSITY OF BOLTON SCHOOL OF ENGINEERING TW32 UNIVERSITY OF BOLTON SCHOOL OF ENGINEERING BENG (HONS) AUTOMOTIVE PERFORMANCE ENGINEERING and BSC (HONS) MOTORSPORT TECHNOLOGY EXAMINATION SEMESTER 2-2015/2016 VEHICLE DYNAMICS AND ADVANCED ELECTRONICS

More information

POWER SYSTEM DYNAMIC SECURITY ASSESSMENT CLASSICAL TO MODERN APPROACH

POWER SYSTEM DYNAMIC SECURITY ASSESSMENT CLASSICAL TO MODERN APPROACH Abstract POWER SYSTEM DYNAMIC SECURITY ASSESSMENT CLASSICAL TO MODERN APPROACH A.H.M.A.Rahim S.K.Chakravarthy Department of Electrical Engineering K.F. University of Petroleum and Minerals Dhahran. Dynamic

More information

5.5 Exercises for This Chapter Two-Axle Vehicle on Cosine Track Two-Axle Vehicle on Generally Periodic Track...

5.5 Exercises for This Chapter Two-Axle Vehicle on Cosine Track Two-Axle Vehicle on Generally Periodic Track... Contents 1 Introduction... 1 1.1 The Basic Function of the Wheel/rail System.... 1 1.2 Significance of Dynamics on the Operation of Rail Vehicles... 2 1.3 On the History of Research in the Field of Railway

More information

Stress Analysis and Validation of Superstructure of 15-meter Long Bus under Normal Operation

Stress Analysis and Validation of Superstructure of 15-meter Long Bus under Normal Operation AIJSTPME (2013) 6(3): 69-74 Stress Analysis and Validation of Superstructure of 15-meter Long Bus under Normal Operation Lapapong S., Pitaksapsin N., Sucharitpwatkul S.*, Tantanawat T., Naewngerndee R.

More information

Simulation of the Stick-Slip Friction between Steering Shafts Using ADAMS/PRE

Simulation of the Stick-Slip Friction between Steering Shafts Using ADAMS/PRE Simulation of the Stick-Slip Friction between Steering Shafts Using ADAMS/PRE Dexin Wang and Yuting Rui Research & Vehicle Technology Ford Motor Company ABSTRACT Cyclic stick-slip friction is a well-known

More information

ARTIFICIAL NEURAL NETWORKS گروه مطالعاتي 17 بهار 92

ARTIFICIAL NEURAL NETWORKS گروه مطالعاتي 17 بهار 92 ARTIFICIAL NEURAL NETWORKS گروه مطالعاتي 17 بهار 92 BIOLOGICAL INSPIRATIONS Some numbers The human brain contains about 10 billion nerve cells (neurons) Each neuron is connected to the others through 10000

More information

University of Bristol - Explore Bristol Research. Publisher's PDF, also known as Version of record

University of Bristol - Explore Bristol Research. Publisher's PDF, also known as Version of record Watanabe, N., & Stoten, D. P. (214). Actuator control for a rapid prototyping railway bogie, using a dynamically substructured systems approach. In Proceedings of 12th International Conference on Motion

More information

An artificial neural networks (ANNs) model is a functional abstraction of the

An artificial neural networks (ANNs) model is a functional abstraction of the CHAPER 3 3. Introduction An artificial neural networs (ANNs) model is a functional abstraction of the biological neural structures of the central nervous system. hey are composed of many simple and highly

More information

Investigation on dynamic behavior of railway track in transition zone

Investigation on dynamic behavior of railway track in transition zone Journal of Mechanical Science and Technology 25 (2) (2) 287~292 wwwspringerlinkcom/content/738494x DOI 7/s22622x Investigation on dynamic behavior of railway track in transition zone JabbarAli Zakeri *

More information

Research on Optimization of Bearing Span of Main Reducer Gear System Based on the Transfer Matrix

Research on Optimization of Bearing Span of Main Reducer Gear System Based on the Transfer Matrix MATEC Web of Conferences 3, () DOI:.5/ matecconf/3 MMME Research on Optimization of Bearing Span of Main Reducer Gear System Based on the Transfer Matrix Feng Yun HUANG, Li TIAN a, Ding WEI and Huan Huan

More information

Estimation of the Pre-Consolidation Pressure in Soils Using ANN method

Estimation of the Pre-Consolidation Pressure in Soils Using ANN method Current World Environment Vol. 11(Special Issue 1), 83-88 (2016) Estimation of the Pre-Consolidation Pressure in Soils Using ANN method M. R. Motahari Department of Civil Engineering, Faculty of Engineering,

More information

Due Date 1 (for confirmation of final grade): Monday May 10 at 11:59pm Due Date 2 (absolute latest possible submission): Friday May 14 at 5pm

Due Date 1 (for  confirmation of final grade): Monday May 10 at 11:59pm Due Date 2 (absolute latest possible submission): Friday May 14 at 5pm ! ME345 Modeling and Simulation, Spring 2010 Case Study 3 Assigned: Friday April 16! Due Date 1 (for email confirmation of final grade): Monday May 10 at 11:59pm Due Date 2 (absolute latest possible submission):

More information

Application of Neuro Fuzzy Reduced Order Observer in Magnetic Bearing Systems

Application of Neuro Fuzzy Reduced Order Observer in Magnetic Bearing Systems Application of Neuro Fuzzy Reduced Order Observer in Magnetic Bearing Systems M. A., Eltantawie, Member, IAENG Abstract Adaptive Neuro-Fuzzy Inference System (ANFIS) is used to design fuzzy reduced order

More information

Analysis and Design of an Electric Vehicle using Matlab and Simulink

Analysis and Design of an Electric Vehicle using Matlab and Simulink Analysis and Design of an Electric Vehicle using Matlab and Simulink Advanced Support Group January 22, 29 23-27: University of Michigan Research: Optimal System Partitioning and Coordination Original

More information

IMPACT OF ROAD SURFACE ROUGHNESS AND ENGINE TORQUE ON THE LOAD OF AUTOMOTIVE TRANSMISSION SYSTEM

IMPACT OF ROAD SURFACE ROUGHNESS AND ENGINE TORQUE ON THE LOAD OF AUTOMOTIVE TRANSMISSION SYSTEM International Journal of echanical Engineering and Technology (IJET) Volume 10, Issue 02, February 2019, pp. 1752 1761, Article ID: IJET_10_02_181 Available online at http://www.iaeme.com/ijmet/issues.asp?jtype=ijet&vtype=10&itype=2

More information

A Nonlinear Dynamic Model for Single-axle Wheelsets with Profiled Wheels and Rails

A Nonlinear Dynamic Model for Single-axle Wheelsets with Profiled Wheels and Rails Proceedings of the International Conference of Control, Dynamic Systems, and Robotics Ottawa, Ontario, Canada, May 15-16 2014 Paper No. 60 A Nonlinear Dynamic Model for Single-axle Wheelsets with Profiled

More information

Artificial Neural Networks. Edward Gatt

Artificial Neural Networks. Edward Gatt Artificial Neural Networks Edward Gatt What are Neural Networks? Models of the brain and nervous system Highly parallel Process information much more like the brain than a serial computer Learning Very

More information

Single-track models of an A-double heavy vehicle combination

Single-track models of an A-double heavy vehicle combination Single-track models of an A-double heavy vehicle combination PETER NILSSON KRISTOFFER TAGESSON Department of Applied Mechanics Division of Vehicle Engineering and Autonomous Systems Vehicle Dynamics Group

More information

ARTIFICIAL NEURAL NETWORK WITH HYBRID TAGUCHI-GENETIC ALGORITHM FOR NONLINEAR MIMO MODEL OF MACHINING PROCESSES

ARTIFICIAL NEURAL NETWORK WITH HYBRID TAGUCHI-GENETIC ALGORITHM FOR NONLINEAR MIMO MODEL OF MACHINING PROCESSES International Journal of Innovative Computing, Information and Control ICIC International c 2013 ISSN 1349-4198 Volume 9, Number 4, April 2013 pp. 1455 1475 ARTIFICIAL NEURAL NETWORK WITH HYBRID TAGUCHI-GENETIC

More information

MODEL-BASED ANALYSIS OF WHEEL SPEED VIBRATIONS FOR ROAD FRICTION CLASSIFICATION USING MF-SWIFT. Antoine Schmeitz, Mohsen Alirezaei

MODEL-BASED ANALYSIS OF WHEEL SPEED VIBRATIONS FOR ROAD FRICTION CLASSIFICATION USING MF-SWIFT. Antoine Schmeitz, Mohsen Alirezaei MODEL-BASED ANALYSIS OF WHEEL SPEED VIBRATIONS FOR ROAD FRICTION CLASSIFICATION USING MF-SWIFT Antoine Schmeitz, Mohsen Alirezaei CONTENTS Introduction Road friction classification from wheel speed vibrations

More information

Chapter 3. Experimentation and Data Acquisition

Chapter 3. Experimentation and Data Acquisition 48 Chapter 3 Experimentation and Data Acquisition In order to achieve the objectives set by the present investigation as mentioned in the Section 2.5, an experimental set-up has been fabricated by mounting

More information

EXAMPLE: MODELING THE PT326 PROCESS TRAINER

EXAMPLE: MODELING THE PT326 PROCESS TRAINER CHAPTER 1 By Radu Muresan University of Guelph Page 1 EXAMPLE: MODELING THE PT326 PROCESS TRAINER The PT326 apparatus models common industrial situations in which temperature control is required in the

More information

A Novel 2-D Model Approach for the Prediction of Hourly Solar Radiation

A Novel 2-D Model Approach for the Prediction of Hourly Solar Radiation A Novel 2-D Model Approach for the Prediction of Hourly Solar Radiation F Onur Hocao glu, Ö Nezih Gerek, and Mehmet Kurban Anadolu University, Dept of Electrical and Electronics Eng, Eskisehir, Turkey

More information

Introduction to Biomedical Engineering

Introduction to Biomedical Engineering Introduction to Biomedical Engineering Biosignal processing Kung-Bin Sung 6/11/2007 1 Outline Chapter 10: Biosignal processing Characteristics of biosignals Frequency domain representation and analysis

More information

Identification of two-mass system parameters using neural networks

Identification of two-mass system parameters using neural networks 3ème conférence Internationale des énergies renouvelables CIER-2015 Proceedings of Engineering and Technology - PET Identification of two-mass system parameters using neural networks GHOZZI Dorsaf 1,NOURI

More information

Dynamic (Vibrational) and Static Structural Analysis of Ladder Frame

Dynamic (Vibrational) and Static Structural Analysis of Ladder Frame Dynamic (Vibrational) and Static Structural Analysis of Ladder Frame Ketan Gajanan Nalawade 1, Ashish Sabu 2, Baskar P 3 School of Mechanical and building science, VIT University, Vellore-632014, Tamil

More information

SOP Release. FEV Chassis Reliable Partner in Chassis Development. FEV Chassis Applications and Activities. Concept Layout. Design

SOP Release. FEV Chassis Reliable Partner in Chassis Development. FEV Chassis Applications and Activities. Concept Layout. Design CHASSIS Reliable Partner in Chassis Development FEV Chassis Applications and Activities Founded in 1978, FEV is an internationally recognized leader in the design and development of internal combustion

More information

Car Dynamics using Quarter Model and Passive Suspension, Part VI: Sprung-mass Step Response

Car Dynamics using Quarter Model and Passive Suspension, Part VI: Sprung-mass Step Response IOSR Journal of Computer Engineering (IOSR-JCE) e-issn: 2278-0661,p-ISSN: 2278-8727, Volume 17, Issue 2, Ver. 1 (Mar Apr. 2015), PP 65-74 www.iosrjournals.org Car Dynamics using Quarter Model and Passive

More information

Estimation of Inelastic Response Spectra Using Artificial Neural Networks

Estimation of Inelastic Response Spectra Using Artificial Neural Networks Estimation of Inelastic Response Spectra Using Artificial Neural Networks J. Bojórquez & S.E. Ruiz Universidad Nacional Autónoma de México, México E. Bojórquez Universidad Autónoma de Sinaloa, México SUMMARY:

More information

Neural Networks and the Back-propagation Algorithm

Neural Networks and the Back-propagation Algorithm Neural Networks and the Back-propagation Algorithm Francisco S. Melo In these notes, we provide a brief overview of the main concepts concerning neural networks and the back-propagation algorithm. We closely

More information

Lecture 4: Perceptrons and Multilayer Perceptrons

Lecture 4: Perceptrons and Multilayer Perceptrons Lecture 4: Perceptrons and Multilayer Perceptrons Cognitive Systems II - Machine Learning SS 2005 Part I: Basic Approaches of Concept Learning Perceptrons, Artificial Neuronal Networks Lecture 4: Perceptrons

More information

Armin Rasch * Abstract

Armin Rasch * Abstract Optimization strategy for the identification of elastomer parameters of truck mountings for the improved adjustment of Multi-Body Simulation data with measured values on rough road conditions Armin Rasch

More information

1330. Comparative study of model updating methods using frequency response function data

1330. Comparative study of model updating methods using frequency response function data 1330. Comparative study of model updating methods using frequency response function data Dong Jiang 1, Peng Zhang 2, Qingguo Fei 3, Shaoqing Wu 4 Jiangsu Key Laboratory of Engineering Mechanics, Nanjing,

More information

Experimental Study about the Applicability of Traffic-induced Vibration for Bridge Monitoring

Experimental Study about the Applicability of Traffic-induced Vibration for Bridge Monitoring Experimental Study about the Applicability of Traffic-induced Vibration for Bridge Monitoring Kyosuke Yamamoto, Riku Miyamoto, Yuta Takahashi and Yukihiko Okada Abstract Traffic-induced vibration is bridge

More information

4. Multilayer Perceptrons

4. Multilayer Perceptrons 4. Multilayer Perceptrons This is a supervised error-correction learning algorithm. 1 4.1 Introduction A multilayer feedforward network consists of an input layer, one or more hidden layers, and an output

More information

Artificial Neural Network Method of Rock Mass Blastability Classification

Artificial Neural Network Method of Rock Mass Blastability Classification Artificial Neural Network Method of Rock Mass Blastability Classification Jiang Han, Xu Weiya, Xie Shouyi Research Institute of Geotechnical Engineering, Hohai University, Nanjing, Jiangshu, P.R.China

More information

THE IMPACT OF SELECTED STRUCTURAL PARAMETERS IN UNSPRUNG INDUSTRIAL VEHICLES ON THEIR LONGITUDINAL OSCILLATIONS

THE IMPACT OF SELECTED STRUCTURAL PARAMETERS IN UNSPRUNG INDUSTRIAL VEHICLES ON THEIR LONGITUDINAL OSCILLATIONS Journal of KONES Powertrain and Transport, Vol. 1, No. 16 2009 THE IMPACT OF SELECTED STRUCTURAL PARAMETERS IN UNSPRUNG INDUSTRIAL VEHICLES ON THEIR LONGITUDINAL OSCILLATIONS Andrzej Kosiara Technical

More information

USING WAVELET NEURAL NETWORK FOR THE IDENTIFICATION OF A BUILDING STRUCTURE FROM EXPERIMENTAL DATA

USING WAVELET NEURAL NETWORK FOR THE IDENTIFICATION OF A BUILDING STRUCTURE FROM EXPERIMENTAL DATA 13 th World Conference on Earthquake Engineering Vancouver, B.C., Canada August 1-6, 24 Paper No. 241 USING WAVELET NEURAL NETWORK FOR THE IDENTIFICATION OF A BUILDING STRUCTURE FROM EXPERIMENTAL DATA

More information

The single track model

The single track model The single track model Dr. M. Gerdts Uniersität Bayreuth, SS 2003 Contents 1 Single track model 1 1.1 Geometry.................................... 1 1.2 Computation of slip angles...........................

More information

Short Term Load Forecasting Based Artificial Neural Network

Short Term Load Forecasting Based Artificial Neural Network Short Term Load Forecasting Based Artificial Neural Network Dr. Adel M. Dakhil Department of Electrical Engineering Misan University Iraq- Misan Dr.adelmanaa@gmail.com Abstract Present study develops short

More information

Effect of Dynamic Interaction between Train Vehicle and Structure on Seismic Response of Structure

Effect of Dynamic Interaction between Train Vehicle and Structure on Seismic Response of Structure Effect of Dynamic Interaction between Train Vehicle and Structure on Seismic Response of Structure Munemasa TOKUNAGA & Masamichi SOGABE Railway Technical Research Institute, Japan SUMMARY: The conventional

More information

Journal of Engineering Science and Technology Review 6 (2) (2013) Research Article. Received 25 June 2012; Accepted 15 January 2013

Journal of Engineering Science and Technology Review 6 (2) (2013) Research Article. Received 25 June 2012; Accepted 15 January 2013 Jestr Journal of Engineering Science and Technology Review 6 () (3) 5-54 Research Article JOURNAL OF Engineering Science and Technology Review www.jestr.org Fault Diagnosis and Classification in Urban

More information

ECE Introduction to Artificial Neural Network and Fuzzy Systems

ECE Introduction to Artificial Neural Network and Fuzzy Systems ECE 39 - Introduction to Artificial Neural Network and Fuzzy Systems Wavelet Neural Network control of two Continuous Stirred Tank Reactors in Series using MATLAB Tariq Ahamed Abstract. With the rapid

More information

CHAPTER INTRODUCTION

CHAPTER INTRODUCTION CHAPTER 3 DYNAMIC RESPONSE OF 2 DOF QUARTER CAR PASSIVE SUSPENSION SYSTEM (QC-PSS) AND 2 DOF QUARTER CAR ELECTROHYDRAULIC ACTIVE SUSPENSION SYSTEM (QC-EH-ASS) 3.1 INTRODUCTION In this chapter, the dynamic

More information

Computational Simulation of Dynamic Response of Vehicle Tatra T815 and the Ground

Computational Simulation of Dynamic Response of Vehicle Tatra T815 and the Ground IOP Conference Series: Earth and Environmental Science PAPER OPEN ACCESS Computational Simulation of Dynamic Response of Vehicle Tatra T815 and the Ground To cite this article: Jozef Vlek and Veronika

More information

Parameter Prediction and Modelling Methods for Traction Motor of Hybrid Electric Vehicle

Parameter Prediction and Modelling Methods for Traction Motor of Hybrid Electric Vehicle Page 359 World Electric Vehicle Journal Vol. 3 - ISSN 232-6653 - 29 AVERE Parameter Prediction and Modelling Methods for Traction Motor of Hybrid Electric Vehicle Tao Sun, Soon-O Kwon, Geun-Ho Lee, Jung-Pyo

More information

ANN Control of Non-Linear and Unstable System and its Implementation on Inverted Pendulum

ANN Control of Non-Linear and Unstable System and its Implementation on Inverted Pendulum Research Article International Journal of Current Engineering and Technology E-ISSN 2277 4106, P-ISSN 2347-5161 2014 INPRESSCO, All Rights Reserved Available at http://inpressco.com/category/ijcet ANN

More information

Keywords- Source coding, Huffman encoding, Artificial neural network, Multilayer perceptron, Backpropagation algorithm

Keywords- Source coding, Huffman encoding, Artificial neural network, Multilayer perceptron, Backpropagation algorithm Volume 4, Issue 5, May 2014 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Huffman Encoding

More information

Application of an Artificial Neural Network Based Tool for Prediction of Pavement Performance

Application of an Artificial Neural Network Based Tool for Prediction of Pavement Performance 0 0 0 0 Application of an Artificial Neural Network Based Tool for Prediction of Pavement Performance Adelino Ferreira, Rodrigo Cavalcante Pavement Mechanics Laboratory, Research Center for Territory,

More information

Reading Group on Deep Learning Session 1

Reading Group on Deep Learning Session 1 Reading Group on Deep Learning Session 1 Stephane Lathuiliere & Pablo Mesejo 2 June 2016 1/31 Contents Introduction to Artificial Neural Networks to understand, and to be able to efficiently use, the popular

More information

Model-Based Diagnosis of Chaotic Vibration Signals

Model-Based Diagnosis of Chaotic Vibration Signals Model-Based Diagnosis of Chaotic Vibration Signals Ihab Wattar ABB Automation 29801 Euclid Ave., MS. 2F8 Wickliffe, OH 44092 and Department of Electrical and Computer Engineering Cleveland State University,

More information

One-Hour-Ahead Load Forecasting Using Neural Network

One-Hour-Ahead Load Forecasting Using Neural Network IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 17, NO. 1, FEBRUARY 2002 113 One-Hour-Ahead Load Forecasting Using Neural Network Tomonobu Senjyu, Member, IEEE, Hitoshi Takara, Katsumi Uezato, and Toshihisa Funabashi,

More information

NONLINEAR BACKSTEPPING DESIGN OF ANTI-LOCK BRAKING SYSTEMS WITH ASSISTANCE OF ACTIVE SUSPENSIONS

NONLINEAR BACKSTEPPING DESIGN OF ANTI-LOCK BRAKING SYSTEMS WITH ASSISTANCE OF ACTIVE SUSPENSIONS NONLINEA BACKSTEPPING DESIGN OF ANTI-LOCK BAKING SYSTEMS WITH ASSISTANCE OF ACTIVE SUSPENSIONS Wei-En Ting and Jung-Shan Lin 1 Department of Electrical Engineering National Chi Nan University 31 University

More information

Identification of modal parameters from ambient vibration data using eigensystem realization algorithm with correlation technique

Identification of modal parameters from ambient vibration data using eigensystem realization algorithm with correlation technique Journal of Mechanical Science and Technology 4 (1) (010) 377~38 www.springerlink.com/content/1738-494x DOI 107/s106-010-1005-0 Identification of modal parameters from ambient vibration data using eigensystem

More information

Time-domain simulation and nonlinear analysis on ride performance of four-wheel vehicles

Time-domain simulation and nonlinear analysis on ride performance of four-wheel vehicles Journal of Physics: Conference Series Time-domain simulation and nonlinear analysis on ride performance of four-wheel vehicles To cite this article: Y S Wang et al 2008 J. Phys.: Conf. Ser. 96 012133 View

More information

Prediction of Hourly Solar Radiation in Amman-Jordan by Using Artificial Neural Networks

Prediction of Hourly Solar Radiation in Amman-Jordan by Using Artificial Neural Networks Int. J. of Thermal & Environmental Engineering Volume 14, No. 2 (2017) 103-108 Prediction of Hourly Solar Radiation in Amman-Jordan by Using Artificial Neural Networks M. A. Hamdan a*, E. Abdelhafez b

More information

Dynamic Tests on Ring Shear Apparatus

Dynamic Tests on Ring Shear Apparatus , July 1-3, 2015, London, U.K. Dynamic Tests on Ring Shear Apparatus G. Di Massa Member IAENG, S. Pagano, M. Ramondini Abstract Ring shear apparatus are used to determine the ultimate shear strength of

More information

A Wavelet Neural Network Forecasting Model Based On ARIMA

A Wavelet Neural Network Forecasting Model Based On ARIMA A Wavelet Neural Network Forecasting Model Based On ARIMA Wang Bin*, Hao Wen-ning, Chen Gang, He Deng-chao, Feng Bo PLA University of Science &Technology Nanjing 210007, China e-mail:lgdwangbin@163.com

More information

Experimental validation of a numerical model for the ground vibration from trains in tunnels

Experimental validation of a numerical model for the ground vibration from trains in tunnels Experimental validation of a numerical model for the ground vibration from trains in tunnels Qiyun Jin; David Thompson; Daniel Lurcock; Martin Toward; Evangelos Ntotsios; Samuel Koroma Institute of Sound

More information

Bearing fault diagnosis based on EMD-KPCA and ELM

Bearing fault diagnosis based on EMD-KPCA and ELM Bearing fault diagnosis based on EMD-KPCA and ELM Zihan Chen, Hang Yuan 2 School of Reliability and Systems Engineering, Beihang University, Beijing 9, China Science and Technology on Reliability & Environmental

More information

Effect of the number of poles on the acoustic noise from BLDC motors

Effect of the number of poles on the acoustic noise from BLDC motors Journal of Mechanical Science and Technology 25 (2) (211) 273~277 www.springerlink.com/content/1738-494x DOI 1.17/s1226-1-1216-4 Effect of the number of poles on the acoustic noise from BLDC motors Kwang-Suk

More information

VIBRATION ANALYSIS OF E-GLASS FIBRE RESIN MONO LEAF SPRING USED IN LMV

VIBRATION ANALYSIS OF E-GLASS FIBRE RESIN MONO LEAF SPRING USED IN LMV VIBRATION ANALYSIS OF E-GLASS FIBRE RESIN MONO LEAF SPRING USED IN LMV Mohansing R. Pardeshi 1, Dr. (Prof.) P. K. Sharma 2, Prof. Amit Singh 1 M.tech Research Scholar, 2 Guide & Head, 3 Co-guide & Assistant

More information

Electric Load Forecasting Using Wavelet Transform and Extreme Learning Machine

Electric Load Forecasting Using Wavelet Transform and Extreme Learning Machine Electric Load Forecasting Using Wavelet Transform and Extreme Learning Machine Song Li 1, Peng Wang 1 and Lalit Goel 1 1 School of Electrical and Electronic Engineering Nanyang Technological University

More information

Stockbridge-Type Damper Effectiveness Evaluation: Part II The Influence of the Impedance Matrix Terms on the Energy Dissipated

Stockbridge-Type Damper Effectiveness Evaluation: Part II The Influence of the Impedance Matrix Terms on the Energy Dissipated 1470 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 18, NO. 4, OCTOBER 2003 Stockbridge-Type Damper Effectiveness Evaluation: Part II The Influence of the Impedance Matrix Terms on the Energy Dissipated Giorgio

More information

Artificial Intelligence

Artificial Intelligence Artificial Intelligence Jeff Clune Assistant Professor Evolving Artificial Intelligence Laboratory Announcements Be making progress on your projects! Three Types of Learning Unsupervised Supervised Reinforcement

More information

2nd International Conference on Electronic & Mechanical Engineering and Information Technology (EMEIT-2012)

2nd International Conference on Electronic & Mechanical Engineering and Information Technology (EMEIT-2012) Estimation of Vehicle State and Road Coefficient for Electric Vehicle through Extended Kalman Filter and RS Approaches IN Cheng 1, WANG Gang 1, a, CAO Wan-ke 1 and ZHOU Feng-jun 1, b 1 The National Engineering

More information

Car Dynamics using Quarter Model and Passive Suspension; Part V: Frequency Response Considering Driver-seat

Car Dynamics using Quarter Model and Passive Suspension; Part V: Frequency Response Considering Driver-seat 357 Car Dynamics using Quarter Model and Passive Suspension; Part V: Frequency Response Considering Driver-seat Galal Ali Hassaan Emeritus Professor, Department of Mechanical Design & Production, Faculty

More information

Influence of electromagnetic stiffness on coupled micro vibrations generated by solar array drive assembly

Influence of electromagnetic stiffness on coupled micro vibrations generated by solar array drive assembly Influence of electromagnetic stiffness on coupled micro vibrations generated by solar array drive assembly Mariyam Sattar 1, Cheng Wei 2, Awais Jalali 3 1, 2 Beihang University of Aeronautics and Astronautics,

More information

*** (RASP1) ( - ***

*** (RASP1) (  - *** 8-95 387 0 *** ** * 87/7/4 : 8/5/7 :. 90/ (RASP).. 88. : (E-mail: mazloumzadeh@gmail.com) - - - - * ** *** 387 0.(5) 5--4-. Hyperbolic Tangent 30 70.. 0/-0/ -0- Sigmoid.(4).(7). Hyperbolic Tangent. 3-3-3-.(8).(

More information

Examining the Adequacy of the Spectral Intensity Index for Running Safety Assessment of Railway Vehicles during Earthquakes

Examining the Adequacy of the Spectral Intensity Index for Running Safety Assessment of Railway Vehicles during Earthquakes October 1-17, 8, Beijing, China Examining the Adequacy of the Spectral Intensity Index for Running Safety Assessment of Railway Vehicles during Earthquakes Xiu LUO 1 and Takefumi MIYAMOTO 1 Dr. Eng., Senior

More information

Multilayer Neural Networks

Multilayer Neural Networks Multilayer Neural Networks Multilayer Neural Networks Discriminant function flexibility NON-Linear But with sets of linear parameters at each layer Provably general function approximators for sufficient

More information

Convergence of Hybrid Algorithm with Adaptive Learning Parameter for Multilayer Neural Network

Convergence of Hybrid Algorithm with Adaptive Learning Parameter for Multilayer Neural Network Convergence of Hybrid Algorithm with Adaptive Learning Parameter for Multilayer Neural Network Fadwa DAMAK, Mounir BEN NASR, Mohamed CHTOUROU Department of Electrical Engineering ENIS Sfax, Tunisia {fadwa_damak,

More information

Rozwiązanie zagadnienia odwrotnego wyznaczania sił obciąŝających konstrukcje w czasie eksploatacji

Rozwiązanie zagadnienia odwrotnego wyznaczania sił obciąŝających konstrukcje w czasie eksploatacji Rozwiązanie zagadnienia odwrotnego wyznaczania sił obciąŝających konstrukcje w czasie eksploatacji Tadeusz Uhl Piotr Czop Krzysztof Mendrok Faculty of Mechanical Engineering and Robotics Department of

More information

Target Cascading: A Design Process For Achieving Vehicle Targets

Target Cascading: A Design Process For Achieving Vehicle Targets Target Cascading: A Design Process For Achieving Vehicle Targets Hyung Min Kim and D. Geoff Rideout The University of Michigan May 24, 2000 Overview Systems Engineering and Target Cascading Hierarchical

More information

A Hybrid Model of Wavelet and Neural Network for Short Term Load Forecasting

A Hybrid Model of Wavelet and Neural Network for Short Term Load Forecasting International Journal of Electronic and Electrical Engineering. ISSN 0974-2174, Volume 7, Number 4 (2014), pp. 387-394 International Research Publication House http://www.irphouse.com A Hybrid Model of

More information

Research Article Vehicle Vibration Analysis in Changeable Speeds Solved by Pseudoexcitation Method

Research Article Vehicle Vibration Analysis in Changeable Speeds Solved by Pseudoexcitation Method Hindawi Publishing Corporation Mathematical Problems in Engineering Volume, Article ID 87, 4 pages doi:.55//87 Research Article Vehicle Vibration Analysis in Changeable Speeds Solved by Pseudoexcitation

More information

Materials Science Forum Online: ISSN: , Vols , pp doi: /

Materials Science Forum Online: ISSN: , Vols , pp doi: / Materials Science Forum Online: 2004-12-15 ISSN: 1662-9752, Vols. 471-472, pp 687-691 doi:10.4028/www.scientific.net/msf.471-472.687 Materials Science Forum Vols. *** (2004) pp.687-691 2004 Trans Tech

More information

Intelligent Modular Neural Network for Dynamic System Parameter Estimation

Intelligent Modular Neural Network for Dynamic System Parameter Estimation Intelligent Modular Neural Network for Dynamic System Parameter Estimation Andrzej Materka Technical University of Lodz, Institute of Electronics Stefanowskiego 18, 9-537 Lodz, Poland Abstract: A technique

More information

Mr. Harshit K. Dave 1, Dr. Keyur P. Desai 2, Dr. Harit K. Raval 3

Mr. Harshit K. Dave 1, Dr. Keyur P. Desai 2, Dr. Harit K. Raval 3 Investigations on Prediction of MRR and Surface Roughness on Electro Discharge Machine Using Regression Analysis and Artificial Neural Network Programming Mr. Harshit K. Dave 1, Dr. Keyur P. Desai 2, Dr.

More information

Unit III. A Survey of Neural Network Model

Unit III. A Survey of Neural Network Model Unit III A Survey of Neural Network Model 1 Single Layer Perceptron Perceptron the first adaptive network architecture was invented by Frank Rosenblatt in 1957. It can be used for the classification of

More information

Modification of a Sophomore Linear Systems Course to Reflect Modern Computing Strategies

Modification of a Sophomore Linear Systems Course to Reflect Modern Computing Strategies Session 3220 Modification of a Sophomore Linear Systems Course to Reflect Modern Computing Strategies Raymond G. Jacquot, Jerry C. Hamann, John E. McInroy Electrical Engineering Department, University

More information

Malaysia. Lumpur, Malaysia. Malaysia

Malaysia. Lumpur, Malaysia. Malaysia Impact Force Identification By Using Modal Transformation Method For Automobile Test Rig Abdul Ghaffar Abdul Rahman,a, Khoo Shin Yee 2,b, Zubaidah Ismail 3,c, Chong Wen Tong 2,d and Siamak oroozi 4,e Faculty

More information

Neuro -Finite Element Static Analysis of Structures by Assembling Elemental Neuro -Modelers

Neuro -Finite Element Static Analysis of Structures by Assembling Elemental Neuro -Modelers Neuro -Finite Element Static Analysis of Structures by Assembling Elemental Neuro -Modelers Abdolreza Joghataie Associate Prof., Civil Engineering Department, Sharif University of Technology, Tehran, Iran.

More information

Limitations of smoothening functions for automotive vibro-impact problems

Limitations of smoothening functions for automotive vibro-impact problems Shock and Vibration 18 (2011) 397 406 397 DOI 10.3233/SAV-2010-0582 IOS Press Limitations of smoothening functions for automotive vibro-impact problems Zhiwei Zhang a,, Rajendra Singh b and Ashley R. Crowther

More information

Application of cepstrum and neural network to bearing fault detection

Application of cepstrum and neural network to bearing fault detection Journal of Mechanical Science and Technology 23 (2009) 2730~2737 Journal of Mechanical Science and Technology www.springerlin.com/content/738-494x DOI 0.007/s2206-009-0802-9 Application of cepstrum and

More information