• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.8 no.2
• /
• pp.60-66
• /
• 2009

The acceleration noise is valuable index to monitor traffic stability. However, the previous study was performed for the acceleration noise of individual vehicle. The consideration of the acceleration noise for vehicle in the network has not been studied yet. This paper proposes a new macroscopic traffic flow monitoring method based on applying network acceleration noise.

• Proceedings of the KIEE Conference
• /
• 2006.07d
• /
• pp.1858-1859
• /
• 2006

This paper presents a new fuzzy-neural-network based interacting multiple model (FNNBIMM) algorithm for tracking a maneuvering target. To effectively handle the unknown target acceleration, this paper regards it as additional noise, time-varying variance to target model. Each sub model characterized by the variance of the overall process noise, which is obtained on the basis of each acceleration interval. Since it is hard to approximate this time-varying variance adaptively owing to the unknown acceleration, the FNN is utilized to precisely approximate this time-varying variance. The gradient descendant method is utilized to optimize each FNN. To show the feasibility of the proposed algorithm, a numerical example is provided.

• Smart Structures and Systems
• /
• v.31 no.5
• /
• pp.501-516
• /
• 2023

Dynamic irregularity and acceleration of bridges subjected to high-speed trains provide crucial information for comprehensive evaluation of the health state of under-track structures. This paper proposes a novel approach for real-time estimation of vertical track dynamic irregularity and bridge acceleration using deep generative adversarial network (GAN) and vibration data from in-service train. The vehicle-body and bogie acceleration responses are correlated with the two target variables by modeling train-bridge interaction (TBI) through least squares generative adversarial network (LSGAN). To realize supervised learning required in the present task, the conventional LSGAN is modified by implementing new loss function and linear activation function. The proposed approach can offer pointwise and accurate estimates of track dynamic irregularity and bridge acceleration, allowing frequent inspection of high-speed railway (HSR) bridges in an economical way. Thanks to its applicability in scenarios of high noise level and critical resonance condition, the proposed approach has a promising prospect in engineering applications.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.4
• /
• pp.472-478
• /
• 2006

This paper presents a new fuzzy-neural-network based interacting multiple model (FNNBIMM) algorithm for tracking a maneuvering target. To effectively handle the unknown target acceleration, this paper regards it as additional noise, time-varying variance to target model. Each sub model characterized by the variance of the overall process noise, which is obtained on the basis of each acceleration interval. Since it is hard to approximate this time-varying variance adaptively owing to the unknown acceleration, the FNN is utilized to precisely approximate this time-varying variance. The error back-propagation method is utilized to optimize each FNN. To show the feasibility of the proposed algorithm, a numerical example is provided.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.39 no.12
• /
• pp.1306-1316
• /
• 1990

A controller using a multilayer neural network is proposed to the dynamic control of a PUMA 560 robot arm. This controller is developed based on an error back-propagation (BP) neural network. Since the neural network can model an arbitrary nonlinear mapping, it is used as a commanded feedforward torque generator. A Proportional Derivative (PD) feedback controller is used in parallel with the feedforward neural network to train the system. The neural network was trained by the current state of the manipulator as well as the PD feedback error torque. No a priori knowledge on system dynamics is needed and this information is rather implicitly stored in the interconnection weights of the neural network. In another experiment, the neural network was trained with the current, past and future positions only without any use of velocity sensors. Form this thim window of position values, BP network implicitly filters out the velocity and acceleration components for each joint. Computer simulation demonstrates such powerful characteristics of the neurocontroller as adaptation to changing environments, robustness to sensor noise, and continuous performance improvement with self-learning.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2004.11a
• /
• pp.619-624
• /
• 2004

In this research, the structural health monitoring method using wavelet packet analysis and artificial neural network (ANN) is developed. Wavelet packet Transform (WPT) is applied to the response acceleration of a 3 element-cantilever beam which is subjected to impulse load and Gaussian random load to decompose the response signal, then the energy of each component is calculated. The first ten largest components in magnitude among the decomposed components are selected as input to an ANN to identify the damage location and severity. This method successfully predicted the amount of damage in the structure when the structure is subjected to impulse load. However, when the beam is subjected to Gaussian random load which can be considered as ambient vibration it did not yield satisfactory results. This method is applicable to structures such as machinery gears that are subjected to repetitive loads.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2003.11a
• /
• pp.619-623
• /
• 2003

In this research, Forced Vibration Test(FVT) on a cable stayed bridge was conducted to examine the validity of the frequency domain pattern recognition method using signal anomaly index and artificial neural network. The considering structure, Samchunpo Bridge, located in Sachun-Shi, Kyungsangnam-Do, is a cable stayed bridge with the 436 meter span. The excitation force was induced by a sudden braking of a fully loaded truck, and vertical acceleration signals were acquired at 14 points. The initial 2-dimensional FE-model was developed from the design documents to prepare the training sets for the artificial neural network, and then the model calibration was performed with the field test data. As a result of the model calibration, we obtained the FFT spectrums from the model simulation, which was similar to those from the vibration test. These tests and the simulation data will be used fur the structural identification using arbitrarily added masses to the bridge.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2006.05a
• /
• pp.1437-1442
• /
• 2006

The international competition in car markets has continuously required the research about the sound quality of a car. The domestic carmakers have also invested a lot of money for the research and development of interior sound quality of passenger cars. Therefore, the aim of this research is to predict the customer's evaluation of a new vehicle. There are two major research works to achieve this goal in this research. The first one is to search questionnaires about the sound quality, which customers prefer, to identify the relationship between these questionnaires and sound metrics that is a psychoacoustics parameters, and to development sound indexes for the questionnaires. All tests for this work is proceed on the road test during acceleration. The second one is to balance the sound component (engine noise, booming noise, road noise and wind noise) of a passenger. This wok will be tested on the constant speed. All of research results will be contributed to the development of brand sound quality of a new passenger car.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.9
• /
• pp.1725-1730
• /
• 2017

The acceleration sensor and the gyroscopic sensor are used as representative sensors to detect repetitive motion and have been used to analyze various sporting components. However, both sensors have problems with noise sensitivity and accumulation of errors. There have been attempts to use two sensors together to overcome hardware problems. The complementary filter has shown successful results in mitigating the problems of both sensors by minimizing the disadvantages of accelerometer and gyroscope sensors and maximizing their advantages. In this paper, we proposed a modified method using neural network to reduce variable. The neural network is an algorithm that can precisely measure even in unexpected environments or situations by pre-learning the number of various cases. The proposed method applies a Neural Network by dividing the repetitive motion into three sections, the first, the middle and the end. As a result, the recognition rate is 96.35%, 98.77%, 96.92% and the accuracy is 97.18%.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2010.10a
• /
• pp.236-241
• /
• 2010

This paper presents a wear diagnosis method for centrifugal impellers by using an accelerometer. The features are calculated from raw and wavelet transformed signals with several statistical methods applied in time or frequency domains. From the effectiveness coefficient test, it is shown that 7th level of wavelet transformed signal is suitable for wear classification problems. A neural network with 5 feature sets is applied to diagnose the wear magnitude of pump impellers. The verification result reveals that high accuracy for the wear diagnosis of impellers can be obtained by using wavelet features transformed from acceleration signals.