• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• v.1
• /
• pp.277-282
• /
• 2006

It is well known that Kalman filtering is an optimal real-time data fusion method for GPS/INS integration. However, it has some limitations in terms of stability, adaptability and observability. A Kalman filter can perform optimally only when its dynamic model is correctly defined and the noise statistics for the measurement and process are completely known. It is found that estimated Kalman filter states could be influenced by several factors, including vehicle dynamic variations, filter tuning results, and environment changes, etc., which are difficult to model. Neural networks can map input-output relationships without apriori knowledge about them; hence a proper designed neural network is capable of learning and extracting these complex relationships with enough training. This paper presents a GPS/INS integrated system that combines Kalman filtering and neural network algorithms to improve navigation solutions during GPS outages. An Extended Kalman filter estimates INS measurement errors, plus position, velocity and attitude errors etc. Kalman filter states, and gives precise navigation solutions while GPS signals are available. At the same time, a multi-layer neural network is trained to map the vehicle dynamics with corresponding Kalman filter states, at the same rate of measurement update. After the output of the neural network meets a similarity threshold, it can be used to correct INS measurements when no GPS measurements are available. Selecting suitable inputs and outputs of the neural network is critical for this hybrid method. Detailed analysis unveils that some Kalman filter states are highly correlated with vehicle dynamic variations. The filter states that heavily impact system navigation solutions are selected as the neural network outputs. The principle of this hybrid method and the neural network design are presented. Field test data are processed to evaluate the performance of the proposed method.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2011.10a
• /
• pp.396-397
• /
• 2011

This paper has proposed a multi-channel low pass filter (LPF) for LTE-Advanced systems. The proposed LPF is an active-RC 5th chebyshev topology with three cut-off frequencies of 5 MHz, 10 MHz, and 40 MHz. A 3-bit tuning circuit has been adopted to prevent variations of each cut-off frequency from process, voltage, and temperature (PVT). To achieve a high cut-off frequency of 40 MHz, an operational amplifier used in the proposed filter has employed a PMOS cross-connection load with a negative impedance. A proposed filter has been implemented in a $0.13-{\mu}m$ CMOS technology and consumes 20.2 mW with a 1.2V supply voltage.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.2
• /
• pp.247-268
• /
• 2011

Due to the application-specific nature of wireless sensor networks, the sensitivity to such a requirement as data reporting interval varies according to the type of application. Such considerations require an application-specific, parameter tuning paradigm allowing us to maximize energy conservation prolonging the operational network lifetime. In this paper, we propose a reporting interval adaptive, sensor control platform for energy-saving data gathering in wireless sensor networks. The ultimate goal is to extend the network lifetime by providing sensors with high adaptability to application-dependent or time-varying, reporting interval requirements. The proposed sensor control platform is based upon a two phase clustering (TPC) scheme which constructs two types of links within each cluster - namely, direct link and relay link. The direct links are used for control and time-critical, sensed data forwarding while the relay links are used only for multi-hop data reporting. Sensors opportunistically use the energy-saving relay link depending on the user reporting, interval constraint. We present factors that should be considered in deciding the total number of relay links and how sensors are scheduled for sensed data forwarding within a cluster for a given reporting interval and link quality. Simulation and implementation studies demonstrate that the proposed sensor control platform can help individual sensors save a significant amount of energy in reporting data, particularly in dense sensor networks. Such saving can be realized by the adaptability of the sensor to the reporting interval requirements.

• Journal of Internet Computing and Services
• /
• v.11 no.1
• /
• pp.1-13
• /
• 2010

It is not easy to set RED(Random Early Detection) parameter according to environment in managing Network Device. Especially, it is more difficult to set parameter in the case of maintaining the constant service rate according to the change of environment. In this paper, we hypothesize the router that has Multi-core CPU in output queue and propose AI RED(Artificial Intelligence RED), which directly induces Genetic Algorithm of Artificial Intelligence in the output queue that is appropriate to the optimization of parameter according to RED environment, which is automatically adaptive to workload. As a result, AI RED Is simpler and finer than FuRED(Fuzzy-Logic-based RED), and RED parameter that AI RED searches through simulations is more adaptive to environment than standard RED parameter, providing the effective service. Consequently, the automation of management of RED parameter can provide a manager with the enhancement of efficiency in Network management.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.11
• /
• pp.29-39
• /
• 2019

As mobile robot applications increase in real life, the need of low cost autonomous driving are gradually increasing. We propose a novel vision-based real-time lane detection and tracking system that supports autonomous driving of mobile robots in constrained tracks which are designed considering indoor driving conditions of mobile robots. Considering the processing of lanes with various shapes and the pre-adjustment of operation parameters, the system structure with multi-operation modes are designed. In parameter tuning mode, thresholds of the color filter is dynamically adjusted based on the geometric property of the lane thickness. And in the unstable input mode of curved tracks and the stable input mode of straight tracks, lane feature pixels are adaptively extracted based on the geometric and temporal characteristics of the lanes and the lane model is fitted using the least-squared method. The track centerline is calculated using lane models and the motion model is simplified and tracked by a linear Kalman filter. In the driving experiments, it was confirmed that even in low-performance robot configurations, real-time processing produces the accurate autonomous driving in the constrained track.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.18 no.3
• /
• pp.106-118
• /
• 2019

In recent years, research and development on image object recognition using artificial intelligence have been actively carried out, and it is expected to be used for road maintenance. Among them, artificial intelligence models for object detection of road surface are continuously introduced. In order to develop such object recognition algorithms, a backbone network that extracts feature maps is essential. In this paper, we will discuss how to select the appropriate neural network. To accomplish it, we compared with 4 different deep neural networks using 6,000 road surface damage images. Based on three evaluation methods for analyzing characteristics of neural networks, we propose a method to determine optimal neural networks. In addition, we improved the performance through optimal tuning of hyper-parameters, and finally developed a light backbone network that can achieve 85.9% accuracy of road surface damage classification.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.31 no.6
• /
• pp.301-308
• /
• 2018

This paper presents an optimal design method of a hybrid structural control system considering multi-hazard. Unlike a typical structural control system in which one system is designed for one specific type of hazard, a simultaneous optimal design method for both active and passive control systems is proposed for the mitigation of seismic and wind induced vibration responses of structures. As a numerical example, an optimal design problem is illustrated for a hybrid mass damper(HMD) and 30 viscous dampers which are installed on a 30 story building structure. In order to solve the optimization problem, a self-adaptive Harmony Search(HS) algorithm is adopted. Harmony Search algorithm is one of the meta-heuristic evolutionary methods for the global optimization, which mimics the human player's tuning process of musical instruments. A self-adaptive, dynamic parameter adjustment algorithm is also utilized for the purpose of broad search and fast convergence. The optimization results shows that the performance and effectiveness of the proposed system is superior with respect to a reference hybrid system in which the active and passive systems are independently optimized.

• Earthquakes and Structures
• /
• v.20 no.3
• /
• pp.353-364
• /
• 2021

Natural hazards like earthquakes, high winds, and tsunami are a threat all the time for multi-story structures. The environmental forces cannot be clogged but the structures can be prevented from these natural hazards by using protective systems. The structural control can be achieved by using protective systems like the passive, active, semi-active, and hybrid protective systems; but the semi-active protective system has gained importance because of its adaptability to the active systems and reliability of the passive systems. Therefore, a semi-active protective system for the earthquake forces has been adopted in this work. Magneto-Rheological (MR) damper is used in the structure as a semi-active protective system; which is connected to the current driver and proposed controller. The Proportional Integral Derivative (PID) controller and reliable PID controller are two proposed controllers, which will actuate the MR damper and the desired force is generated to mitigate the vibration of the structural response subjected to the earthquake. PID controller and reliable PID controller are designed and tuned using Ziegler-Nichols tuning technique along with the MR damper simulated in Simulink toolbox and MATLAB to obtain the reduced vibration in a three-story benchmark structure. The earthquake is considered to be uncertain; where the proposed control algorithm works well during the presence of earthquake; this paper considers robustness to provide satisfactory resilience against this uncertainty. In this work, two different earthquakes are considered like El-Centro and Northridge earthquakes for simulation with different controllers. In this paper performances of the structure with and without two controllers are compared and results are discussed.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.12
• /
• pp.3904-3922
• /
• 2022

As a research hotspot, pedestrian detection has a wide range of applications in the field of computer vision in recent years. However, current pedestrian detection methods have problems such as insufficient detection accuracy and large models that are not suitable for large-scale deployment. In view of these problems mentioned above, a lightweight pedestrian detection and early warning method using a new model called you only look once (Yolov5) is proposed in this paper, which utilizing advantages of Yolov5s model to achieve accurate and fast pedestrian recognition. In addition, this paper also optimizes the loss function of the batch normalization (BN) layer. After sparsification, pruning and fine-tuning, got a lot of optimization, the size of the model on the edge of the computing power is lower equipment can be deployed. Finally, from the experimental data presented in this paper, under the training of the road pedestrian dataset that we collected and processed independently, the Yolov5s model has certain advantages in terms of precision and other indicators compared with traditional single shot multiBox detector (SSD) model and fast region-convolutional neural network (Fast R-CNN) model. After pruning and lightweight, the size of training model is greatly reduced without a significant reduction in accuracy, and the final precision reaches 87%, while the model size is reduced to 7,723 KB.

• Proceedings of the Korean Society of Crop Science Conference
• /
• 2022.10a
• /
• pp.88-88
• /
• 2022

Detection of stress responses in crops is important to diagnose crop growth and evaluate yield. Also, the multi-spectral sensor is effectively known to evaluate stress caused by nutrient and moisture in crops or biological agents such as weeds or diseases. Therefore, in this experiment, multispectral images were taken by an unmanned aerial vehicle(UAV) under field condition. The experiment was conducted in the long-term fertilizer field in the National Institute of Crop Science, and experiment area was divided into different status of NPK(Control, N-deficiency, P-deficiency, K-deficiency, Non-fertilizer). Total 11 vegetation indices were created with RGB and NIR reflectance values using python. Variations in nutrient content in plants affect the amount of light reflected or absorbed for each wavelength band. Therefore, the objective of this experiment was to evaluate vegetation indices derived from multispectral reflectance data as input into machine learning algorithm for the classification of nutritional deficiency in rice. RandomForest model was used as a representative ensemble model, and parameters were adjusted through hyperparameter tuning such as RandomSearchCV. As a result, training accuracy was 0.95 and test accuracy was 0.80, and IPCA, NDRE, and EVI were included in the top three indices for feature importance. Also, precision, recall, and f1-score, which are indicators for evaluating the performance of the classification model, showed a distribution of 0.7-0.9 for each class.