JICE 2017 December

JICE December 2017 academic articles

Showing all 10 results

  • pdf-icon-round

    Deep Learning based Handwritten Digit Recognition

    $15.00

    Abstract: Neural network and depth learning have been widely used in the field of image processing. Good recognition results are often required for complex network models. But the complex network model makes training difficult and takes a long time. In order to obtain a higher recognition rate with a simple model, the BP neural network and the convolutional neural network are studied separately and verified on the MNIST data set. In order to improve the recognition results further, a combined depth network is proposed and validated on the MNIST dataset. The experimental results show that the recognition effect of the combined depth network is obviously better than that of a single network. A more accurate recognition result is achieved by the combined network.


  • pdf-icon-round

    Water Bloom Warning Model Based on Random Forest

    $15.00

    Abstract: Based on the random forest classification algorithm, a warning model of water bloom is proposed. Using the collected data, Select the water quality, meteorological factors which like Chlorophyll a (Chl-a), water temperature (T), PH, nitrogen and phosphorus ratio (TN:TP), chemical oxygen demand (COD), total nitrogen (TN), total phosphorus (TP), dissolved oxygen Light (E) and so on as the impact factor and use them establish a warning model for Water bloom. And compared with the prediction accuracy of neural network model and SVM model. The results show that the water bloom warning model is established by using stochastic forest classification algorithm, the prediction accuracy is slightly higher than other algorithms. And the random forest algorithm has the characteristics of high robustness, China good performance, strong practicability can effectively carry out water bloom early warning.


  • pdf-icon-round

    EKF based Sliding Mode Control for a Quadrotor Attitude Stabilization

    $15.00

    Abstract: In recent years, the interest in unmanned aerial vehicles (UAVs) has been increasing around the world. These vehicles are used in various applications from military operations to civilian tasks. Quadrotor, also called as a quadcopter, is one of the different types of UAVs. Quadrotor can fly more stable than helicopter and the flight control is more convenient. In UAVs, the most basic and salient point is the attitude control for stability. This paper estimates quadrotor’s attitude by extended Kalman filter (EKF) and presents the design procedure of a sliding mode control (SMC) to focus on stabilization. The performance and effectiveness of the proposed system are verified through a simulation study.


  • pdf-icon-round

    Study of detection algorithm of pedestrians by image analysis with a crossing request when gazing at a pedestrian crossing signal

    $15.00

    Abstract: Despite the advancement of information and transportation systems, inconvenient pedestrian crossing buttons remain common. In accordance with intelligent transportation systems (ITS), it is necessary to improve pedestrian crossing systems. Therefore, in this study, the proposed system adopts signal gaze, which is more natural compared to pressing a pedestrian crossing button, as a crossing request. A compact camera is inserted in a traffic light to view the other side of the crosswalk. The image data is analyzed in real time to identify all people who have a crossing request. An algorithm with three detectors using Haar-like feature quantities was developed and an evaluation experiment was conducted, considering three conditions: daytime, nighttime, and shade. The detection rate of crossing requests was 100% within 5 s. Although the detection rate was extremely high, there was a problem of incorrectly detecting non-humans. Therefore, in this research, we evaluated the system when detecting non-humans in order to determine the causes. As a result, it became clear that the detection rate changes rapidly depending on the waiting time for a traffic light and also when crossing the crosswalk; however, the system continues to detect the incorrectly detected background.


  • pdf-icon-round

    Emergence of Robust Cooperative States by Iterative Internalizations of Opponents’ Personalized Values in Minority Game

    $15.00

    Abstract: Adolescence is a period in which individuals begin facing some challenging choices. Through these choices, and social interactions with peers, adolescent individuals develop their “personalized values” that are the foundations of  their actions. It is known that the adolescent brains have high plasticity, and that the adolescent brains change dynamically, other than that of an adult brain. This study discusses the type of behavior that emerges from adolescents, as
    well as adult individuals with elevated plasticities. To realize this, we adopt the minority game (MG), which is one of the tasks concerning the choice described above. We implement Elman-nets with different learning rates that express different degrees of the plasticity as the player models in the MG. Our simulation results showed that it is a possibility that robust cooperative states emerge by iteratively internalizing the opponent players’ personalized values among players that have a network of reference relationship, irrespective of their varying degrees of plasticity.


  • pdf-icon-round

    A Sanshin Musical Performance Assistive Device for People with Physical Disabilities of the Extremities

    $15.00

    Abstract: This paper describes a sanshin musical performance assistive device for people with physical disabilities of the extremities. Sanshin is an Okinawan three-stringed musical instrument. We had developed portable assistive devices to press strings against board of sanshin and its controllers. In the experiments, some participants played the sanshin using the assistive devices. It was found from the experimental results that two persons with muscular dystrophies could appropriately performed music using the sanshins with the previous assistive devices, and that two able-bodied persons could smoothly and steadily play the sanshins using the improved assistive device with the photoelectric sensor-based controller


  • pdf-icon-round

    TDOA based Geolocation using IRLS Algorithm

    $15.00

    Abstract: The geolocation method using the wireless signal processing system is widely used in many industrial areas. The time difference of arrival (TDOA) method is one of the most commonly used geolocation methods. The TDOA signal based geolocation system can estimate the position of a mobile object by using at least three base stations in two dimensional spaces. In geolocation problem, the precise estimation of mobile’s position is the most significant issue.The measurement noise that is contained in measured TDOA data causes the estimation inaccuracy in a mobile  geolocation. In this paper, the objective function that represents the scalar error of position estimation is formulated using the concept of p Lp -norm approximation. Also, we suggest the iterative reweighted least square (IRLS) scheme for minimizing of the objective function. The optimal solution can be obtained using the limited measurement data through the reweighted iteration process of the IRLS scheme.

    Keywords: Geolocation; Time difference of arrival; Iterated reweighted least square; Obtimization
    method


  • pdf-icon-round

    Analyzing the Advantages of Utilizing State Representations in a Probabilistic Reversal Learning Task

    $15.00

    Abstract: Cognitive flexibility is the ability to adaptively change behaviors in the face of dynamically changing circumstances. To explore the neural basis and computational account of this ability, a probabilistic reversal learning task was employed as the experimental paradigm. Recent studies suggest that a subject may utilize not only a reward history but also a “state representation” of a task to successfully solve one. However, the specific advantages or impact of state representations in task solving are still not fully understood. In this study, we investigated this matter by computer simulations, in which we used two types of reinforcement learning models, a model with state representations and one without. As a result of the simulations, we found that state representations make a learning agent robust against an increasingly difficult task, especially when the number of sampling time in each state is reduced. Based on the results, we propose a hypothesis for the acquisition process of state representations and discuss the experimental design to test it.


  • pdf-icon-round

    Developing A Simplified Maintenance & Rehabilitation Activity Prioritization Tool for Afghanistan Roads

    $15.00

    Abstract: Roads are one of the most important factor of life, and maintenance & rehabilitation of them are very vital and challenging for a country. Afghanistan is one of those countries which face the challenges of low-budget, computerized office works and skilled personnel. Regarding to budget limitation pavement maintenance and rehabilitation activity prioritization is obligatory. Currently, a technology based, and simplified maintenance activity prioritization tool are an essential need of the country. The aim of this research is to develop a tool which prioritize the maintenance and rehabilitation activities by considering some factors such as pavement condition index, road width, traffic volume, residential importance as well as maintenance and rehabilitation cost. Since characterizing a model that presents each one of those variables was difficult, a simplified model named TOPSIS was denoted for the issue of prioritization. TOPSIS model lets you have a more precise ranking for the outcome. Considering the problem, Visual Basic have the ability to easily code any type of model and present a graphical display of the model. A source code was developed and Visual Basic was used for computations coding, graphical display of results and generating reports. The developed model indicates that more than two criteria/weighs are very important for prioritizing the alternatives/activities. The developed tool can prioritize the maintenance and rehabilitation activities and generate different database for further use in ArcGIS.


  • pdf-icon-round

    A Comparative Study on the Performances of Q-Learning and Neural Q-Learning Agents toward Analysis of Emergency of Communication

    $15.00

    Abstract: In this paper, we suppose the gesture theory that is one theory on the origin of language, which tries to establish that speech originated from gestures. Based on the theory,  we assume that “actions” having some purposes can be used as “symbols” in the communication through a learning process. The purpose of this study is to clarify what abilities of agents and what conditions are necessary to acquire usages of the actions as the symbols. To investigate them, we adopt a collision avoidance game and compare the performances of Q-learning agents with that of Neural Q-learning agents. In our simulation, we found that the Neural Q-learning agent’s ability to reach the goal place is higher than the Q-learning agent’s one. In contrast, the Neural Q-learning agent’s ability to avoid collisions is lower than the Q-learning agent’s one. If the inconsistencies in the learning data sets of the Neural Q-learning agent, however, can be resolved, the agent has enough potential to improve its ability for collision avoidance. Therefore, we conclude that the most suitable agent to analyze the emergence of communication is the Neural Q-learning agent who changed a feed forward type neural network into a recurrent type neural network that can resolve the inconsistencies in the learning data sets.