Search

Recently, motorcycle accidents have increased as the number of motorcycle drivers has increased. Although the head and neck are the body parts most frequently injured when a motorcycle accident occurs, there is a lack of research on the protection afforded to the neck by the safety equipment used by motorcycle drivers. This study presents an airbag system that uses artificial intelligence to prevent injury to the neck of a motorcycle driver. It uses a six-axis sensor, the MPU6050 sensor, which measures acceleration and angular velocity in real time as the user moves. The angles are obtained by using the measured acceleration and angular velocity, and the accident situation is judged by AI, which analyzes the acceleration and angle data. Because data is needed for AI to learn, data by type were collected through experiments. In this study, we compare the judgement performance of a parallel neural networks–convolutional neural network and a parallel neural network.

Injuries to the head and the neck are the most frequent in the event of motorcycle accidents. But enough research has not been done to protect the neck. This paper presents an airbag system that recognizes the accident situation with Artificial Intelligence to protect the driver's neck area from motorcycle accident situations when driving. In some papers with similar themes, most of them are judged based on a critical point. However, in the case of an accident judgment using the critical point, a malfunction may occur such that the airbag operates when a similar operation is performed, or the airbag does not operate due to failing to pass the critical point at the time of an accident. Artificial intelligence was used to avoid malfunctions and inconveniences. Artificial intelligence can solve the problem of malfunction that occurs when it is judged as a critical point and can solve the inconvenience of commercialized products. The CNN presented in this paper can solve these two problems, and the accuracy of accident judgment is as high as 95.75%. Through the MPU 6050 sensor, it operates the airbag by determining the accident situation using the Artificial Intelligence that was learned in advance through the information on acceleration and angular velocity of the driver's movements that were measured in real time. To make Artificial Intelligence learn, the data were collected by dividing several types of accidents on motorcycles. In this paper, the Artificial Intelligence made by Convolutional Neural Networks (CNN) method and the Artificial Intelligence made by Neural Networks (NN) method is compared, and it is confirmed that the performance such as Test Accuracy or Train Accuracy of CNN is better.

The traffic accident occurrence rate is increasing relative to the increase in the number of people using personal mobility device (PM). This paper proposes an airbag system with a more efficient algorithm to decide the deployment of a wearable bike airbag in case of an accident. The existing wearable airbags are operated by judging the accident situations using the thresholds of sensors. However, in this case, the judgment accuracy can drop against various motions. This study used the long short-term memory (LSTM) model using the sensor values of the inertial measurement unit (IMU) as input values to judge accident occurrences, which obtains data in real time from the three acceleration-axis and three angular velocity-axis sensors on the driver motion states and judges whether or not an accident has occurred using the obtained data. The existing neural network (NN) or convolutional neural network (CNN) model judges only the input data. This study confirmed that this model has a higher judgment accuracy than the existing NN or CNN by giving strong points even in “past information” through LSTM by regarding the driver motion as time-series data.

Given their wide applicability, vehicle positioning systems are being actively studied currently. As the conventional global positioning system (GPS) cannot determine the vehicle position inside tunnels and other GPS-shaded areas, we propose to improve vehicle positioning using a fuzzy logic system that can retrieve the position of a moving vehicle in such areas. The proposed system analyzes the chromaticity and frequency components of active LED illumination, which is being commonly installed at locations like inside tunnels. We verify the system performance in a test environment that resembles a tunnel. As the proposed method uses the chromaticity and frequency of LED lights for vehicle positioning, it is suitable for areas where GPS reception is poor or unavailable.

The common way to obtain a stereoscopic image of a subject at a distance is to place two cameras on the parallel axis rather than crossing axis. To find the IAD and maximum focal length, left and right images are obtained by varying the IAD of cameras and the focal length of the camera lens and the depth budget for the obtained images is analyzed through post production. Then, the database for IAD and focal length of the camera lens with the depth range that does not cause visual fatigue and visual discomfort are developed. These data are used to design fuzzy control and deduce the IAD and focal length of the camera lens to shoot a subject at a distance, and the function of the fuzzy control is confirmed through the actual shooting within the range of deduced IAD and focal length of the camera lens. 키워드 : 원거리 입체영상, 수평식 리그, 깊이범위, 깊이 총량, 퍼지추론 Key word : Remote Stereoscopic Photographs, Parallel Rig, IAD, Depth Range, Depth Budget, Fuzzy InferenceJournal of the Korea Institute of Information andCommunication Engineering

This paper describes a new method for a digital gas metal arc welding (GMAW) system. The GMAW system is an arc welding process that incorporates the GMAW power source (PS-GMAW) with a wire feed unit (WFU). The PS-GMAW requires an electric power of constant voltage. A constant magnitude is maintained for the arc current by controlling the wire-feed speed of the WFU. A mathematical model is derived, and a self-tuning fuzzy proportional-integral-derivative (PID) controller is designed and applied to control the welding current. The electrode wire feeding mechanism with this controller is driven by a DC motor, which can compensate for both the molten part of the electrode and undesirable fluctuations in the arc length during the welding process. By accurately maintaining the output welding current and welding voltage at constant values during the welding process, excellent welding results can be obtained. Simulation and experimental results are shown to prove the effectiveness of the proposed controller.

This paper focuses on the feasibility of IEEE 802.11b wireless LAN for real-time wireless industrial network. In general, it has been known that IEEE 802.11b is not suitable for industrial networking because its medium access control method is the contention-based CSMA/CA (Carrier Sensing Multiple Access with Collision Avoidance) that exhibits unstable performance under heavy traffic and unbounded delay distribution. As an attempt to circumvent the probabilistic nature of IEEE 802.11b, this paper presents an enhanced four-layer architecture using the Network Driver Interface Specification (NDIS) and a virtual polling algorithm. Also, the enhanced four-layer architecture's performance is compared with that of conventional IEEE 802.11b. Based on the experimental results, it is found that the enhanced four-layer architecture for IEEE 802.11b is a very promising alternative for wireless industrial networking.

In this paper, a sensorless vector control method is introduced where rotor speed is estimated based on extended Kalman filter (EKF). The EKF is a recursive optimum stochastic state estimator. The proposed EKF is designed to obtain a small estimation speed error in both transient and steady state over a wide speed range. A major challenge at very low and zero speed is the lost coupling effect from the rotor to the stator, which makes the information on rotor variables unobservable on the stator side. To solve this problem, in the proposed EKF, the load torque and the rotor speed with unmeasurable information are estimated simultaneously. The rotor speed is considered via the equation of motion and the estimation of load torque, on the other hand, is performed as a constant parameter. Hardware and software for experiment of the AC induction motor drive are introduced. The experimental results are presented to show the effectiveness of the proposed system.