• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.8
• /
• pp.1996-2015
• /
• 2023

High-definition (HD) maps can provide precise road information that enables an autonomous driving system to effectively navigate a vehicle. Recent research has focused on leveraging semantic segmentation to achieve automatic annotation of HD maps. However, the existing methods suffer from low recognition accuracy in automatic driving scenarios, leading to inefficient annotation processes. In this paper, we propose a novel semantic segmentation method for automatic HD map annotation. Our approach introduces a new encoder, known as the convolutional transformer hybrid encoder, to enhance the model's feature extraction capabilities. Additionally, we propose a multi-level fusion module that enables the model to aggregate different levels of detail and semantic information. Furthermore, we present a novel decoupled boundary joint decoder to improve the model's ability to handle the boundary between categories. To evaluate our method, we conducted experiments using the Bird's Eye View point cloud images dataset and Cityscapes dataset. Comparative analysis against stateof-the-art methods demonstrates that our model achieves the highest performance. Specifically, our model achieves an mIoU of 56.26%, surpassing the results of SegFormer with an mIoU of 1.47%. This innovative promises to significantly enhance the efficiency of HD map automatic annotation.

• Journal of Biosystems Engineering
• /
• v.27 no.6
• /
• pp.537-546
• /
• 2002

In this study an automatic soil hardness measuring system mountable on agricultural tractors was developed to improve the accuracy of manual soil hardness testers by a constant penetrating rate, right direction of the cone-penetrometer and the isolation of vibration from the operator. This was necessary to supply similar experimental condition for performance test of new model and comparative experiment. The results of the study are summaried as follows; 1. The system consisted of a sensing part of soil hardness, a driving part of the measuring system and an attaching part between the tractor and the measuring system. 2. The allowable limit value of the system developed was set to 392N to protect from breaking the serve motor and the coupling used in this system. 3. The driving shaft penetrated into soil by 0.3m to measure soil hardness. The soil hardness was measured at the depth of 0.3m from the soil surface but the penetrating work was stopped and the driving shaft was pulled out to protect the system when the value of the soil hardness was too big on foreign substances like stones or straws. 4. Two values measured by automatic measuring system developed in this research and manual penetrometer were compared by statistics hypothesis testing method. When two people measured the soil hardness at the depth of 0.1 and 0.15m by manual cone penetrometer, there was no relationship between two values by two people but the values at the same depths by automatic measuring system developed showed similarity. The automatic system, therefore, developed in this research was proper for measuring soil hardness.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.15 no.8
• /
• pp.4742-4748
• /
• 2014

A semi-automatic clutch was developed for drivers of vehicles with manual transmission. The clutch is operated by pressing a switch on the gear stick without stepping on a clutch pedal when the driver wants to shift gears. To automatic control a clutch, driving information is provided by sensors installed under the vehicle. On the other hand, sensors are prone to failure under severe driving conditions and a long time is needed to install or repair these sensors in the vehicle. In this paper, a semi-automatic clutch that received driving information by CAN communication from the ECU was developed and a pneumatic actuator was used to operate the clutch. The semi-automatic clutch by a pneumatic cylinder was operated with a supply air pressure of more than 3bar.

• Applied Chemistry for Engineering
• /
• v.24 no.3
• /
• pp.274-278
• /
• 2013

Automatic transmission fluid (ATF) is used as an automatic transmission in the vehicle or as a characterized fluid for automatic transmission. Recently, vehicle manufacturers usually guarantee for changing fluids over 80000~100000 km mileage or no exchange. However, most drivers usually change ATF below every 50000 km driving distance when driving in Republic of Korea according to a survey from the Korea Institute of Petroleum Management which can cause both a serious environmental contamination by the used ATF and an increase in the cost of driving. In this study, various physical properties such as flash point, pour point, kinematic viscosity, dynamic viscosity at low temperature, total acid number and four-ball test were investigated for both fresh ATF and used ATF after the actual vehicle driving distance of 50000 km and 100000 km. It was shown that most physical properties were suitable for the specification of ATF, but the foam characteristics of the used oil after running 100000 km was unsuitable for the specification of fresh ATF. Therefore, the exchange cycle of ATF every 80000~100000 km driving distance is recommended considering great positive contributions to preventing environmental pollution and reducing driving cost.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.717-722
• /
• 2003

A driving simulator is a computer-controlled tool to study an interface between a driver and vehicle response by enabling the driver to participate in judging vehicle characteristics. Using the driving simulator, human factor study, vehicle system development and other research can be effectively done under controllable, reproducible and non-dangerous conditions. An Adaptive Cruise Control (ACC) system is generally regarded as a system that can be achieved in the near future without the demanding infrastructure components and technologies. ACC system is an automatic vehicle following system with no human engagement in the longitudinal vehicle direction. And the influence of the driver is substantial in developing the system. Driving characteristic is very different according to the accident riskiness, gender, age and so on. In this research, experiments have been carried out to investigate driving characteristics with the ACC system, using a driving simulator. Participants are 21 male and 19 female. Driving characteristics such as preferred headway-time, lane keeping ability, eye direction, and head movement have been observed and compared between the driving with ACC and the driving without ACC.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.1
• /
• pp.33-39
• /
• 2014

Automatic driving needs many functions such as the obstacle recognition, the lane recognition, and the lane change, etc. In this paper, we realized a system which automatically drove the three-kinds of vehicle driving course, to introduce and apply the concept of 'color recognition' that expands the scope of 'lane recognition' for vehicle driving. We made the reduced each course compared with RC(Radio Control) car size, and controlled the steering considering the position and the slope of the detection line and the speed. Because the RC car does not have the brake function, we consider the speed and the position of the detection line to stop the RC car.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.16 no.2
• /
• pp.180-189
• /
• 2017

Some people who take a driver's license test retake it again because of license cancellation, but most of them take the test for the first time to drive the car. Driving a car is directly linked to life, and the initial correct driving habits are more important than anything else. In particular, it is very important to obtain a license by evaluating the correct driving ability based on objective and fair standards when learning the first driving, because many people acquire a driving license while entering the society for the first time. In this paper, we propose the S / W design and its main functions that can emit high quality drivers through efficient, fair and accurate automated scoring. Through this, it is proposed to improve the automatic grading driver's license system, to prevent traffic accidents, and to reduce traffic accidents through proper driving.

• Structural Engineering and Mechanics
• /
• v.63 no.6
• /
• pp.761-769
• /
• 2017

Accurately extracting the axle distribution information of a passing vehicle from bridge dynamic responses experiences a key and challenging step in non-pavement bridge weigh-in-motion (BWIM). In this article, the wavelet transformation is adopted and the wavelet coefficient curve is used as a substitute for dynamic response. The driving frequency is introduced and expanded to multi-axle vehicle, and the wavelet coefficient curve on specific scale corresponding to the driving frequency is confirmed to contain obvious axle information. On this basis, an automatic method for axle distribution information identification is proposed. The specific wavelet scale can be obtained through iterative computing, and the false peaks due to bridge vibration can be eliminated through cross-correlation analysis of the wavelet coefficients of two measure points. The integrand function that corresponds to the maximum value of the cross-correlation function is used to identify the peaks caused by axles. A numerical application of the proposed axle information identification method is carried out. Numerical results demonstrate that this method acquires precise axle information from the responses of an axle-insensitive structure (e.g., girder) and decreases the requirement of sensitivity structure of BWIM. Finally, an experimental study on a full-scale simply supported bridge is also conducted to verify the effectiveness of this method.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.13 no.2
• /
• pp.140-146
• /
• 2013

In this paper, we describe a vision sensor-based driving algorithm for indoor automatic guided vehicles (AGVs) that facilitates a path tracking task using two mono cameras for navigation. One camera is mounted on vehicle to observe the environment and to detect markers in front of the vehicle. The other camera is attached so the view is perpendicular to the floor, which compensates for the distance between the wheels and markers. The angle and distance from the center of the two wheels to the center of marker are also obtained using these two cameras. We propose five movement patterns for AGVs to guarantee smooth performance during path tracking: starting, moving straight, pre-turning, left/right turning, and stopping. This driving algorithm based on two vision sensors gives greater flexibility to AGVs, including easy layout change, autonomy, and even economy. The algorithm was validated in an experiment using a two-wheeled mobile robot.

• Journal of Biomedical Engineering Research
• /
• v.18 no.3
• /
• pp.315-323
• /
• 1997

Driving is a complex vigilance task that includes improper lookout, excessive speed and inattention. The primary objective of this research is to detect driver drowsiness so that the driver can be alerted to an impending traffic accident in performance. We developed the automatic detection and warning system during drowsy driving. A drowsiness detection system must be able to monitor driver status and detect the detrimental changes of a driver performance. Eyeblink has been found to be a reliable factor of drowsiness detection in earlier studies. As an additional parameter, we also considered the yawning which often occurs in a low vigilance state and predicts the drowsy state. We used a computer vision method to extract the eyeblink and yawning in the face image sequences. When the drowsy state was detected, the driver was refreshed by alarming device and menthol scent generator after deciding the warning level by fuzzy logic. For the evaluation of our system, we measured the physiological parameters such as EOG and EEG. The results indicated that it is possible to detect and alert the driver drowsiness temporarily or continuously by using our system.