• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.21 no.5
• /
• pp.274-286
• /
• 2022

Multi-object tracking has been studied for a long time under computer vision and plays a critical role in applications such as autonomous driving and driving assistance. Multi-object tracking techniques generally consist of a detector that detects objects and a tracker that tracks the detected objects. Various publicly available datasets allow us to train a detector model without much effort. However, there are relatively few publicly available datasets for training a tracker model, and configuring own tracker datasets takes a long time compared to configuring detector datasets. Hence, the detector is often developed separately with a tracker module. However, the separated tracker should be adjusted whenever the former detector model is changed. This study proposes a system that can train a model that performs detection and tracking simultaneously using only the detector training datasets. In particular, a Siam network with augmentation is used to compose the detector and tracker. Experiments are conducted on public datasets to verify that the proposed algorithm can formulate a real-time multi-object tracker comparable to the state-of-the-art tracker models.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.1
• /
• pp.77-84
• /
• 2023

In autonomous robots, obstacle avoidance is a key feature. Potential Field is the most widely used method in this field. Such method requires real-time calculation of the nearest point of the obstacle from the robot, which involves difficulty of reliably segmenting the obstacle region from the distance sensor data profile. In this paper, Active Min-Depth Filter is introduced to obtain the nearest point of each obstacle using real-time calculation but without segmentation. Through simulations on various sensor noise environments, the robustness of the Active Min-Depth Filter could be confirmed, and successful results were obtained by applying real-world moving robots.

• The Journal of Korea Robotics Society
• /
• v.12 no.3
• /
• pp.297-305
• /
• 2017

As the development of autonomous vehicles becomes realistic, many automobile manufacturers and components producers aim to develop 'completely autonomous driving'. ADAS (Advanced Driver Assistance Systems) which has been applied in automobile recently, supports the driver in controlling lane maintenance, speed and direction in a single lane based on limited road environment. Although technologies of obstacles avoidance on the obstacle environment have been developed, they concentrates on simple obstacle avoidances, not considering the control of the actual vehicle in the real situation which makes drivers feel unsafe from the sudden change of the wheel and the speed of the vehicle. In order to develop the 'completely autonomous driving' automobile which perceives the surrounding environment by itself and operates, ability of the vehicle should be enhanced in a way human driver does. In this sense, this paper intends to establish a strategy with which autonomous vehicles behave human-friendly based on vehicle dynamics through the reinforcement learning that is based on Q-learning, a type of machine learning. The obstacle avoidance reinforcement learning proceeded in 5 simulations. The reward rule has been set in the experiment so that the car can learn by itself with recurring events, allowing the experiment to have the similar environment to the one when humans drive. Driving Simulator has been used to verify results of the reinforcement learning. The ultimate goal of this study is to enable autonomous vehicles avoid obstacles in a human-friendly way when obstacles appear in their sight, using controlling methods that have previously been learned in various conditions through the reinforcement learning.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.12
• /
• pp.67-74
• /
• 2019

In this paper, we design and implement an analytics platform based on bulk cargo DTG data for crackdown on overloaded trucks. DTG(digital tachograph) is a device that stores the driving record in real time; that is, it is a device that records the vehicle driving related data such as GPS, speed, RPM, braking, and moving distance of the vehicle in one second unit. The fast processing of DTG data is essential for finding vehicle driving patterns and analytics. In particular, a big data analytics platform is required for preprocessing and converting large amounts of DTG data. In this paper, we implement a big data analytics framework based on cargo DTG data using Spark, which is an open source-based big data framework for crackdown on overloaded trucks. As the result of implementation, our proposed platform converts real large cargo DTG data sets into GIS data, and these are visualized by a map. It also recommends crackdown points.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.10
• /
• pp.3955-3971
• /
• 2020

Due to the significant increase in the use of autonomous car technology, it is essential to integrate this technology with high-precision digital map data containing more precise and accurate roadway information, as compared to existing conventional map resources, to ensure the safety of self-driving operations. While existing map technologies may assist vehicles in identifying their locations via Global Positioning System, it is however difficult to update the environmental changes of roadways in these maps. Roadway vision algorithms can be useful for building autonomous vehicles that can avoid accidents and detect real-time location changes. We incorporate a hybrid architectural design that combines unsupervised classification of vision data with supervised joint fusion classification to achieve a better noise-resistant algorithm. We identify, via a deep learning approach, an intelligent hybrid fusion algorithm for fusing multimodal vision feature data for roadway classifications and characterize its improvement in accuracy over unsupervised identifications using image processing and supervised vision classifiers. We analyzed over 93,000 vision frame data collected from a test vehicle in real roadways. The performance indicators of the proposed hybrid fusion algorithm are successfully evaluated for the generation of roadway digital maps for autonomous vehicles, with a recall of 0.94, precision of 0.96, and accuracy of 0.92.

• Journal of Auto-vehicle Safety Association
• /
• v.14 no.2
• /
• pp.57-69
• /
• 2022

Recently, as interest in self-driving cars has increased worldwide, research and development on the Advanced Driver Assist System is actively underway. Among them, the purpose of Adaptive Cruise Control (ACC) is to minimize the driver's driving fatigue through the control of the vehicle's longitudinal speed and relative distance. In this study, for the research of the ACC test in the real environment, the real-road test was conducted based on domestic-road test scenario proposed in preceding study, considering ISO 15622 test method. In this case, the distance measurement method using the dual camera was verified by comparing and analyzing the result of using the dual camera and the result of using the measurement equipment. As a result of the comparison, two results could be derived. First, the relative distance after stabilizing the ACC was compared. As a result of the comparison, it was found that the minimum error rate was 0.251% in the first test of scenario 8 and the maximum error rate was 4.202% in the third test of scenario 9. Second, the result of the same time was compared. As a result of the comparison, it was found that the minimum error rate was 0.000% in the second test of scenario 10 and the maximum error rate was 9.945% in the second test of scenario 1. However, the average error rate for all scenarios was within 3%. It was determined that the representative cause of the maximum error occurred in the dual camera installed in the test vehicle. There were problems such as shaking caused by road surface vibration and air resistance during driving, changes in ambient brightness, and the process of focusing the video. Accordingly, it was determined that the result of calculating the distance to the preceding vehicle in the image where the problem occurred was incorrect. In the development stage of ADAS such as ACC, it is judged that only dual cameras can reduce the cost burden according to the above derivation of test results.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.11 no.1
• /
• pp.13-27
• /
• 2012

According to the development of Intelligent Transportation System technology, VMS(Variable Message Signs) are operating on highway. But, VMS display information which don't reflect traffic condition and driver's human factor. So driver can't read VMS message during limited time, it makes to reduce VMS's reliability. This paper presents a model for VMS message reading time and distance considering traffic condition and human factor. We built driving simulator by Winroad package which is able to copy real driving condition. Subjects were comprised of 20 people who reflect domestic driver's condition such as sex. We did regression analysis with experiment results and draw the model. The model could be possible to develop message- set considering traffic condition and human factor.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.07a
• /
• pp.7-8
• /
• 2019

In this paper, we applied the migration technique to real-time tasks that have relatively low criticality but still important to be dropped by the mixed-criticality scheduling algorithms. The proposed drop and migrate algorithm analyzes the schedulability by calculating CPU utilization and response time of using task migration. We provide analysis to guarantee the deadline of LO-tasks, by transforming the response time equation specified with migration time. The transformed response time equation was able to analyze the migration schedulability. This algorithm can be used with various mixed-criticality schedulers as a supplementary method. We expect this algorithm will be used for scheduling LO-tasks such as communication task that requires safety guarantee especially in platooning and autonomous driving by utilizing the advantages of multiple node connectivities.

• KIISE Transactions on Computing Practices
• /
• v.24 no.2
• /
• pp.59-66
• /
• 2018

In this paper, we propose a system to detect various damage automatically inflicted on road pavement by collecting and analyzing data from acceleration and camera sensors in real time. The proposed system sends the collected images, acceleration signals, and GPS coordinates to the road manager and the database in the remote server, shortly after detecting the damage to the road pavement. Our study makes three key contributions. The proposed system 1) enables road managers to maintain road conditions quickly, accurately, and conveniently; 2) allows road mangers to take care of various kinds of damage to the road pavement at the initial stage; and finally 3) even makes it possible to track the damage, which suggests that the integration of a high-level decision support function becomes affordable. We tested the sensitivity and precision of the proposed system against real-time data obtained from the vehicles driving on the highway at an average speed of 100 km/h. With ten iterations, the proposed system achieved an average sensitivity of 74% and an average precision of 84% in road pavement damage detection, which is comparable with the best competing schemes.

• Journal of IKEEE
• /
• v.21 no.4
• /
• pp.358-367
• /
• 2017

Advanced Driver Assistance Systems(ADAS), such as Front Collision Warning System (FCWS) are currently being developed. FCWS require high processing speed because it must operate in real time while driving. In addition, a low-power system is required to operate in an automobile embedded system. In this paper, FCWS is implemented in CPU-FPGA architecture in embedded system to enable real-time processing. The lane detection enabled the use of the Inverse Transform Perspective (IPM) and sliding window methods to operate at fast speed. To detect the vehicle, a Convolutional Neural Network (CNN) with high recognition rate and accelerated by parallel processing in FPGA is used. The proposed architecture was verified using Intel FPGA Cyclone V SoC(System on Chip) with ARM-Core A9 which operates in low power and on-board FPGA. The performance of FCWS in HD resolution is 44FPS, which is real time, and energy efficiency is about 3.33 times higher than that of high performance PC enviroment.