• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.6
• /
• pp.313-330
• /
• 2021

In an autonomous driving system equipped with a camera and lidar, depth completion techniques enable dense depth estimation. In particular, using self-supervised learning it is possible to train the depth completion network even without ground truth. In actual autonomous driving, such depth completion should have very short latency as it is the input of other algorithms. So, rather than complicate the network structure to increase the accuracy like previous studies, this paper focuses on network latency. We design a U-Net type network with RegNet encoders optimized for GPU computation. Instead, this paper presents several techniques that can increase accuracy during the process of self-supervised learning. The proposed techniques increase the robustness to unreliable lidar inputs. Also, they improve the depth quality for edge and sky regions based on the semantic information extracted in advance. Our experiments confirm that our model is very lightweight (2.42 ms at 1280x480) but resistant to noise and has qualities close to the latest studies.

• Korean Journal of Agricultural Science
• /
• v.48 no.3
• /
• pp.617-627
• /
• 2021

In the field of on-road motor vehicles, the level for autonomous driving technology is defined according to J3016, proposed by Society of Automotive Engineers (SAE) International. However, in the field of agricultural machinery, different standards are applied by country and manufacturer, without a standardized classification for autonomous driving technology which makes it difficult to clearly define and accurately evaluate the autonomous driving technology, for agricultural machinery. In this study, a method to classify the autonomy levels for autonomous agricultural machinery (ALAAM) is proposed by modifying the SAE International J3016 to better characterize various agricultural operations such as tillage, spraying and harvesting. The ALAAM was classified into 6 levels from 0 (manual) to 5 (full automation) depending on the status of operator and autonomous system interventions for each item related to the automation of agricultural tasks such as straight-curve path driving, path-implement operation, operation-environmental awareness, error response, and task area planning. The core of the ALAAM classification is based on the relative roles between the operator and autonomous system for the automation of agricultural machines. The proposed ALAAM is expected to promote the establishment of a standard to classify the autonomous driving levels of self-propelled agricultural machinery.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.5
• /
• pp.203-208
• /
• 2023

Recently, with the growing interest in the field of autonomous driving, many researchers have been focusing on developing autonomous driving software platforms. In particular, we have concentrated on developing 3D object detection models that can improve real-time performance. In this paper, we introduce a self-constructed 3D LiDAR dataset specific to domestic environments and propose a VariFocal-based CenterPoint for the 3D object detection model, with improved performance over the previous models. Furthermore, we present experimental results comparing the performance of the 3D object detection modules using our self-built and public dataset. As the results show, our model, which was trained on a large amount of self-constructed dataset, successfully solves the issue of failing to detect large vehicles and small objects such as motorcycles and pedestrians, which the previous models had difficulty detecting. Consequently, the proposed model shows a performance improvement of about 1.0 mAP over the previous model.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.5
• /
• pp.889-898
• /
• 2021

Currently, due to the demographic cliff phenomenon in Republic of Korea, A serious defense vacuum could occur due to the lack of South Korean military's personal strength. As a result, The South Korean military has a possibility to implement the polices the prepare for military provocations and preemptive strikes by the North Korean military while resolving the South Korean defense vacuum caused by the shrinking population. It seems like that the only way for the South Korean military to solve the shortage of personal strength due to the population decline is to reduce the number of Mechanized Units(MU) other than, infantry and automate, and autonomous driving the weapons system of the Mechanized Units(MU). In this paper, we propose the use of the virtual autonomous driving of the self propelled artillery K-9's in self selection of the position and occupation of position and self positioning in the position. At the same time in this paper, the self propelled artillery K-9 model robot is used to simulate and the explain about the operation method, necessity and feasibility in the self propelled artillery K-9. In addition, this paper predicted the problems that would arise if the South Korean military deployed autonomous driving self propelled K-9, in real combat.

• The Journal of the Convergence on Culture Technology
• /
• v.5 no.1
• /
• pp.361-369
• /
• 2019

With the commercialization of the ADAS functions, the need for the experience of the autonomous driving system is increasing, and the role of the artificial intelligence agent is attracting attention. This study is an autonomous driving experience experiment that verifies the effect of self-efficacy and agent embodiment. Through a simulator experiment, we measured the effect of existence of self-efficacy and agent embodiment on social presence, perceived risk, and perceived ease of use. Results show that self-efficacy had a positive effect on social presence and perceived risk, and agent embodiment negatively affected perceived ease of use. Based on the results of the study, we proposed guidelines for agent design that can increase the acceptance of the semi-autonomous driving system.

• Journal of Advanced Navigation Technology
• /
• v.27 no.2
• /
• pp.182-189
• /
• 2023

Depth estimation is a key technology in 3D map generation for autonomous driving of vehicles, robots, and drones. The existing sensor-based method has high accuracy but is expensive and has low resolution, while the camera-based method is more affordable with higher resolution. In this study, we propose self-attention-based unsupervised monocular depth estimation for UAV camera system. Self-Attention operation is applied to the network to improve the global feature extraction performance. In addition, we reduce the weight size of the self-attention operation for a low computational amount. The estimated depth and camera pose are transformed into point cloud. The point cloud is mapped into 3D map using the occupancy grid of Octree structure. The proposed network is evaluated using synthesized images and depth sequences from the Mid-Air dataset. Our network demonstrates a 7.69% reduction in error compared to prior studies.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.62 no.1
• /
• pp.83-89
• /
• 2013

In order to reduce the power consumption in international standard IEC62087, the luminance efficiency should be improved at the low discharge load rather than at the high discharge load. Thus, this paper analysed the characteristics of the discharge at the panels with ITO Gap of $65{\mu}m$, $80{\mu}m$, and $100{\mu}m$ in 50-inch PDP with FHD resolution. It was well known that the long gap panel improves the luminance and the luminous efficiency. However, it is very difficult to drive the panel due to high driving voltage. When the normal driving method was applied at the panel with ITO gap of $100{\mu}m$, the phenomenon of the double peak was generated in the sustain period. We confirmed that main factor of the double peak is the self-erasing discharge. When the CLHS driving method was applied at the panel with ITO gap of $100{\mu}m$, the self-erasing discharge was improved in the sustain period. Also, the $V_S$ and $V_A$ minimum voltage of the CLHS driving method decreased about 9V and 12V compared with those of the normal driving method. Moreover, when the CLHS driving method was applied to the panel with ITO gap of $100{\mu}m$, the luminance and the luminous efficiency increased compared with those of the normal driving method. The luminance and the luminous efficiency greatly increased at the low discharge load. The less discharge load, the higher increase rate of the luminance and the luminous efficiency. Especially, the luminous efficiency at ITO gap of $100{\mu}m$ increased about 26.3% at the discharge load of 4% compared with that at ITO gap of $65{\mu}m$.

• Journal of Auto-vehicle Safety Association
• /
• v.15 no.2
• /
• pp.42-48
• /
• 2023

In recent years, automated vehicles have garnered attention in the multidisciplinary research field, promising increased safety on the road and new opportunities for passengers. High-Definition (HD) maps have been in development for many years as they offer roadmaps with inch-perfect accuracy and high environmental fidelity, containing precise information about pedestrian crossings, traffic lights/signs, barriers, and more. Demonstrating autonomous driving requires verification of driving on actual roads, but this can be challenging, time-consuming, and costly. To overcome these obstacles, creating HD maps of real roads in a simulation and conducting virtual driving has become an alternative solution. However, existing HD maps using high-precision data are expensive and time-consuming to build, which limits their verification in various environments and on different roads. Thus, it is challenging to demonstrate autonomous driving on anything other than extremely limited roads and environments. In this paper, we propose a new and simple method for implementing HD maps that are more accessible for autonomous driving demonstrations. Our HD map combines the CARLA simulator and OpenStreetMap (OSM) data, which are both open-source, allowing for the creation of HD maps containing high-accuracy road information globally with minimal dependence. Our results show that our easily accessible HD map has an accuracy of 98.28% for longitudinal length on straight roads and 98.42% on curved roads. Moreover, the accuracy for the lateral direction for the road width represented 100% compared to the manual method reflected with the exact road data. The proposed method can contribute to the advancement of autonomous driving and enable its demonstration in diverse environments and on various roads.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.9
• /
• pp.2129-2137
• /
• 2015

Lots of research and development works have been actively focused on the self-driving vehicles, locally and globally. In order to implement the self-driving vehicles, lots of fundamental core technologies need to be successfully developed and, specially, it is noted that traffic lights detection and recognition system is an essential part of the computer vision technologies in the self-driving vehicles. Up to nowadays, most conventional algorithm for detecting and recognizing traffic lights are mainly based on the color signal analysis, but these approaches have limits on the performance improvements that can be achieved due to the color signal noises and environmental situations. In order to overcome the performance limits, this paper introduces the morphological analysis for the traffic lights recognition. That is, by considering the color component analysis and the shape analysis such as rectangles and circles simultaneously, the efficiency of the traffic lights recognitions can be greatly increased. Through several simulations, it is shown that the proposed method can highly improve the recognition rate as well as the mis-recognition rate.

• The Journal of the Convergence on Culture Technology
• /
• v.8 no.1
• /
• pp.551-563
• /
• 2022

This study aims to identify the effect on the user experiences when the embodied agent in a self-driving car interacts with emotional expressions by using 'interjection'. An experimental study was designed with two conditions: the inclusion of injections in the agent's conversation feedbacks (with interjections vs. without interjections) and the type of conversation (task-oriented conversation vs. social-oriented conversation). The online experiment was conducted with the four video clips of conversation scenario treatments and measured intimacy, likability, trust, social presence, perceived anthropomorphism, and future intention to use. The result showed that when the agent used interjection, the main effect on social presence was found in both conversation types. When the agent did not use interjection in the task-oriented conversation, trust and future intention to use were higher than when the agent talked with emotional expressions. In the context of the conversation with the AI agent in a self-driving car, we found only the effect of adding emotional expression by using interjection on the enhancing social presence, but no effect on the other user experience factors.