• 한국ITS학회 논문지
• /
• 제22권1호
• /
• pp.276-290
• /
• 2023

현재 국내에서는 자율주행차량의 상용화를 목표로 다양한 노력을 기울이고 있으며 자율주행차량이 운영 가이드라인에 따라 안전하고 신속하게 주행할 수 있는 연구들이 대두되고 있다. 본 연구는 자율주행차량의 경로탐색을 미시적인 관점으로 바라보며 Deep Q-Learning을 통해 자율주행차량의 차로변경을 학습시켜 효율성을 입증하고자 한다. 이를 위해 SUMO를 사용하였으며, 시나리오는 출발지에서 랜덤 차로로 출발하여 목적지의 3차로까지 차로변경을 통해 우회전하는 것으로 설정하였다. 연구 결과 시뮬레이션 기반의 차로변경과 Deep Q-Learning을 적용한 시뮬레이션 기반의 차로변경으로 구분하여 분석하였다. 평균 통행 속도는 Deep Q-Learning을 적용한 시뮬레이션의 경우가 적용하지 않은 경우에 비해 약 40% 향상되었으며 평균 대기 시간은 약 2초, 평균 대기 행렬 길이는 약 2.3대 감소하였다.

• 한국정보전자통신기술학회논문지
• /
• 제14권2호
• /
• pp.134-139
• /
• 2021

현재 자율주행차량 시장은 3레벨 자율주행차량을 상용화하고 있으나, 여전히 운전자의 주의를 필요로 한다. 3레벨 자율주행 이후 4레벨 자율주행차량에서 가장 주목되는 부분은 차량의 안정성이다. 3레벨과 다르게 4레벨 이후의 자율주행차량은 운전자의 부주의까지 포함하여 자율주행을 실시해야 하기 때문이다. 따라서 본 논문에서는 운전자가 부주의한 상황에서 긴급상황을 알리고 운전자의 반응을 인식하는 자율차량 안전을 위한 긴급상황 알림 및 운전자 반응 확인 시스템을 제안한다. 긴급상황 알림 및 운전자 반응 확인 시스템은 긴급상황 전달 모듈을 사용하여 긴급상황을 텍스트화하여 운전자에게 음성으로 전달하며 운전자 반응 확인 모듈을 사용하여 긴급상황에 대한 운전자의 반응을 인식하고 운전 권한을 운전자에게 넘길지 결정한다. 실험 결과, 긴급상황 전달 모듈의 HMM은 RNN보다 25%, LSTM보다 42.86% 빠른 속도로 음성을 학습했다. 운전자 반응 확인 모듈의 Tacotron2는 deep voice보다 약 20ms, deep mind 보다 약 50ms 더 빨리 텍스트를 음성으로 변환했다. 따라서 긴급상황 알림 및 운전자 반응 확인 시스템은 효율적으로 신경망 모델을 학습시키고, 실시간으로 운전자의 반응을 확인할 수 있다.

• 한국ITS학회 논문지
• /
• 제22권1호
• /
• pp.311-326
• /
• 2023

본 연구는 자율주행차량과 일반차량의 인지 방식과 범위의 차이를 비교하고 미시적 교통시뮬레이션을 활용한 교통안전시설의 설치 및 운영 개선방안을 제안하는 것을 목적으로 하였다. 연구에서는 기존 교통안전표지 설치·관리 업무편람을 검토하고 교통안전시설 중 안전표지를 차량의 행동 변화에 따라 분류하였다. 이후, 분류된 시설에 대해 가상환경에서 시뮬레이션 실험을 통해 안전표지의 설치 위치를 변경해가며 최적 설치지점을 추론하여 개선방안을 제시하였다. 본 연구는 일반 운전자의 시인성을 기준으로 설치된 교통안전시설이 자율주행차량과 일반차량이 혼재된 상황에서 도로 효율성과 안전성에 어떠한 영향을 주는지 확인하였다. 본 연구를 통해 도출한 교통안전시설의 인지에 대한 임곗값은 교통안전시설의 설치·관리에 관한 규칙의 개정 근거로 활용될 수 있다는 데 의의가 있다.

• 농업과학연구
• /
• 제48권3호
• /
• 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.

• 한국컴퓨터정보학회논문지
• /
• 제23권11호
• /
• pp.167-173
• /
• 2018

The Korean government announced its goal of commercializing autonomous vehicle by year 2020. With such changes, it is expecting to decrease car accident mortality by half. To commercialize autonomous car, not only worries on safety of autonomous vehicle has to be solved but at the same time, institutional system has to be clear to distinguish legal responsibilities in case of accident. This paper will present the legal improvement direction of the introduction of autonomous vehicles as follows. First, it is necessary to re-establish concept of 'driver' institutionally. Second, it is appropriate to focus on Level 3 autonomous vehicle which is about to be commercialized in year 2020 and organize legal responsibility. Third, we should have a clear understanding on how level 3 autonomous vehicle will be commercialized in the future. Fourth, it is necessary to revise The Traffic Law, Act on Special Cases concerning the Settlement of Traffic Accident, and Automobile Accident Compensation Security Law in line with level 3 autonomous vehicle. Fifth, it is necessary to review present car insurance system. Sixth, present Product Liability Law is limited to movable products (Article 2), however, it is necessary to include intangible product which is software. Seventh, we should review on making special law related to autonomous car including civil, criminal, administrative, and insurance perspectives.

• 전자통신동향분석
• /
• 제35권6호
• /
• pp.1-11
• /
• 2020

Recently, autonomous things, which are pieces of equipment or devices that grasp the context of circumstances on their own and perform actions appropriate for the situation in the surrounding environment, are attracting much research interest. This is because autonomous things are expected to be able to interact with humans more naturally, supersede humans in many tasks, and further solve problems by themselves by collaborating with each other without human intervention. This prospect leans heavily on AI as deep learning has delivered astonishing breakthroughs recently and broadened its range of applications. This paper surveys application trends in deep learning-based AI techniques for autonomous things, especially autonomous driving vehicles, because they present a wide range of problems involving perception, decision, and actions that are very common in other autonomous things.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2004년도 ICCAS
• /
• pp.1126-1130
• /
• 2004

This paper describes the design, development, and performance testing of an autonomous ground vehicle that was developed to participate in the DARPA Grand Challenge that was held in March 2004. The authors of this paper are members of Team CIMAR which was one of twenty five teams selected by DARPA to participate in a competition to develop an autonomous vehicle that can navigate from near Los Angeles to near Las Vegas at speeds averaging twenty miles per hour. Most of the event was held on open terrain and trails in a rocky desert environment. This paper describes the overall system design and the performance of the system at the event.

• 한국정밀공학회지
• /
• 제17권4호
• /
• pp.121-128
• /
• 2000

In this paper, the control algorithm fur an autonomous vehicle is studied and applied to an actual 2 wheel-driven vehicle system. In order to control a nonholonomic system, the kinematic model for an autonomous vehicle is constructed by relative velocity relationship about the virtual point at distance from the vehicle's frame. And the optimal controller that based on the kinematic model is operated on purpose to track a reference vehicle's path. The actual system is designed with named 'HYAVI' and the system controller is applied. Because all the results of simulation don't satisfy the driving conditions of HYAVI, a reformed control algorithm that satisfies an actual autonomous vehicle is applied at HYAVI. At the results of actual experiments, the path tracking works very well by the reformed control algorithm. An autonomous vehicle that applied this control algorithm can be easily used for a path generation algorithm.

• 제어로봇시스템학회논문지
• /
• 제18권3호
• /
• pp.251-257
• /
• 2012

Unmanned ground vehicles have important military, reconnaissance, and materials handling application. Many of these applications require the UGVs to move at high speeds through uneven, natural terrain with various compositions and physical parameters. This paper presents a framework for high speed autonomous navigation based on the integrated real time traversability. Specifically, the proposed system performs real-time dynamic simulation and calculate maximum traversing velocity guaranteeing safe motion over rough terrain. The architecture of autonomous navigation is firstly presented for high-speed autonomous navigation. Then, the integrated real time traversability, which is composed of initial velocity profiling step, dynamic analysis step, road classification step and stable velocity profiling step, is introduced. Experimental results are presented that demonstrate the method for a $6{\times}6$ autonomous vehicle moving on flat terrain with bump.

• 대한임베디드공학회논문지
• /
• 제16권6호
• /
• pp.285-292
• /
• 2021

Autonomous vehicles are divided into an upper controller that calculates control value through cognitive judgment and a lower controller that appropriately transmits its control value to an actuator. Here, the longitudinal control in a lower controller has a problem as the road slopes due to the property of the Acceleration sensor to output the acceleration as the slope of the device. Therefore, in this paper, a sigmoid function is proposed to determine the slope to compensate for this problem. Through the experiment, Checked performance by comparing the existing table model with the proposed model.