• Proceedings of the KAIS Fall Conference
• /
• 2004.06a
• /
• pp.209-212
• /
• 2004

실난수 발생기는 소프트웨어로 구현된 의사난수 발생기와는 달리 비결정적인 출력열을 제공하고, 알고리즘의 분해나 노출에 따른 위협이 없으며 의사난수발생기의 시드 값 관리가 불필요하다는 장점 등을 가지고 있다. 그러나 전자회고 잡음을 이용하는 경우 실난수 발생기는 온도나 주변 전자회로 기판 잡음 등으로 인해 지속적으로 편이성을 갖지 않는 안정된 출력 난수열을 얻는 것은 어렵다. 본 논문에서는 실난수 발생기에서 추출된 출력 난수열이 지속적으로 안정화된 출력 수열을 얻기 위한 방안을 연구하였다. 제시한 방안은 출력 난수열 "0"나 "1" 비트의 편이성을 가질 경우 FIPS 140-1에서 제시하고 있는 랜덤성의 조건을 만족하도록 출력 난수열의 랜덤성을 개선하였다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.42 no.5
• /
• pp.1058-1068
• /
• 2017

In V2X communication, a random based resource selection scheme is needed with considerations of cases that support devices without sidelink reception capabilities and require reduction of UE's power consumption. In this paper, as improvement of D2D's resource section scheme that one TRP is repeated to data subframe pool within a PSCCH period, it is proposed that different TRPs is applied for enhanced random resource selection based on pseudo-random sequence having UE-specific seed value. By results of proposed scheme's performance by numerical analysis, it is confirmed that collision probability among resources allocated to each UE for data TB is reduced, and a number of UEs which can avoid resource collision as much as possible and have simultaneous resource allocation is increased.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.350-350
• /
• 2023

SCE-UA(Shuffled Complex Evolution-University Arizona)기법은 최적해 탐색 알고리즘으로 개념적 강우유출 모형(conceptual rainfall runoff model)의 보정을 위한 도구로 개발되었다. SCE-UA기법은 메타휴리스틱 방법의 일종으로 최적해를 구하기 위하여 여러번 목적함수 값을 계산해야 한다. 이 때 목적함수 계산 횟수와 해의 수렴과 관련된 제어 매개변수가 존재하며, 사용자가 적절한 값을 입력해주어야 한다. 이 연구에서는 SCE-UA와 관련된 제어 매개변수의 기능에 대해서 검토하였다. 그리고 집합체 수의 변화에 따라서 검사함수인 Ackley function의 전역해를 얼마나 잘 탐색하는지 검토하였다. 검토 결과 랜덤 시드에 따라서 전역해 탐색 결과가 달라졌으며, 집합체의 수가 증가할수록 목적함수 계산 횟수는 증가하는 경향을 나타내었다. 검사함수의 차원(결정 변수의 수)이 증가하면 전역해의 탐색률이 감소하며, 집합체의 수가 많아지면 전역해를 더 잘 찾는 경향이 나타나지만, 목적함수 계산 횟수는 더 많아지게 되는 것을 확인할 수 있었다. 2차원인 경우 집합체의 수가 7개 이상일 때 탐색 성공률은 90% 이상이 되었지만, 10차원인 경우 집합체의 수가 시험 최대값인 20개일 때의 전역해 탐색률은 37%에 그쳤다. 이 연구의 결과는 SCE-UA 기법의 설정 매개변수에 관한 기본 개념을 이해하고, 사용자가 설정 매개변수 선정 시에 활용할 수 있을 것이다.

• Journal of Internet Computing and Services
• /
• v.23 no.3
• /
• pp.47-56
• /
• 2022

The safe driving reward system aims to reduce the loss of life and property by reducing the occurrence of accidents by motivating safe driving and encouraging active participation by providing direct reward to vehicle drivers who have performed safe driving. In the case of the existing digital tachograph, the goal is to limit dangerous driving by recording the driving status of the vehicle whereas the safe driving reward system is a support measure to increase the effect of accident prevention and induces safe driving with financial reward when safe driving is performed. In other words, in an area where accidents due to speeding are high, direct reward is provided to motivate safe driving to prevent traffic accidents when safe driving instructions such as speed compliance, maintaining distance between vehicles, and driving in designated lanes are performed. Since these safe operation data and reward histories must be managed transparently and safely, the reward evidences and histories were constructed using the closed blockchain Hyperledger Fabric. However, while transparency and safety are guaranteed in the blockchain system, low data processing speed is a problem. In this study, the sequential block generation speed was as low as 10 TPS(transaction per second), and as a result of applying the acceleration function a high-performance network of 1,000 TPS or more was implemented.

• Journal of Digital Convergence
• /
• v.18 no.10
• /
• pp.233-238
• /
• 2020

In the cloud environment, IoT devices using sensors and wearable devices are being applied in various environments, and technologies that accurately determine the information generated by IoT devices are being actively studied. However, due to limitations in the IoT environment such as power and security, information generated by IoT devices is very weak, so financial damage and human casualties are increasing. To accurately collect and analyze IoT information, this paper proposes a topographic information-based key management technique that considers IoT information errors. The proposed technique allows IoT layout errors and groups topographic information into groups of dogs in order to secure connectivity of IoT devices in the event of arbitrary deployment of IoT devices in the cloud environment. In particular, each grouped terrain information is assigned random selected keys from the entire key pool, and the key of the terrain information contained in the IoT information and the probability-high key values are secured with the connectivity of the IoT device. In particular, the proposed technique can reduce information errors about IoT devices because the key of IoT terrain information is extracted by seed using probabilistic deep learning.