Recently, SDAS(Advanced driver-assistance system) are being installed to assist driving of vehicles and improve driver convenience. LDWS(Lane departure warning system) and FCWS(Forward collision warning system) are the core of the technology. Among these, FCWS is evaluated as a key assistive technology to prevent vehicle crashes. Accordingly, many algorithms are being developed and tested to improve detection speed and actual detection algorithms are being commercialized. In this paper, We propose the design of a system that optimizes FCWS speed by considering the AI performance of the terminal device when implemented as an embedded system.

PURPOSES : This study develops a model that can estimate travel speed of each movement flow using deep-learning-based probe vehicles at urban intersections. METHODS : Current technologies cannot determine average travel speeds for all vehicles passing through a specific real-world area under obseravation. A virtual simulation environment was established to collect information on all vehicles. A model estimate turning speeds was developed by deep learning using probe vehicles sampled during information processing time. The speed estimation model was divided into straight and left-turn models, developed as fully-offset, non-offset, and integrated models. RESULTS : For fully-offset models, speed estimation for both straight and left-turn models achieved MAPE within 10%. For non-offset models, straight models using data drawn from four or more probe vehicles achieved a MAPE of less than 15%. The MAPE for left turns was approximately 20%. CONCLUSIONS : Using probe-vehicle data(PVD), a deep learning model was developed to estimate speeds each movement flow. This, confirmed the viability of real-time signal control information processing using a small number of probe vehicles.

선박이 접안할 때 발생하는 접안에너지에 가장 영향력이 큰 요소는 접안속도이며, 과도한 경우 사고로 이어질 수 있다. 접안속도의 결정에 영향을 미치는 요소는 다양하지만 기존 연구에서는 일반적으로 선박 크기에 제한하여 분석하였다. 따라서 본 연구에서는 다양한 선박 접안속도의 영향요소를 반영하여 분석하고 그에 따른 중요도를 도출하고자 한다. 분석에 활용한 데이터는 국내 한 탱커부두의 선박 접안속도를 실측한 것을 바탕으로 하였다. 수집된 데이터를 활용하여 머신러닝 분류 알고리즘인 의사결정나무(Decision Tree), 랜덤포 레스트(Random Forest), 로지스틱회귀(Logistic Regression), 퍼셉트론(Perceptron)을 비교분석하였다. 알고리즘 평가 방법으로는 혼동 행렬에 따른 모델성능 평가지표를 사용하였다. 분석 결과, 가장 성능이 좋은 알고리즘으로는 퍼셉트론이 채택되었으며 그에 따른 접안속도 영향 요인의 중요도는 선박 크기(DWT), 부두 위치(Jetty No.), 재화상태(State) 순으로 나타났다. 이에 따라 선박 접안 시, 선박의 크기를 비롯하여 부두 위치, 재화 상태 등 다양한 요인을 고려하여 접안속도를 설계하여야 한다.

This research sought to analyze the characteristics of breast movement at the treadmill activity levels. It also examined the effect of wearing a sports bra in reducing breast displacement. The subjects for the data collection were females in their 20s (n=2) with C-cup size breast. The experimental conditions were three different moving speeds (4 km/h, jogging: 7 km/h, and sprinting: 10 km/h) and two types of sports bras. Three dimensional breast displacement was measured. The displacement of the right nipple point was measured with a 3D motion analyzer. The results show that the breasts were greatly displaced from the walking speed (4 km/h) when subjects did not wear any bra. Whereas their breast displacement distance decreased remarkably when they wore sports bras. The nipple point moved 42~44 mm in the vertical direction at walking speed with naked condition. But it was reduced by 80% after wearing sports bras. When subjects running (7 km/h, 10 km/h) without any bra, the nipple point moved 122~141 mm. However it was reduced by 60~70% when they wore sports bras. The apartment time (time delay) between at the highest point of the upper body and the nipple was 0.25 seconds at the running speeds (7 km/h, 10 km/h) without wearing any bra. After wearing sports bras, the time delay was cut to 0.06~0.12 seconds. These results implies that without wearing any bra the skin surrounding the breasts might be seriously pulled at running activity. The functional sports bra suppress breast movement. It might prevent the sagging of breasts by preventing the damage of the Cooper's ligaments.

보드게임에서는 많은 경우의 수의 말들과 많은 상태공간들을 가지고 있다. 그러므로 게임은 학습을 오래 하여야 한다. 본 논문에서는 Q러닝 알고리즘을 이용했다. 그러나 강화학습은 학습초기에 학습속도가 느려지는 단점이 있다. 그러므로 학습을 하는 동안에 같은 최선의 값이 있을 때, 게임트리를 고려한 문제영역의 지식을 활용한 휴리스틱을 사용하여 학습의 속도향상을 시도하였다. 기존 구현된 말과 개선하여 구현된 말을 비교하기 위하여 보드게임을 제작했다. 그래서 일방적으로 공격하는 말과 승부를 겨루게 하였다. 개선된 말은 게임트리를 고려하여 상대방 말을 공격하였다. 실험결과 개선하여 구현된 말이 학습속도적인 면에서 향상됨 것을 알수 있었다