http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Yonghwan Jeong 대한기계학회 2022 JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY Vol.36 No.2
This paper presents a personalized lane keeping system for an autonomous vehicle using recurrent neural network (RNN) with long short term memory (LSTM) cell. The proposed algorithm is trained by datasets collected by manual driving of three drivers. The collected driving data is analyzed for the target lane offset and responsiveness to the road curvature of each driver. 178744 and 76605 datasets are used to train and validate the LSTMRNN based model. An encoder is used to standardize the input feature to improve the accuracy of network training. The proposed lane keeping algorithm for each driver has been evaluated through prediction accuracy analysis and simulation study using MATLAB/Simulink and Carsim. 99.7 % of the prediction error of steering wheel angle was bounded between -0.87 deg to 0.89 deg with mean of 0.01 deg. The simulation results show that the proposed algorithm precisely modeled the lane keeping characteristics of three drivers.
Yonghwan Jeong 제어로봇시스템학회 2021 제어로봇시스템학회 국제학술대회 논문집 Vol.2021 No.10
This paper presents a path tracking control algorithm for four-wheel steering autonomous vehicles. The proposed controller utilized the direct yaw moment control approach for path tracking of the autonomous vehicle. The path tracking algorithm consists of three parts, desired yaw-rate decision, adaptive sliding mode control, and control allocation. The desired yaw rate has been determined based on the geometric relationship between a reference path and vehicle position. A required yaw-moment to tracking the desired yaw-rate has been decided by using an adaptive sliding mode control approach to compensate for the linear tire model assumption. Since the four-wheel steering vehicle is an overactuated system, the optimization-based control allocation has been introduced to determine the front and rear-wheel steering inputs, respectively, considering a control effort, actuator limit, ride comfort, and body slip. The simulation study has been conducted to compare the path tracking performance of the base controller using front steering and the proposed algorithm. It has been shown from the simulations that the path tracking performance has been improved in a driving situation with a high curvature and a steep road slope.
정용환(Yonghwan Jeong),이경준(Kyoungjun Lee),정혁진(Hyok-Jin Chong),고봉철(Bong-Chul Ko),이경수(Kyongsu Yi) 대한기계학회 2015 대한기계학회 춘추학술대회 Vol.2015 No.11
This paper presents a vehicle sensor fault tolerant algorithm for automated vehicles. The proposed algorithm consists of a fault detection algorithm and virtual sensor. The fault detection algorithm is designed to monitor the health of steering wheel angle, yaw-rate, and wheel speed sensors. Three yaw-rate estimators using different sensor measurements are used to construct a bank of residuals. A fault of vehicle sensor leads to increase the unique subset of residuals and an adaptive threshold is used to identify the increase of residuals. The virtual sensor is composed of steering wheel angle, yaw-rate, and wheel speed estimators which uses measurements from normal operating sensors. After a faulty sensor is identified, measurement from faulty sensor is replaced by measurement from virtual sensor. The fault tolerant performance and reliability of the proposed algorithm have been validated via computer simulation.