http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
조향각 개선을 위한 단일모터 및 솔레노이드 기반 선형 제어 4륜 조향 구조 설계
변수환,박재병 제어로봇시스템학회 2021 제어로봇시스템학회 각 지부별 자료집 Vol.2021 No.12
This paper proposes improved mobile robot steering structure capable of in-place rotation and counter-steering with a single motor and a solenoid actuator for linear control. In previous research, for steering, our proposed structure changes linear movement on prismatic joint to rotation movement. Therefore, to calculate steering angle, we have to use non-linear function. Also, by this structural limitation, steering range of previous structure is constrained. So, we propose the improved structure that replaced rack gears to regular gears. It makes steering angle of the robot up to ±180 deg, so it can move laterally and we can calculate steering angle linearly. Furthermore, it has simpler mechanical structure than previous one. So we expect that the robot can be manufactured with cheaper price and higher maneuverability.
단일 모터 및 선형 구동기 기반의 기동성이 향상된 4 륜 선형 조향 구조 설계
변수환(SooHwan Byeon),박재병(Jaebyung Park) 제어로봇시스템학회 2022 제어·로봇·시스템학회 논문지 Vol.28 No.4
In this paper, we propose a new steering structure to improve the structural robustness and maneuverability of the original steering structure designed in our previous study. Most mobile robot platforms are designed with a differential drive or all-wheel steering (AWS) models. Differential-drive robots have a simple structure and are easy to manufacture, but this structure causes considerable slipping when the robots rotate. The AWS model has such a high maneuverability that it can reduce slip, but this structure requires as many steering motors as driving motors. To address these problems, in our previous research, we designed the steering structure to be as highly maneuverable as AWS models using a single steering motor and a solenoid actuator. However, the original structure requires nonlinear control and is too complex to analyze for selecting the proper steering motor. Thus, in this work, we design a steering structure with a simpler mechanical structure that has structural robustness and is easy to analyze, control, and manufacture. In addition, the proposed structure can steer faster than the original one and even move a four-wheel robot horizontally.
변수환 ( Soo Hwan Byun ),소현준 ( Hyun Jun Soh ),유정훈 ( Jeong Hoon Yoo ) 정보저장시스템학회 2012 정보저장시스템학회논문집 Vol.8 No.2
In spite of many advantages, the practical application of the thin film solar cell is restricted due to its low efficiency compared with the bulk type solar cells. This study intends to adopt the surface plasmon effect using nano particles to solve the low efficiency problem in thin film solar cells. By inserting Ag nano-particles in the absorbing layer of a thin film solar cell, the poynting vector value of the absorbing layer is increased due to the strong energy field. Increasing the value may give thin film solar cells chance to absorb more energy from the incident beam so that the efficiency of the thin film solar cell can be improved. In this work, we have designed the optimal shape of Ag nano-particle in the absorbing laser of a basic type thin film solar cell using the finite element analysis commercial package COMSOL. Design parameters are set to the particle diameter and the distance between each Ag nano-particle and by changing those parameters using the full factorial design variable set-up, we can determine optimal design of Ag nano-particles for maximizing the poynting vector value in the absorbing layer.
특징이 적은 환경에서의 AMCL 성능 향상을 위한 리샘플링 기법
변수환(SooHwan Byeon),박재병(Jaebyung Park) 대한전기학회 2021 대한전기학회 워크샵 Vol.2021 No.4
In this paper, we describe how to improve the performance of AMCL(Adaptive Monte Carlo Localization) algorithm in an environment with fewer features. This algorithm is one of the popular algorithms for the robot localization problem. It updates candidate samples of the robot"s pose based on existing samples. However, it causes a problem that cumulates errors while the robot moves in an environment with fewer features, so it shows worse performance. Robots based on the AMCL algorithm run recovery behavior when they are stuck or lose their way, but they often fail to find their pose even after running the recovery behavior. The original algorithm proposed a method to solve this kind of problem, but it was not good enough to find the robot’s pose in an environment with few features. Thus, we propose a method to sample candidate particles with a higher probability than other samples proposed by the original AMCL. We sampled random particles with constant probability and variance based on the robot’s pose and sensors’ noise. Original AMCL couldn"t localize the robot"s pose exactly after it passed through around 30 meters corridor, but it could with the proposed algorithm.
특징이 적은 환경에서의 AMCL 성능 향상을 위한 리샘플링 기법
변수환(SooHwan Byeon),박재병(Jaebyung Park) 대한전기학회 2021 정보 및 제어 심포지엄 논문집 Vol.2021 No.4
In this paper, we describe how to improve the performance of AMCL(Adaptive Monte Carlo Localization) algorithm in an environment with fewer features. This algorithm is one of the popular algorithms for the robot localization problem. It updates candidate samples of the robot’s pose based on existing samples. However, it causes a problem that cumulates errors while the robot moves in an environment with fewer features, so it shows worse performance. Robots based on the AMCL algorithm run recovery behavior when they are stuck or lose their way, but they often fail to find their pose even after running the recovery behavior. The original algorithm proposed a method to solve this kind of problem, but it was not good enough to find the robot’s pose in an environment with few features. Thus, we propose a method to sample candidate particles with a higher probability than other samples proposed by the original AMCL. We sampled random particles with constant probability and variance based on the robot’s pose and sensors’ noise. Original AMCL couldn’t localize the robot’s pose exactly after it passed through around 30 meters corridor, but it could with the proposed algorithm.