2020 |
안찬우, 박재흥 수직 주차에서 AVM 을 이용한 동시적 위치추정 및 지도 생성 Conference/Workshop The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea, 06.11.2020. @misc{안찬우2020, title = {수직 주차에서 AVM 을 이용한 동시적 위치추정 및 지도 생성}, author = {안찬우, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/12/20201231_034353.pdf}, year = {2020}, date = {2020-11-06}, booktitle = {수직주차에서 을이용한동시적위치추정및지도생성}, journal = {자동차공학회}, address = {The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea}, abstract = {: In this paper, we propose a Simultaeneous Localization And Mapping(SLAM) method for vertical parking scenario. Most commercial car uses ultrasonic sensor to localize vehicle's position in parking lot. However, vehicle cannot be localized when there are not any obstacles around ego vehicle. The proposed method uses simple features from parking lines and localizes vehicle's position without any assumption about obstacles around ego vehicle using only Around View Monitor(AVM). At first, cross point is proposed as a feature which can be used for tracking. Cross points on parking lines are detected using Deep Learning and parking lines are extracted using RANSAC and Canny Edge. Using detected cross points and parking lines, relative poses between frames are tracked using Iterative Closeset Point(ICP) and landmark based localization. Global position of detected cross point is used as landmark's position and position of intersection from detected lines is used as local position of the corresponding cross point. Using landmark based localization, relative poses are calculated and accumulated errors are recovered. From above method, we build a 2D map which is continuously optimized. We evaluated our proposed approach on our own dataset and compared our result with LOAM(Lidar Odometry and Mapping in Real-time).}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } : In this paper, we propose a Simultaeneous Localization And Mapping(SLAM) method for vertical parking scenario. Most commercial car uses ultrasonic sensor to localize vehicle's position in parking lot. However, vehicle cannot be localized when there are not any obstacles around ego vehicle. The proposed method uses simple features from parking lines and localizes vehicle's position without any assumption about obstacles around ego vehicle using only Around View Monitor(AVM). At first, cross point is proposed as a feature which can be used for tracking. Cross points on parking lines are detected using Deep Learning and parking lines are extracted using RANSAC and Canny Edge. Using detected cross points and parking lines, relative poses between frames are tracked using Iterative Closeset Point(ICP) and landmark based localization. Global position of detected cross point is used as landmark's position and position of intersection from detected lines is used as local position of the corresponding cross point. Using landmark based localization, relative poses are calculated and accumulated errors are recovered. From above method, we build a 2D map which is continuously optimized. We evaluated our proposed approach on our own dataset and compared our result with LOAM(Lidar Odometry and Mapping in Real-time). |
안준우, 김민수, 박재흥 클로소이드(Clothoid)를 사용한 주차 경로 후-처리를 통한 경로 추종 오차 및 승차감 개선 Conference/Workshop The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea, 06.11.2020. @misc{안준우2020, title = {클로소이드(Clothoid)를 사용한 주차 경로 후-처리를 통한 경로 추종 오차 및 승차감 개선}, author = {안준우, 김민수, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/12/20201231_033430.pdf}, year = {2020}, date = {2020-11-06}, journal = {자동차공학회}, address = {The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea}, abstract = {최근 자율 주차 기술이 활발히 연구되고 있다. 다양하고 복잡한 상황에서 주차 경로를 생성하기 위 해 Hybrid-A* 혹은 RRT* (Rapidly-exploring random tree*) 알고리즘들을 이용하여 경로를 생성할 수 있다. 이 알고리즘들은 주차 시작 지점에서 주차할 지점에 도달할 때까지 장애물이 없는 지점에 트리 를 확장해 나가면서 경로를 생성한다. 트리를 확장할 때, 트리의 노드(node)와 노드 사이 경로는 차량 의 non-holonomic 제약 조건을 만족해야 하며 최단길이를 가진다. 이 때, 계산시간을 줄이기 위해서 단순화된 확장모델인 Reed-shepp curve를 사용한다. 이 curve는 최소회전 반경을 가지는 원과 직선의 조합으로 경로를 생성한다. 그러나, 원과 원 그리고 직선과 원이 만나는 지점의 곡률이 연속적이지 않 다. 게다가, 두 Reed-shepp curve가 만나는 지점에서의 곡률도 불연속하다. 생성된 경로의 곡률이 불연속한 경우, 두가지 문제점이 존재한다. 첫번째로, 불연속한 지점에서 추 종 오차가 발생한다. 이 이유는, 이 지점에서 곡률의 변화가 크며, 곡률을 따라가기 위해 제어되는 차 량의 조향각이 순간적으로 크게 변할 수 없기 때문이다. 예를 들어, 차량이 최소회전 반경을 가지는 원을 추종하고 있다가 직선을 추종하게 되었을 때, 원을 추종하는 조향각과 직선을 추종하기 위한 조 향각간 차이가 순간적으로 크게 발생하게 되고 조향각이 회전하는데 시간이 필요하기 때문에 직선 경 로를 벗어나게 된다. 이 때, 두번째 문제인 승차감이 좋지 못한 문제가 발생한다. 경로에 수렴하기 위 해 순간적으로 조향각이 크게 변하게 되고 제어 성능에 따라서 진동(oscillation)이 발생할 수 있다. 생성된 경로의 곡률을 연속적으로 만들기 위해, 본 연구에서는 클로소이드(clothoid)를 이용하여 경 로를 후처리한다. 클로소이드는 곡선 길이에 비례하여 곡률이 증대하는 성질을 가진 곡선으로, 차량이 등속주행시 등각속도로 핸들을 회전할 수 있다. 따라서, 조향각이 갑자기 크게 변하는 경우가 없어진 다. 생성된 경로에서 전/후진 전환지점을 기준으로 경로 세그먼트(segment)를 나눠 각 세그먼트 별로 클로소이드를 적용한다. 예를 들어, 주차 시작 지점에서 주차할 지점까지 경로를 생성하였을 때, 전/ 후진이 한번 발생되었다면 2개의 경로 세그먼트가 생성된다. 각 세그먼트의 시작 지점 자세(pose)에서 끝 지점의 자세까지 클로소이드를 적용한다.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } 최근 자율 주차 기술이 활발히 연구되고 있다. 다양하고 복잡한 상황에서 주차 경로를 생성하기 위 해 Hybrid-A* 혹은 RRT* (Rapidly-exploring random tree*) 알고리즘들을 이용하여 경로를 생성할 수 있다. 이 알고리즘들은 주차 시작 지점에서 주차할 지점에 도달할 때까지 장애물이 없는 지점에 트리 를 확장해 나가면서 경로를 생성한다. 트리를 확장할 때, 트리의 노드(node)와 노드 사이 경로는 차량 의 non-holonomic 제약 조건을 만족해야 하며 최단길이를 가진다. 이 때, 계산시간을 줄이기 위해서 단순화된 확장모델인 Reed-shepp curve를 사용한다. 이 curve는 최소회전 반경을 가지는 원과 직선의 조합으로 경로를 생성한다. 그러나, 원과 원 그리고 직선과 원이 만나는 지점의 곡률이 연속적이지 않 다. 게다가, 두 Reed-shepp curve가 만나는 지점에서의 곡률도 불연속하다. 생성된 경로의 곡률이 불연속한 경우, 두가지 문제점이 존재한다. 첫번째로, 불연속한 지점에서 추 종 오차가 발생한다. 이 이유는, 이 지점에서 곡률의 변화가 크며, 곡률을 따라가기 위해 제어되는 차 량의 조향각이 순간적으로 크게 변할 수 없기 때문이다. 예를 들어, 차량이 최소회전 반경을 가지는 원을 추종하고 있다가 직선을 추종하게 되었을 때, 원을 추종하는 조향각과 직선을 추종하기 위한 조 향각간 차이가 순간적으로 크게 발생하게 되고 조향각이 회전하는데 시간이 필요하기 때문에 직선 경 로를 벗어나게 된다. 이 때, 두번째 문제인 승차감이 좋지 못한 문제가 발생한다. 경로에 수렴하기 위 해 순간적으로 조향각이 크게 변하게 되고 제어 성능에 따라서 진동(oscillation)이 발생할 수 있다. 생성된 경로의 곡률을 연속적으로 만들기 위해, 본 연구에서는 클로소이드(clothoid)를 이용하여 경 로를 후처리한다. 클로소이드는 곡선 길이에 비례하여 곡률이 증대하는 성질을 가진 곡선으로, 차량이 등속주행시 등각속도로 핸들을 회전할 수 있다. 따라서, 조향각이 갑자기 크게 변하는 경우가 없어진 다. 생성된 경로에서 전/후진 전환지점을 기준으로 경로 세그먼트(segment)를 나눠 각 세그먼트 별로 클로소이드를 적용한다. 예를 들어, 주차 시작 지점에서 주차할 지점까지 경로를 생성하였을 때, 전/ 후진이 한번 발생되었다면 2개의 경로 세그먼트가 생성된다. 각 세그먼트의 시작 지점 자세(pose)에서 끝 지점의 자세까지 클로소이드를 적용한다. |
이양우, 김민수, 박재흥 AVM이미지와 ROTATE-YOLOv3를 이용한 주차 영역 인식 Conference/Workshop The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea, 06.11.2020. @misc{이양우2020, title = {AVM이미지와 ROTATE-YOLOv3를 이용한 주차 영역 인식}, author = {이양우, 김민수, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/12/20AKSAE_D083.pdf}, year = {2020}, date = {2020-11-06}, address = {The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea}, abstract = {최근 자율 발렛 주차 기술에 대한 연구가 활발히 진행되고 있다. 자율 발렛 주차 기술을 수 행하기 위한 중요한 기술 중 하나는 빈 주차 영역을 인식하는 기술이다. 기존의 방법들은 경계 선을 검출해 주차선을 인식한 다음 주차 영역을 인식하는 방법을 사용한다. 이와 같은 주차선 인식 방법은 두가지 단점이 있다. 첫번째는 Radon변환과 같은 방법으로 주차선을 찾는 과정에 서 계산 시간이 많이 소요돼 실시간으로 계산되기 어렵다. 두번째는 빛이나 그림자에 의해 주 차선이 선명하지 않다면 주차선 검출 기반 방법은 주차 영역을 인식하기 어렵다. 위 두가지 단점을 다루기 위해, 본 논문에서는 Around View Monitor(AVM) 이미지에서 딥러닝 을 통해 주차 영역을 인식하는 방법을 사용한다. 주차 영역은 하나의 경계 상자로 간주되며, 객체 상자를 인식하는 ROTATE-YOLOv3를 사용한다. ROTATE-YOLOv3는 기존의 YOLOV3를 변형한 객 체 인식 방법으로, 경계 상자의 회전 값을 추가해 학습할 수 있다는 장점이 있다. 여기서 회전 값은 학습 이미지의 하단선을 기준으로 경계 상자의 하단선의 틀어진 각도를 의미한다. 실제 차량으로 취득한 AVM 이미지를 통해 실험을 진행했다. }, keywords = {}, pubstate = {published}, tppubtype = {presentation} } 최근 자율 발렛 주차 기술에 대한 연구가 활발히 진행되고 있다. 자율 발렛 주차 기술을 수 행하기 위한 중요한 기술 중 하나는 빈 주차 영역을 인식하는 기술이다. 기존의 방법들은 경계 선을 검출해 주차선을 인식한 다음 주차 영역을 인식하는 방법을 사용한다. 이와 같은 주차선 인식 방법은 두가지 단점이 있다. 첫번째는 Radon변환과 같은 방법으로 주차선을 찾는 과정에 서 계산 시간이 많이 소요돼 실시간으로 계산되기 어렵다. 두번째는 빛이나 그림자에 의해 주 차선이 선명하지 않다면 주차선 검출 기반 방법은 주차 영역을 인식하기 어렵다. 위 두가지 단점을 다루기 위해, 본 논문에서는 Around View Monitor(AVM) 이미지에서 딥러닝 을 통해 주차 영역을 인식하는 방법을 사용한다. 주차 영역은 하나의 경계 상자로 간주되며, 객체 상자를 인식하는 ROTATE-YOLOv3를 사용한다. ROTATE-YOLOv3는 기존의 YOLOV3를 변형한 객 체 인식 방법으로, 경계 상자의 회전 값을 추가해 학습할 수 있다는 장점이 있다. 여기서 회전 값은 학습 이미지의 하단선을 기준으로 경계 상자의 하단선의 틀어진 각도를 의미한다. 실제 차량으로 취득한 AVM 이미지를 통해 실험을 진행했다. |
김민수, 안준우, 김민성, 박재흥 후보 궤적을 이용한 비구조화된 환경에서의 교차로 인식 및 주행 Conference/Workshop The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea, 06.11.2020. @misc{김민수2020, title = {후보 궤적을 이용한 비구조화된 환경에서의 교차로 인식 및 주행}, author = {김민수, 안준우, 김민성, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/12/20AKSAE_D035_2.pdf}, year = {2020}, date = {2020-11-06}, address = {The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea}, abstract = {In this paper, we present intersection detection and navigation method in the unstructured environment. The presented method uses only a camera image. It generates candidate trajectories in the segmented camera image and decides whether the vehicle enters an intersection using a histogram of the candidate trajectories. In addition, the vehicle selects the target trajectory and tracksit with considering the vehicle velocity, static obstacles, and navigation information. The proposed method was tested and evaluated in a simulation.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this paper, we present intersection detection and navigation method in the unstructured environment. The presented method uses only a camera image. It generates candidate trajectories in the segmented camera image and decides whether the vehicle enters an intersection using a histogram of the candidate trajectories. In addition, the vehicle selects the target trajectory and tracksit with considering the vehicle velocity, static obstacles, and navigation information. The proposed method was tested and evaluated in a simulation. |
Junhyung Kim, Jaesug Jung, Jaeheung Park Implementation of Position-based Walking to Torque-controlled Humanoid Conference/Workshop 2020 The 20th International Conference on Control, Automation and Systems, 13.10.2020. @misc{Kim2020b, title = {Implementation of Position-based Walking to Torque-controlled Humanoid}, author = {Junhyung Kim, Jaesug Jung, Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/11/20201123_060312.pdf}, year = {2020}, date = {2020-10-13}, journal = {2020 The 20th International Conference on Control, Automation and Systems}, volume = {0}, number = {0}, pages = {0}, publisher = {2020 The 20th International Conference on Control, Automation and Systems}, address = {2020 The 20th International Conference on Control, Automation and Systems}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
Beomyeong Park, Jaeheung Park Walking Pattern Generation using MPC with minimization of COM Velocity Fluctuation Conference/Workshop 2020 IEEE International Conference on Control, Automation and System, 13.10.2020. @misc{Park2020walking, title = {Walking Pattern Generation using MPC with minimization of COM Velocity Fluctuation}, author = {Beomyeong Park and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/11/WB5.4.pdf}, year = {2020}, date = {2020-10-13}, publisher = {2020 IEEE International Conference on Control, Automation and System}, address = {2020 IEEE International Conference on Control, Automation and System}, abstract = {The center of mass(COM) trajectory obtained through the predetermined Zero Momentum Point(ZMP) reference and Linear Inverted Pendulum Model(LIPM) has been effectively used for the humanoid robot walking. However, a reference in which ZMP is fixed at one point of the support foot during single support(SSP) can cause COM velocity fluctuation and discontinuity of COM acceleration. In order to solve these disadvantages, a moving ZMP trajectory has been used in which the ZMP moves forward from behind the support foot. However, the distance the ZMP moves within the support foot was determined experimentally or proportional to the foot size or stride. This makes it difficult to find a moving ZMP trajectory that can effectively reduce the COM velocity fluctuation. This paper proposes to generate a COM trajectory by directly considering the minimization of COM velocity fluctuation in the cost function of MPC. Through the constraint, the stability of the robot was maintained by keeping the ZMP in the support foot. The proposed method was verified through simulation and experiment.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } The center of mass(COM) trajectory obtained through the predetermined Zero Momentum Point(ZMP) reference and Linear Inverted Pendulum Model(LIPM) has been effectively used for the humanoid robot walking. However, a reference in which ZMP is fixed at one point of the support foot during single support(SSP) can cause COM velocity fluctuation and discontinuity of COM acceleration. In order to solve these disadvantages, a moving ZMP trajectory has been used in which the ZMP moves forward from behind the support foot. However, the distance the ZMP moves within the support foot was determined experimentally or proportional to the foot size or stride. This makes it difficult to find a moving ZMP trajectory that can effectively reduce the COM velocity fluctuation. This paper proposes to generate a COM trajectory by directly considering the minimization of COM velocity fluctuation in the cost function of MPC. Through the constraint, the stability of the robot was maintained by keeping the ZMP in the support foot. The proposed method was verified through simulation and experiment. |
박수한, 장근우, 김승연, 김상현, 박재흥 주어진 힘에 대해 관절 토크를 최소화하는 역기구학 Conference/Workshop The 15th Korea Robotics Society Annual Conference, Pyeongchang, Korea, 18.08.2020. @misc{ParkIKTorque2020, title = {주어진 힘에 대해 관절 토크를 최소화하는 역기구학}, author = {박수한, 장근우, 김승연, 김상현, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/02/20200206_042053.pdf}, year = {2020}, date = {2020-08-18}, address = {The 15th Korea Robotics Society Annual Conference, Pyeongchang, Korea}, abstract = {When handling an object, people usually optimize their pose to ease work load. It is possible because the required joint torque to realize target wrench can be changed by different joint configurations. Using this property, we propose an inverse kinematics method to lessen the joint torque while the robot handles an object. The proposed method minimizes the norm of joint torque ensuring position and orientation constraints by using non-linear optimization method. As a result, we could obtain a joint configuration which does not violate the constraints and achieves the smaller norm of joint torque than that from conventional inverse kinematics algorithm. }, keywords = {}, pubstate = {published}, tppubtype = {presentation} } When handling an object, people usually optimize their pose to ease work load. It is possible because the required joint torque to realize target wrench can be changed by different joint configurations. Using this property, we propose an inverse kinematics method to lessen the joint torque while the robot handles an object. The proposed method minimizes the norm of joint torque ensuring position and orientation constraints by using non-linear optimization method. As a result, we could obtain a joint configuration which does not violate the constraints and achieves the smaller norm of joint torque than that from conventional inverse kinematics algorithm. |
이상엽, 이해성, 김승연, 박재흥 Failure Recovery Strategy for Linear Triple Peg-in-Hole Assembly Conference/Workshop The 15th Korea Robotics Society Annual Conference, Kangwon, Pyeongchang, 18.08.2020. @misc{Lee2020, title = {Failure Recovery Strategy for Linear Triple Peg-in-Hole Assembly}, author = {이상엽, 이해성, 김승연, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/02/20200206_054426.pdf}, year = {2020}, date = {2020-08-18}, booktitle = {Failure Recovery Strategy for Linear Triple Peg-in-Hole Assembly}, address = {The 15th Korea Robotics Society Annual Conference, Kangwon, Pyeongchang}, abstract = {In this paper, we propose a failure recovery strategy for the linear triple peg-in-hole which is one of assembly tasks of autonomous furniture assembly. When the search surface is smaller than the search range, the assembly may fail by falling the part out of the search surface. The failure recovery strategy consists of a failure detection phase and a rotation phase to recover the status of the peg part to its pre-failure state. In the rotation phase based on torque control, the position of the part is compliant and the right peg of the part becomes the center of rotation. The proposed strategy is verified by assembling the back parts of the IKEA chair using a seven degree of freedom manipulator. In 100 trials, the parts were assembled 100 percent and the average assembly time was 6.44 seconds.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this paper, we propose a failure recovery strategy for the linear triple peg-in-hole which is one of assembly tasks of autonomous furniture assembly. When the search surface is smaller than the search range, the assembly may fail by falling the part out of the search surface. The failure recovery strategy consists of a failure detection phase and a rotation phase to recover the status of the peg part to its pre-failure state. In the rotation phase based on torque control, the position of the part is compliant and the right peg of the part becomes the center of rotation. The proposed strategy is verified by assembling the back parts of the IKEA chair using a seven degree of freedom manipulator. In 100 trials, the parts were assembled 100 percent and the average assembly time was 6.44 seconds. |
이해성, 이상엽, 백지영, 박재흥 조립형 가구의 다양한 부품 조립을 위한 동작모듈의 설계 Conference/Workshop The 15th Korea Robotics Society Annual Conference, Kangwon, Pyeongchang, 18.08.2020. @misc{LeeKroc2020, title = {조립형 가구의 다양한 부품 조립을 위한 동작모듈의 설계}, author = {이해성, 이상엽, 백지영, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/05/20200529_081453.pdf}, year = {2020}, date = {2020-08-18}, address = {The 15th Korea Robotics Society Annual Conference, Kangwon, Pyeongchang}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
Suhan Park, Jiyeong Baek, Seungyeon Kim, Jaeheung Park Rigid Grasp Candidate Generation for Assembly Tasks Conference/Workshop IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM), Boston, USA (Virtual Conference), 06.07.2020. @misc{Park2020Rigid, title = {Rigid Grasp Candidate Generation for Assembly Tasks}, author = {Suhan Park, Jiyeong Baek, Seungyeon Kim and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/08/grasp_aim2020_revised.pdf}, year = {2020}, date = {2020-07-06}, address = {IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM), Boston, USA (Virtual Conference)}, abstract = {In this study, a new method is proposed to generate rigid grasp candidates for assembly tasks. Although the majority of the recent research is aimed at the generation of candidates for grasping several types of objects, our method focuses on tasks requiring precise and rigid contact. For instance, assembling furniture requires higher grasping force and precision than bin picking tasks or pick and place tasks. This means that the pose information of the object should be accurate for assembly tasks, and the object should not slip while being held. Given these constraints, we generated possible grasp pose candidates.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this study, a new method is proposed to generate rigid grasp candidates for assembly tasks. Although the majority of the recent research is aimed at the generation of candidates for grasping several types of objects, our method focuses on tasks requiring precise and rigid contact. For instance, assembling furniture requires higher grasping force and precision than bin picking tasks or pick and place tasks. This means that the pose information of the object should be accurate for assembly tasks, and the object should not slip while being held. Given these constraints, we generated possible grasp pose candidates. |
Haeseong Lee, Jaeheung Park Contact States Estimation Algorithm Using Fuzzy Logic in Peg-in-holeAssembly Conference/Workshop 17th International Conference on Ubiquitous Robots (UR), Kyoto, Japan, 22.06.2020. @misc{Lee@fuzzy, title = {Contact States Estimation Algorithm Using Fuzzy Logic in Peg-in-holeAssembly}, author = {Haeseong Lee, Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/07/Contact-States-Estimation-Algorithm-Using-Fuzzy-Logic-in-Peg-in-hole-Assembly.pdf}, year = {2020}, date = {2020-06-22}, address = {17th International Conference on Ubiquitous Robots (UR), Kyoto, Japan}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
Jaeheung Park Keynote: Compliant Whole-body Control for Real-World Interactions Conference/Workshop IEEE ICRA 2020 Virtual Conference, 2020-06-11, 11.06.2020. @misc{Park2020bb, title = {Keynote: Compliant Whole-body Control for Real-World Interactions}, author = {Jaeheung Park}, url = {https://ieeetv.ieee.org/icra-2020-keynote-compliant-whole-body-control-for-real-world-interactions}, year = {2020}, date = {2020-06-11}, address = {IEEE ICRA 2020 Virtual Conference, 2020-06-11}, abstract = {As more complicated robots such as humanoid robots are expected to operate in complex environments, whole-body control has greatly progressed for robots operating in multiple contact scenarios. Now, various approaches can be used to create the compliant whole-body motion of robots, depending on hardware and control methods. In this talk, I would like to share my experience of implementing compliant whole-body motion control. The first one is on a torque-controlled robot using a whole-body controller based on the operational space control framework, and the second one is on a position-controlled robot by creating compliant motion using disturbance estimation. Implementation issues will be also discussed in these approaches. Finally, other exciting work in our research group will be briefly presented.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } As more complicated robots such as humanoid robots are expected to operate in complex environments, whole-body control has greatly progressed for robots operating in multiple contact scenarios. Now, various approaches can be used to create the compliant whole-body motion of robots, depending on hardware and control methods. In this talk, I would like to share my experience of implementing compliant whole-body motion control. The first one is on a torque-controlled robot using a whole-body controller based on the operational space control framework, and the second one is on a position-controlled robot by creating compliant motion using disturbance estimation. Implementation issues will be also discussed in these approaches. Finally, other exciting work in our research group will be briefly presented. |
안준휘, 이호상, 박재흥 간접적ZMP최적화를 통한 작업공간 전신 토크제어 Conference/Workshop 11.02.2020. @misc{안준휘2020, title = {간접적ZMP최적화를 통한 작업공간 전신 토크제어}, author = {안준휘, 이호상, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/02/20200206_020540.pdf}, year = {2020}, date = {2020-02-11}, abstract = {This paper shows wholebody dynamics torque controller through quadratic programming. To minimize zmp(zero moment point), we suggest indirect method to optimize zmp through quadratic programming. Through suggested controller, better contact stability was achieved and 1Khz of control rate for 31dof humanoid robot is also achieved. }, keywords = {}, pubstate = {published}, tppubtype = {presentation} } This paper shows wholebody dynamics torque controller through quadratic programming. To minimize zmp(zero moment point), we suggest indirect method to optimize zmp through quadratic programming. Through suggested controller, better contact stability was achieved and 1Khz of control rate for 31dof humanoid robot is also achieved. |
박수민, 박재흥 ZMP를 고려한 수직 방향의 무게 중심 궤적 생성 Conference/Workshop The 15th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, 10.02.2020. @misc{Park2020, title = {ZMP를 고려한 수직 방향의 무게 중심 궤적 생성}, author = {박수민, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/06/SuminPark_KROC2020.pdf}, year = {2020}, date = {2020-02-10}, urldate = {2020-02-10}, address = {The 15th Korea Robotics Society Annual Conference, Pyeongchang ,Korea}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
박범영, 김동현, 김준형, 박재흥 조이스틱을 사용한 휴머노이드 로봇 보행 전략(Humanoid robot walking strategy using joystick) Conference/Workshop The 15th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, 10.02.2020. @misc{Park2019b, title = {조이스틱을 사용한 휴머노이드 로봇 보행 전략(Humanoid robot walking strategy using joystick)}, author = {박범영, 김동현, 김준형, 박재흥}, url = {http://dyros.snu.ac.kr/20200205_014756/}, year = {2020}, date = {2020-02-10}, address = {The 15th Korea Robotics Society Annual Conference, Pyeongchang ,Korea}, abstract = {In this paper, we describe a strategy for giving humanoid robot walking commands with joysticks. Using the joystick to give a humanoid robot a walking command is more intuitive than using a traditional UI keyboard or mouse. Robot walking using joysticks requires a walking command system different from traditional robot walking. In this paper, we describe the structure of the system that the robot can move forward or stop under the command of joystick. In addition, the paper introduces the determination of the position of the foot for walking or stopping according to the joystick command. The results are verified by simulation.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this paper, we describe a strategy for giving humanoid robot walking commands with joysticks. Using the joystick to give a humanoid robot a walking command is more intuitive than using a traditional UI keyboard or mouse. Robot walking using joysticks requires a walking command system different from traditional robot walking. In this paper, we describe the structure of the system that the robot can move forward or stop under the command of joystick. In addition, the paper introduces the determination of the position of the foot for walking or stopping according to the joystick command. The results are verified by simulation. |
2019 |
Minsoo Kim, Joonwoo Ahn, Jaeheung Park Global Planning Method for Visiting Roads with Parking Spaces in Priority using Rural Postman Problem Conference/Workshop 2019 IEEE Intelligent Transportation Systems Conference (ITSC), Auckland, New Zealand, 18.11.2019. @misc{Kim2019, title = {Global Planning Method for Visiting Roads with Parking Spaces in Priority using Rural Postman Problem}, author = {Minsoo Kim, Joonwoo Ahn and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/12/20191216_054629.pdf}, doi = {10.1109/ITSC.2019.8917318}, isbn = {978-1-5386-7024-8}, year = {2019}, date = {2019-11-18}, booktitle = {Global Planning Method for Visiting Roads with Parking Spaces in Priority using Rural Postman Problem}, journal = {2019 IEEE Intelligent Transportation Systems Conference (ITSC)}, pages = {6}, address = {2019 IEEE Intelligent Transportation Systems Conference (ITSC), Auckland, New Zealand}, school = {Colleage of Convergence, Seoul National University}, abstract = {Global planning is one of the key components in autonomous driving. At a parking lot, the purpose of the global planning is to find an empty parking space. Therefore, the autonomous vehicle is to visit roads with parking spaces until the empty space is found. In this paper, to achieve the purpose, the global planning is formulated as the RuralPostman Problem (RPP) and a genetic algorithm (GA) is used to solve the RPP. However, GA produces a plan to randomly visit roads, which may not be efficient. Therefore, in this paper, we propose a cost function considering the parking availability, order of visiting, length of each road segment, and forward-backward switching of a vehicle. The cost function enables generating a plan by prioritizing visiting the roads with parking spaces. Experimental results show that solving the RPP with the proposed cost function generates a more efficient plan than that produced by solving the RPP without the cost function.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } Global planning is one of the key components in autonomous driving. At a parking lot, the purpose of the global planning is to find an empty parking space. Therefore, the autonomous vehicle is to visit roads with parking spaces until the empty space is found. In this paper, to achieve the purpose, the global planning is formulated as the RuralPostman Problem (RPP) and a genetic algorithm (GA) is used to solve the RPP. However, GA produces a plan to randomly visit roads, which may not be efficient. Therefore, in this paper, we propose a cost function considering the parking availability, order of visiting, length of each road segment, and forward-backward switching of a vehicle. The cost function enables generating a plan by prioritizing visiting the roads with parking spaces. Experimental results show that solving the RPP with the proposed cost function generates a more efficient plan than that produced by solving the RPP without the cost function. |
Jaesug Jung, Donghyeon Kim,, Jaeheung Park Operational Space Control Framework for Torque Controlled Humanoid Robots with Joint Elasticity Conference/Workshop 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems, 03.11.2019. @misc{Jung2019, title = {Operational Space Control Framework for Torque Controlled Humanoid Robots with Joint Elasticity}, author = {Jaesug Jung, Donghyeon Kim, and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/01/20200102_015900.pdf}, year = {2019}, date = {2019-11-03}, booktitle = {Operational Space Control Framework for Torque Controlled Humanoid Robots with Joint Elasticity}, journal = {2019 IEEE/RSJ International Conference on Intelligent Robots and Systems}, address = {2019 IEEE/RSJ International Conference on Intelligent Robots and Systems}, abstract = {Torque controlled robots have the capability of implementing compliant behavior with back-drivability. In practice, however, joint elasticity often prevents an accurate position tracking performance of a robot. In particular, humanoid robots are influenced more by elasticity because of a long kinematic chain between the feet and hands. In this paper, we present a new inverse dynamics based control approach for torque controlled humanoid robots with joint elasticity. When formulating the operational space control framework, feedback control consists of only motor-related parts with measured motor angle values, and the link dynamics is compensated by the feedforward terms. The experiment results of the proposed approach show a noticeable improvement in the position tracking performance in 6-DoF manipulator. Finally, the proposed method was applied to a torque controlled biped robot for walking. Both stiff motion control of the CoM and compliant motion control of the foot were simultaneously achieved, demonstrating the advantage of the torque controlled robot.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } Torque controlled robots have the capability of implementing compliant behavior with back-drivability. In practice, however, joint elasticity often prevents an accurate position tracking performance of a robot. In particular, humanoid robots are influenced more by elasticity because of a long kinematic chain between the feet and hands. In this paper, we present a new inverse dynamics based control approach for torque controlled humanoid robots with joint elasticity. When formulating the operational space control framework, feedback control consists of only motor-related parts with measured motor angle values, and the link dynamics is compensated by the feedforward terms. The experiment results of the proposed approach show a noticeable improvement in the position tracking performance in 6-DoF manipulator. Finally, the proposed method was applied to a torque controlled biped robot for walking. Both stiff motion control of the CoM and compliant motion control of the foot were simultaneously achieved, demonstrating the advantage of the torque controlled robot. |
Sumin Park, Jaeheung Park Vertical COM Motion Generation to Reduce Slipping and Mechanical Work during Walking Conference/Workshop IEEE-RAS International Conference on Humanoid Robots (Humanoids), Toronto, Canada, Kanako Miura Award, 17.10.2019. @misc{Park2019b, title = {Vertical COM Motion Generation to Reduce Slipping and Mechanical Work during Walking}, author = {Sumin Park and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/11/Vertical-COM-motion.pdf}, year = {2019}, date = {2019-10-17}, journal = {IEEE-RAS International Conference on Humanoid Robots (Humanoids), Toronto, Canada, Kanako Miura Award}, address = {IEEE-RAS International Conference on Humanoid Robots (Humanoids), Toronto, Canada, Kanako Miura Award}, abstract = {A robot foot can slip when the horizontal shear force acting on the foot exceeds the frictional force between the foot and the ground. In the linear inverted-pendulum (LIP) model, the vertical height of the center of mass (COM) is kept constant, and the vertical force is always equal to the gravitational force at any walking speed. However, the horizontal force increases upon increasing the walking speed. This restriction on the vertical force in the LIP model can cause the robot foot to slip at fast walking speeds, as the horizontal force can exceed the frictional force, which is proportional to the vertical force. In this study, we present an optimization method to generate vertical COM motion to maintain the utilized coefficient of friction (uCOF) less than the available coefficient of friction between the foot and the ground, and to minimize the mechanical work of the COM. Vertical motions at various speeds are generated using the proposed optimization method. Subsequently, the generated COM motion patterns are used as reference trajectories of the COM in robot simulation. Optimization and simulation results demonstrate that the mechanical work and uCOF decrease because of the vertical motion. This study suggests a way to generate slip-safe and energy-efficient COM patterns, which, in turn, overcome the limitations of the LIP model by adding vertical COM motion.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } A robot foot can slip when the horizontal shear force acting on the foot exceeds the frictional force between the foot and the ground. In the linear inverted-pendulum (LIP) model, the vertical height of the center of mass (COM) is kept constant, and the vertical force is always equal to the gravitational force at any walking speed. However, the horizontal force increases upon increasing the walking speed. This restriction on the vertical force in the LIP model can cause the robot foot to slip at fast walking speeds, as the horizontal force can exceed the frictional force, which is proportional to the vertical force. In this study, we present an optimization method to generate vertical COM motion to maintain the utilized coefficient of friction (uCOF) less than the available coefficient of friction between the foot and the ground, and to minimize the mechanical work of the COM. Vertical motions at various speeds are generated using the proposed optimization method. Subsequently, the generated COM motion patterns are used as reference trajectories of the COM in robot simulation. Optimization and simulation results demonstrate that the mechanical work and uCOF decrease because of the vertical motion. This study suggests a way to generate slip-safe and energy-efficient COM patterns, which, in turn, overcome the limitations of the LIP model by adding vertical COM motion. |
Jaeheung Park Keynote talks: Whole-body-control Conference/Workshop IEEE-RAS International Conference on Humanoid Robots (Humanoids), Toronto, Canada, 16.10.2019. @misc{JPark2019, title = {Keynote talks: Whole-body-control}, author = {Jaeheung Park}, year = {2019}, date = {2019-10-16}, address = {IEEE-RAS International Conference on Humanoid Robots (Humanoids), Toronto, Canada}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
Sanghyun Kim, Keunwoo Jang, Suhan Park, Yisoo Lee, Sang Yup Lee, Jaeheung Park Whole-body Control of Non-holonomic Mobile Manipulator Based on Hierarchical Quadratic Programming and Continuous Task Transition Conference/Workshop IEEE International Conference on Advanced Robotics & Mechatronics (ARM), Osaka, Japan, Best Conference Paper Award, 03.07.2019. @misc{ICARM2019, title = {Whole-body Control of Non-holonomic Mobile Manipulator Based on Hierarchical Quadratic Programming and Continuous Task Transition}, author = {Sanghyun Kim and Keunwoo Jang and Suhan Park and Yisoo Lee and Sang Yup Lee and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/04/root_final.pdf http://dyros.snu.ac.kr/wp-content/uploads/2019/07/ARM2019_award-1024x768.jpg http://dyros.snu.ac.kr/wp-content/uploads/2019/07/20190720_221033-1024x512.jpg}, year = {2019}, date = {2019-07-03}, journal = {IEEE International Conference on Advanced Robotics & Mechatronics (ARM), Osaka, Japan, Best Conference Paper Award}, publisher = {IEEE International Conference on Advanced Robotics & Mechatronics (ARM), Osaka, Japan, Best Conference Paper Award}, address = {IEEE International Conference on Advanced Robotics & Mechatronics (ARM), Osaka, Japan, Best Conference Paper Award}, abstract = {Whole-body manipulation for a non-holonomic mobile manipulator has potential in various fields including service and industrial robotics because it can provide a great deal of manipulation capability and mobility. This paper presents the whole-body controller of the non-holonomic mobile manipulator for achieving complex tasks effectively. The proposed controller is designed based on Hierarchical Quadratic Programming (HQP). Thus, the controller can deal with both equality and inequality tasks. Moreover, by applying the continuous task transition strategy, our controller can compute the continuous control input when the priorities of the tasks are changed. The proposed control scheme was implemented on a mobile base with a 7 DoFs robotic arm and its performance was validated during various experimental scenarios.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } Whole-body manipulation for a non-holonomic mobile manipulator has potential in various fields including service and industrial robotics because it can provide a great deal of manipulation capability and mobility. This paper presents the whole-body controller of the non-holonomic mobile manipulator for achieving complex tasks effectively. The proposed controller is designed based on Hierarchical Quadratic Programming (HQP). Thus, the controller can deal with both equality and inequality tasks. Moreover, by applying the continuous task transition strategy, our controller can compute the continuous control input when the priorities of the tasks are changed. The proposed control scheme was implemented on a mobile base with a 7 DoFs robotic arm and its performance was validated during various experimental scenarios. |
Sang Yup Lee, Mingon Kim, Jaeheung Park Kinesthetic Teaching of Position-Controlled Robot Manipulators Using Disturbance Observer Conference/Workshop Forthcoming Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. @misc{Lee2019, title = {Kinesthetic Teaching of Position-Controlled Robot Manipulators Using Disturbance Observer}, author = {Sang Yup Lee, Mingon Kim, Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/07/20190703_071119.pdf}, year = {2019}, date = {2019-06-24}, journal = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea}, address = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea}, abstract = {Manufacturing robots very often require reprogramming tasks due to the frequent change in the production process. Kinesthetic teaching is one of the efficient ways to re-programming the task for robots. In this paper, the interaction is estimated by the disturbance observer based on the position controllers at each joint without additional sensors. Then, the compliant motion is created in the direction of the observed disturbance. The experiment demonstrates the capability of kinesthetic teaching through the physical guidance.}, keywords = {}, pubstate = {forthcoming}, tppubtype = {presentation} } Manufacturing robots very often require reprogramming tasks due to the frequent change in the production process. Kinesthetic teaching is one of the efficient ways to re-programming the task for robots. In this paper, the interaction is estimated by the disturbance observer based on the position controllers at each joint without additional sensors. Then, the compliant motion is created in the direction of the observed disturbance. The experiment demonstrates the capability of kinesthetic teaching through the physical guidance. |
Joonwoo Ahn, Jaeheung Park End-To-End Driving in Unstructured Environments Using Conditional Imitation Learning Conference/Workshop Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. @misc{JW2019, title = {End-To-End Driving in Unstructured Environments Using Conditional Imitation Learning}, author = {Joonwoo Ahn and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/07/UR19_0204_FI_End-To-End-Driving-in-Unstructured-Environments-Using-Conditional-Imitation-Learning-4.pdf}, year = {2019}, date = {2019-06-24}, journal = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea}, address = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea}, abstract = {We propose a method using conditional imitation learning for autonomous driving on unstructured environments. This method is to train a driving policy to imitate an expert action about the input information. Also, the driving policy is classified according to conditional driving purposes. We show the effectiveness of the proposed method in Carla simulator.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } We propose a method using conditional imitation learning for autonomous driving on unstructured environments. This method is to train a driving policy to imitate an expert action about the input information. Also, the driving policy is classified according to conditional driving purposes. We show the effectiveness of the proposed method in Carla simulator. |
Minsoo Kim, Jaeheung Park Reinforcement Learning to Drive in Unstructured Environments Conference/Workshop Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. @misc{MS2019, title = {Reinforcement Learning to Drive in Unstructured Environments}, author = {Minsoo Kim and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/07/UR19_0205_FI_Reinforcement-Learning-to-Drive-in-Unstructured-Environments.pdf}, year = {2019}, date = {2019-06-24}, journal = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019}, address = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea}, abstract = {In this paper, we present a method to drive in unstructured environments using reinforcement learning. We propose construction of state and reward in reinforcement learning for safe driving at unstructured environments. Experimental results show the performance of the proposed method in a simulation.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this paper, we present a method to drive in unstructured environments using reinforcement learning. We propose construction of state and reward in reinforcement learning for safe driving at unstructured environments. Experimental results show the performance of the proposed method in a simulation. |
Chanwoo Ahn, Jaeheung Park Real-Time Segmentation of Drivable Area Using Supervised Learning Conference/Workshop Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. @misc{CW2019, title = {Real-Time Segmentation of Drivable Area Using Supervised Learning }, author = {Chanwoo Ahn and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/07/UR19_0220_FI_Real-Time-Segmentation-of-Drivable-Area-Using-Supervised-Learning-.pdf}, year = {2019}, date = {2019-06-24}, journal = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots)}, address = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea}, abstract = {Supervised learning is used to determine the driving range of images in self-driving. However, conventional methods cannot be used for actual vehicle driving due to their low accuracy and high computational time. Therefore, in this paper, we propose methods to address these problems by, i) simplification of class, ii) image resizing and iii) transmitting frequency tuning.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } Supervised learning is used to determine the driving range of images in self-driving. However, conventional methods cannot be used for actual vehicle driving due to their low accuracy and high computational time. Therefore, in this paper, we propose methods to address these problems by, i) simplification of class, ii) image resizing and iii) transmitting frequency tuning. |
Jee-eun Lee, Jaeheung Park Kinematic Parameter Calibration for Humanoid Robot Using Relative Pose Measurement in Walking Motion Conference/Workshop Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. @misc{JE2019, title = {Kinematic Parameter Calibration for Humanoid Robot Using Relative Pose Measurement in Walking Motion }, author = {Jee-eun Lee and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/07/UR19_0083_FI_Kinematic-Parameter-Calibration-for-Humanoid-Robot-Using-Relative-Pose-Measurement-in-Walking-Motion-.pdf}, year = {2019}, date = {2019-06-24}, journal = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea}, address = {Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea}, abstract = {Robots in use often need to be re-calibrated, due to the mechanical part deformation resulted from unexpected collisions. However, current calibration algorithms are not easy to be applied to humanoids and they require complex experiment setup. This is because humanoids have difficulties in gathering absolute pose measurement since they do not have a fixed frame on its body. On the other hand, the relative pose over time can be easily measured by a camera or a lidar that are usually attached to the humanoids’ head. In this paper, we present a convenient and low-cost calibration method for humanoids by using relative pose measurements in walking motion. Our framework starts from the assumption that the sole of the support foot at every single footstep can be thought of as the temporary fixed frame while walking. Then, based on the decision of the supporting foot, we model the forward kinematics of an open chain from the foot to the head. In the description of a kinematic model, we use the local product of exponentials (POE) formula to take advantages of its simplicity and continuity in computation. Then, we solve the least square problem formulated to minimize the error between the actual measurement and nominal value computed from the corresponding joint angle data. Lastly, simulation results will be discussed to verify the feasibility of this framework.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } Robots in use often need to be re-calibrated, due to the mechanical part deformation resulted from unexpected collisions. However, current calibration algorithms are not easy to be applied to humanoids and they require complex experiment setup. This is because humanoids have difficulties in gathering absolute pose measurement since they do not have a fixed frame on its body. On the other hand, the relative pose over time can be easily measured by a camera or a lidar that are usually attached to the humanoids’ head. In this paper, we present a convenient and low-cost calibration method for humanoids by using relative pose measurements in walking motion. Our framework starts from the assumption that the sole of the support foot at every single footstep can be thought of as the temporary fixed frame while walking. Then, based on the decision of the supporting foot, we model the forward kinematics of an open chain from the foot to the head. In the description of a kinematic model, we use the local product of exponentials (POE) formula to take advantages of its simplicity and continuity in computation. Then, we solve the least square problem formulated to minimize the error between the actual measurement and nominal value computed from the corresponding joint angle data. Lastly, simulation results will be discussed to verify the feasibility of this framework. |
Jaeheung Park Continuous Task Transition Approach for Robots with Prioritized Tasks Conference/Workshop IEEE ICRA 2019 Workshop, Montreal, Canada, 2019-05-24, 24.05.2019. @misc{Park2019b, title = {Continuous Task Transition Approach for Robots with Prioritized Tasks}, author = {Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/07/Jaeheung-Park-ICRA2019-workshop-presenta.pdf}, year = {2019}, date = {2019-05-24}, journal = {IEEE ICRA 2019 Workshop, Montreal, Canada}, address = {IEEE ICRA 2019 Workshop, Montreal, Canada, 2019-05-24}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
Mingon Kim, Daegyu Lim, Jaeheung Park Online Walking Pattern Generation for Humanoid Robot with Compliant Motion Control Conference/Workshop IEEE International Conference on Robotics and Automation (ICRA), Montreal, Canada, 20.05.2019. @misc{ICRA2019Kim, title = {Online Walking Pattern Generation for Humanoid Robot with Compliant Motion Control}, author = {Mingon Kim and Daegyu Lim and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/01/icra2019_mg_f7.pdf}, year = {2019}, date = {2019-05-20}, journal = {IEEE International Conference on Robotics and Automation (ICRA), Montreal, Canada}, publisher = {IEEE International Conference on Robotics and Automation (ICRA), Montreal, Canada}, address = {IEEE International Conference on Robotics and Automation (ICRA), Montreal, Canada}, abstract = {Compliant motion of humanoids is one of the most important characteristics for interaction with the human and environment in the real world. Especially during walking, the compliant motion ensures stable contact between the foot and the ground, but the walking stability is degraded due to the position tracking performance and unknown disturbance. To address this problem of possible instability of humanoid walking with compliant motion control, this paper proposes a real-time walking pattern generation considering the motion control performance of the robot. The dynamics model of the robot with the motion controller is described as a second order system approximating position tracing performance with a linear inverted pendulum model for the relation between Zero-Moment Point (ZMP) and Center of Mass (CoM). The CoM trajectory is calculated using the preview control based on the dynamics model and the current state of the robot. Therefore, even if the robot has the low tracking performance due to compliant motion control of the robot, the walking stability can be ensured. The proposed method was implemented on our humanoid robot, DYROS-JET, and its performance was demonstrated by improved stability during walking.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } Compliant motion of humanoids is one of the most important characteristics for interaction with the human and environment in the real world. Especially during walking, the compliant motion ensures stable contact between the foot and the ground, but the walking stability is degraded due to the position tracking performance and unknown disturbance. To address this problem of possible instability of humanoid walking with compliant motion control, this paper proposes a real-time walking pattern generation considering the motion control performance of the robot. The dynamics model of the robot with the motion controller is described as a second order system approximating position tracing performance with a linear inverted pendulum model for the relation between Zero-Moment Point (ZMP) and Center of Mass (CoM). The CoM trajectory is calculated using the preview control based on the dynamics model and the current state of the robot. Therefore, even if the robot has the low tracking performance due to compliant motion control of the robot, the walking stability can be ensured. The proposed method was implemented on our humanoid robot, DYROS-JET, and its performance was demonstrated by improved stability during walking. |
Jongseong Ko, Jaehoon Sim, Jaeheung Park Design of Configuration-changeable Modular Robot Test Platform Conference/Workshop The 34rd Institute of Control, Robotics and Systems Annual Conference (ICROS 2019), Gyeongju, Korea, 17.05.2019. @misc{Ko2019, title = {Design of Configuration-changeable Modular Robot Test Platform}, author = {Jongseong Ko, Jaehoon Sim, Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/05/Ko2019.pdf}, year = {2019}, date = {2019-05-17}, journal = {The 34rd Institute of Control, Robotics and Systems Annual Conference (ICROS 2019), Gyeongju, Korea}, address = {The 34rd Institute of Control, Robotics and Systems Annual Conference (ICROS 2019), Gyeongju, Korea}, abstract = {In this paper, we present the development result of multi-purpose test platform hardware which has a simple and robust modular structure and can be modified into various forms according to research needs. Through the proposed standardized module, it is expected that more reliable and consistent data can be measured than before. In addition, the research period can be shortened and the cost can be saved.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this paper, we present the development result of multi-purpose test platform hardware which has a simple and robust modular structure and can be modified into various forms according to research needs. Through the proposed standardized module, it is expected that more reliable and consistent data can be measured than before. In addition, the research period can be shortened and the cost can be saved. |
Junhyung Kim, Jaehoon Sim, Jaeheung Park Humanoid Fall Recovery Strategy using Air Thrust Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 23.01.2019. @misc{Kim2019, title = {Humanoid Fall Recovery Strategy using Air Thrust}, author = {Junhyung Kim, Jaehoon Sim and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/02/TP2-29.pdf}, year = {2019}, date = {2019-01-23}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, abstract = {본 논문에서는 휴머노이드 로봇의 측면에 외력이 가해졌을 때 공기 추력을 이용하여 보행 중 가해진 외력에 대한 휴 머노이드 로봇의 넘어짐 방지하는 전략을 제안한다. 휴머노이드 로봇의 양쪽 어깨에 붙어 있는 공기 추력 구동기는 넘 어짐을 방지하기 위하여 공기 추력을 발생시킨다. 휴머노이드 로봇의 상태가 넘어지고 있는지를 판단하기 위하여 캡처 포인트를 사용하였다. 캡처포인트가 지지다각형을 넘어가면 휴머노이드 로봇이 넘어지는 것으로 판단되면 공기 추력 구동기가 휴머노이드 로봇이 넘어지는 방향의 반대 방향으로 힘을 발생시켜 넘어짐을 회복한다. 본 논문에서 제안하는 넘어짐 회복 전략을 휴머노이드 로봇 DYROS-JET 를 이용한 실제 실험을 통하여 검증하였다. 또한 실험 영상은 https://youtu.be/ij6EltOWbf8 을 통해 확인 할 수 있다.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } 본 논문에서는 휴머노이드 로봇의 측면에 외력이 가해졌을 때 공기 추력을 이용하여 보행 중 가해진 외력에 대한 휴 머노이드 로봇의 넘어짐 방지하는 전략을 제안한다. 휴머노이드 로봇의 양쪽 어깨에 붙어 있는 공기 추력 구동기는 넘 어짐을 방지하기 위하여 공기 추력을 발생시킨다. 휴머노이드 로봇의 상태가 넘어지고 있는지를 판단하기 위하여 캡처 포인트를 사용하였다. 캡처포인트가 지지다각형을 넘어가면 휴머노이드 로봇이 넘어지는 것으로 판단되면 공기 추력 구동기가 휴머노이드 로봇이 넘어지는 방향의 반대 방향으로 힘을 발생시켜 넘어짐을 회복한다. 본 논문에서 제안하는 넘어짐 회복 전략을 휴머노이드 로봇 DYROS-JET 를 이용한 실제 실험을 통하여 검증하였다. 또한 실험 영상은 https://youtu.be/ij6EltOWbf8 을 통해 확인 할 수 있다. |
Sangyup Lee, Suhan Park, Keunwoo Jang, Sanghyun Kim, Seugnyeon Lee, Jaeheung Park A Comparative Study of Tracking Control with Closed Loop Inverse Kinematics Algorithm Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 23.01.2019. @misc{Kim2019, title = {A Comparative Study of Tracking Control with Closed Loop Inverse Kinematics Algorithm}, author = {Sangyup Lee, Suhan Park, Keunwoo Jang, Sanghyun Kim, Seugnyeon Lee and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/02/20190221_030023.pdf}, year = {2019}, date = {2019-01-23}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
장근우, 김상현, 박수한, 이상엽, 박재흥 Self-collision avoidance algorithm for robot manipulator using continuous task transition algorithm Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, , 22.01.2019. @misc{장근우2019, title = {Self-collision avoidance algorithm for robot manipulator using continuous task transition algorithm}, author = {장근우, 김상현, 박수한, 이상엽, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/01/20190130_055223.pdf}, year = {2019}, date = {2019-01-22}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, }, abstract = {본 논문에서는 매니퓰레이터의 내부 링크 간 자가 충돌을 방지하는 알고리즘을 제안한다. 알고리즘은 해의 불연속성 없 이 작업의 삽입이나 삭제가 가능한 작업 전이 알고리즘을 기반으로 한다. 또한, 알고리즘은 매니퓰레이터의 내부 링크 간 거리 값에 따라 일정 거리 이하부터 서로 멀어지는 방향의 작업을 삽입하여 충돌을 회피한다. 알고리즘의 검증은 7자 유도 매니퓰레이터를 통해 이루어졌다. 또한, 실험 영상은 https://youtu.be/FH2wKkQfCIc 을 통해 확인할 수 있다.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } 본 논문에서는 매니퓰레이터의 내부 링크 간 자가 충돌을 방지하는 알고리즘을 제안한다. 알고리즘은 해의 불연속성 없 이 작업의 삽입이나 삭제가 가능한 작업 전이 알고리즘을 기반으로 한다. 또한, 알고리즘은 매니퓰레이터의 내부 링크 간 거리 값에 따라 일정 거리 이하부터 서로 멀어지는 방향의 작업을 삽입하여 충돌을 회피한다. 알고리즘의 검증은 7자 유도 매니퓰레이터를 통해 이루어졌다. 또한, 실험 영상은 https://youtu.be/FH2wKkQfCIc 을 통해 확인할 수 있다. |
박효종, 심재훈, 박수한, 박재흥 SOEM을 이용한 실시간 휴머노이드 로봇 제어 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. @misc{박효종2019, title = {SOEM을 이용한 실시간 휴머노이드 로봇 제어}, author = {박효종, 심재훈, 박수한, 박재흥}, year = {2019}, date = {2019-01-22}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
백지영, 심재훈, 박재흥 휴머노이드 구동을 위한 딥러닝 음성인식 인터페이스 개발 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. @misc{백지영2019, title = {휴머노이드 구동을 위한 딥러닝 음성인식 인터페이스 개발 }, author = {백지영, 심재훈, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/02/20190221_025629.pdf}, year = {2019}, date = {2019-01-22}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
김규리, 성은호, 심재훈, 박재흥 로봇을 제어하는 족동식 사용자 인터페이스 개발 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. @misc{김규리2019, title = {로봇을 제어하는 족동식 사용자 인터페이스 개발}, author = {김규리, 성은호, 심재훈, 박재흥}, year = {2019}, date = {2019-01-22}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
박규진, 임규범, 박재흥 카메라센서와 라이다 센서 융합에 기반한 개선된 주차공간 검출 시스템 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. @misc{박규진2019, title = {카메라센서와 라이다 센서 융합에 기반한 개선된 주차공간 검출 시스템 }, author = {박규진, 임규범, 박재흥}, year = {2019}, date = {2019-01-22}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
김소연, 김승연, 박재흥 로봇 핸드 제어를 위한 실시간 손 동작 영상 인식 시스템 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. @misc{김소연2019, title = {로봇 핸드 제어를 위한 실시간 손 동작 영상 인식 시스템 }, author = {김소연, 김승연, 박재흥}, year = {2019}, date = {2019-01-22}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
박범영, 심원섭, 김영주, 정용찬, 이은규, 이우현, 김은총, 박재흥 Experiential boarding robot system using industrial robot Conference/Workshop 22.01.2019. @misc{박범영2019, title = {Experiential boarding robot system using industrial robot}, author = {박범영, 심원섭, 김영주, 정용찬, 이은규, 이우현, 김은총, 박재흥}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2020/02/20200205_013850.pdf}, year = {2019}, date = {2019-01-22}, abstract = {본 논문은 산업용 로봇에 기반을 두어 사람이 실제로 탑승하여 로봇을 움직이며 체험할 수 있는 탑승 로봇 시스템에 대해 설명한다. 제안하는 시스템은 KUKA사의 KR-120 로봇을 이용한 체험용 탑승 로봇을 사용한다. 로봇의 말단부에는 사람이 탑승할 수 있는 탑승부가 부착되어 있다. 또한 탑승부에는 로봇에 탑승한 조작자가 로봇을 조종하기 위해 사용하는 조이스틱이 장착되어 있다. 시스템을 검증하는 실험은 사람이 로봇에 탑승 후 로봇을 직접 조종하며 진행되었다. 또한 현재 시스템이 가지고 있는 문제를 파악하고, 이를 보완하기 위한 추후 연구의 방향에 대해 설명한다}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } 본 논문은 산업용 로봇에 기반을 두어 사람이 실제로 탑승하여 로봇을 움직이며 체험할 수 있는 탑승 로봇 시스템에 대해 설명한다. 제안하는 시스템은 KUKA사의 KR-120 로봇을 이용한 체험용 탑승 로봇을 사용한다. 로봇의 말단부에는 사람이 탑승할 수 있는 탑승부가 부착되어 있다. 또한 탑승부에는 로봇에 탑승한 조작자가 로봇을 조종하기 위해 사용하는 조이스틱이 장착되어 있다. 시스템을 검증하는 실험은 사람이 로봇에 탑승 후 로봇을 직접 조종하며 진행되었다. 또한 현재 시스템이 가지고 있는 문제를 파악하고, 이를 보완하기 위한 추후 연구의 방향에 대해 설명한다 |
황순욱, 김동현, 정재석, 박재흥 속도 관측기를 이용한 토크 기반 로봇 제어에서의 위치 추종 성능 향상 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. @misc{황순욱2019, title = {속도 관측기를 이용한 토크 기반 로봇 제어에서의 위치 추종 성능 향상}, author = {황순욱, 김동현, 정재석, 박재흥}, year = {2019}, date = {2019-01-22}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
이지민, 이호상, 안준휘, 박재흥 비전 센서와 능동적 접촉 인지 제어 방법을 융합한 외부 물체의 접촉 인지 및 모서리 예측 전략 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. @misc{이지민2019, title = {비전 센서와 능동적 접촉 인지 제어 방법을 융합한 외부 물체의 접촉 인지 및 모서리 예측 전략}, author = {이지민, 이호상, 안준휘, 박재흥}, year = {2019}, date = {2019-01-22}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
안준휘, 이호상, 박재흥 힘-토크 센서와 전신 동역학 모델을 융합한 보행 로봇의 접촉력 추정 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. @misc{안준휘2019, title = {힘-토크 센서와 전신 동역학 모델을 융합한 보행 로봇의 접촉력 추정}, author = {안준휘, 이호상, 박재흥}, year = {2019}, date = {2019-01-22}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
Seungyeon Kim, Jaeheung Park In-hand rolling motion generation using underactuated fingers with Four-bar linkage chains Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, 21.01.2019. @misc{Kim2019k, title = {In-hand rolling motion generation using underactuated fingers with Four-bar linkage chains}, author = {Seungyeon Kim and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/07/MP1-12.pdf}, year = {2019}, date = {2019-01-21}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
Joowan Kim, Wonje Choi, Jaeheung Park Design Improvement of Compliant Anti-Gravity Rehabilitation Robot for Lumbar Pain Patients Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, 21.01.2019. @misc{JWKim2019, title = {Design Improvement of Compliant Anti-Gravity Rehabilitation Robot for Lumbar Pain Patients}, author = {Joowan Kim, Wonje Choi and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/02/KRoC-19-0139.pdf}, year = {2019}, date = {2019-01-21}, address = {The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea}, abstract = {요통을 위한 운동 치료로 요부 안정화 운동은 가장 권장되는 치료 방법이다. 하지만 요부 안정화 운동은 맨몸 운동으로, 힘이 부족한 노인 및 장애 환자들에게 부하 조절이 힘들며 상,하지에 무리를 줄 수 있는 단점이 있어 효과적인 운동 효과를 보기 힘들다. 이러한 단점을 보완하기 위해 맨몸 운동과 근사한 운동효과를 제공하고 정량적인 운동 부하 조절이 가능한 유연한 항중력 재활 로봇을 개발하였고 선행 실험으로 근전도를 활용한 운동 평가를 진행한 결과, 대부분의 운동은 맨몸 운동과 큰 유사성을 가졌지만 특정 자세 운동에서 근육 활성화 유사성이 상대적으로 낮은 결과가 나왔다. 선행 실험 연구를 토대로 본 논문은 맨몸 요부 안정화 운동과 동등한 운동효과를 내기 위하여, 기 개발된 허리 재활 로봇의 기구적 단점을 보완한 설계 개선점에 대해 기술하였다. }, keywords = {}, pubstate = {published}, tppubtype = {presentation} } 요통을 위한 운동 치료로 요부 안정화 운동은 가장 권장되는 치료 방법이다. 하지만 요부 안정화 운동은 맨몸 운동으로, 힘이 부족한 노인 및 장애 환자들에게 부하 조절이 힘들며 상,하지에 무리를 줄 수 있는 단점이 있어 효과적인 운동 효과를 보기 힘들다. 이러한 단점을 보완하기 위해 맨몸 운동과 근사한 운동효과를 제공하고 정량적인 운동 부하 조절이 가능한 유연한 항중력 재활 로봇을 개발하였고 선행 실험으로 근전도를 활용한 운동 평가를 진행한 결과, 대부분의 운동은 맨몸 운동과 큰 유사성을 가졌지만 특정 자세 운동에서 근육 활성화 유사성이 상대적으로 낮은 결과가 나왔다. 선행 실험 연구를 토대로 본 논문은 맨몸 요부 안정화 운동과 동등한 운동효과를 내기 위하여, 기 개발된 허리 재활 로봇의 기구적 단점을 보완한 설계 개선점에 대해 기술하였다. |
Suhan Park, Jaehoon Sim, Jaeheung Park System Design of Humanoid Robot DYROS-JET Conference/Workshop 2019 IEEE/SICE International Symposium on System Integration, Paris, France, 16.01.2019. @misc{SII2019PARK, title = {System Design of Humanoid Robot DYROS-JET}, author = {Suhan Park and Jaehoon Sim and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/01/0170.pdf}, year = {2019}, date = {2019-01-16}, address = {2019 IEEE/SICE International Symposium on System Integration, Paris, France}, abstract = {To perform various tasks for humans, humanoid robots are designed to comprise multiple joints that provide them with both mobility and manipulability enabled by a kinematic structure similar to that of humans. Herein, we introduce a system design created to control a robot with numerous degrees of freedom. Our communication system is constructed to simultaneously control multiple joints and improve control performance by fully exploiting the control frequency. We also consider the design of the input/output interface for multiple sensors that are essential for stable walking and the manipulation of humanoid robots. A fail-safe algorithm is implemented for the software modules and the controller, and the proposed system design is experimentally verified.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } To perform various tasks for humans, humanoid robots are designed to comprise multiple joints that provide them with both mobility and manipulability enabled by a kinematic structure similar to that of humans. Herein, we introduce a system design created to control a robot with numerous degrees of freedom. Our communication system is constructed to simultaneously control multiple joints and improve control performance by fully exploiting the control frequency. We also consider the design of the input/output interface for multiple sensors that are essential for stable walking and the manipulation of humanoid robots. A fail-safe algorithm is implemented for the software modules and the controller, and the proposed system design is experimentally verified. |
2018 |
Minsung Kim, Gyubeom Im, Joonwoo Ahn, Minsoo Kim, Jaeheung Park Autonomous parking algorithm using automatic selection of forward and backward switching points via model predictive control Conference/Workshop The Korean Society of Automotive Engineers Annual Autumn Conference, Kangwon, Korea, 14.11.2018. @misc{Kim2018b, title = {Autonomous parking algorithm using automatic selection of forward and backward switching points via model predictive control}, author = {Minsung Kim, Gyubeom Im, Joonwoo Ahn, Minsoo Kim, Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/07/18AKSAE_D052-2.pdf}, year = {2018}, date = {2018-11-14}, address = {The Korean Society of Automotive Engineers Annual Autumn Conference, Kangwon, Korea}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
Joonwoo Ahn, Minsoo Kim, Gyubeom Im, Minsung Kim, Jaeheung Park Development of Autonomous Valet Parking System and Simulation Result Conference/Workshop The Korean Society of Automotive Engineers Annual Autumn Conference, Kangwon , Korea, 14.11.2018. @misc{Ahn2018b, title = {Development of Autonomous Valet Parking System and Simulation Result }, author = {Joonwoo Ahn, Minsoo Kim, Gyubeom Im, Minsung Kim, Jaeheung Park }, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/07/20190704_055015.pdf}, year = {2018}, date = {2018-11-14}, address = {The Korean Society of Automotive Engineers Annual Autumn Conference, Kangwon , Korea}, abstract = {In this paper, we propose an efficient and safe method for autonomous valet parking system with a driving strategy considering uncertain environment in parking spaces. The proposed method consists of a vehicle controller with decision-making algorithms, a global and local path generation algorithm, and a three-dimensional LiDAR-based parking lot detection algorithm. The proposed method was validated via simulation results.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this paper, we propose an efficient and safe method for autonomous valet parking system with a driving strategy considering uncertain environment in parking spaces. The proposed method consists of a vehicle controller with decision-making algorithms, a global and local path generation algorithm, and a three-dimensional LiDAR-based parking lot detection algorithm. The proposed method was validated via simulation results. |
Gyubeom Im, Minsung Kim, Joonwoo Ahn, Minsoo Kim, Jaeheung Park V-Rep Simulator for Autonomous Vehicle Research in Parking lot Environment Conference/Workshop The Korean Society of Automotive Engineers Annual Autumn Conference, Kangwon, Korea, 14.11.2018. @misc{Im2018, title = {V-Rep Simulator for Autonomous Vehicle Research in Parking lot Environment}, author = {Gyubeom Im, Minsung Kim, Joonwoo Ahn, Minsoo Kim, Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/07/ksae2018_edward.pdf}, year = {2018}, date = {2018-11-14}, address = {The Korean Society of Automotive Engineers Annual Autumn Conference, Kangwon, Korea}, abstract = {In this paper, we propose an efficient and safe method for autonomous valet parking system with a driving strategy considering uncertain environment in parking spaces. The proposed method consists of a vehicle controller with decision-making algorithms, a global and local path generation algorithm, and a three-dimensional LiDAR-based parking lot detection algorithm. The proposed method was validated via simulation results.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this paper, we propose an efficient and safe method for autonomous valet parking system with a driving strategy considering uncertain environment in parking spaces. The proposed method consists of a vehicle controller with decision-making algorithms, a global and local path generation algorithm, and a three-dimensional LiDAR-based parking lot detection algorithm. The proposed method was validated via simulation results. |
Seungyeon Kim, Jaeheung Park Gripper with Thumb Adduction/Abduction Joint for Enhanced In-hand Orientation Manipulation Conference/Workshop 2018 IEEE-RAS International Conference on Humanoid Robots (Humanoids), Beijing, China, 06.11.2018. @misc{Kim2018b, title = {Gripper with Thumb Adduction/Abduction Joint for Enhanced In-hand Orientation Manipulation}, author = {Seungyeon Kim and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/01/Gripper-with-TAA-joint.pdf}, year = {2018}, date = {2018-11-06}, address = {2018 IEEE-RAS International Conference on Humanoid Robots (Humanoids), Beijing, China}, abstract = {A new gripper with a Thumb Adduction/Abduction joint (TAA joint) is presented in this study. The purpose of the proposed gripper is to enable the orientation control of an object along all axes using in-hand manipulation and the last wrist joint. Using TAA joint motion, it was possible to achieve not only stable grasping but also in-hand orientation control. The gripper can produce 13 out of 16 human grasp motions according to hand taxonomy, including four of the most used hand postures [1][2]. The grasping performance was evaluated using various objects. Some of them were tools for specific tasks such as electric drill, screwdriver, and hammer. The others were objects in daily life such as ball, CD, pen, cup, wallet, and card. To evaluate the capability of in-hand orientation control, the maximum ranges of rotations were also estimated using circular cross-sectional objects. The gripper can consecutively rotate the object along the roll and yaw directions, and this was demonstrated using a spherical object.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } A new gripper with a Thumb Adduction/Abduction joint (TAA joint) is presented in this study. The purpose of the proposed gripper is to enable the orientation control of an object along all axes using in-hand manipulation and the last wrist joint. Using TAA joint motion, it was possible to achieve not only stable grasping but also in-hand orientation control. The gripper can produce 13 out of 16 human grasp motions according to hand taxonomy, including four of the most used hand postures [1][2]. The grasping performance was evaluated using various objects. Some of them were tools for specific tasks such as electric drill, screwdriver, and hammer. The others were objects in daily life such as ball, CD, pen, cup, wallet, and card. To evaluate the capability of in-hand orientation control, the maximum ranges of rotations were also estimated using circular cross-sectional objects. The gripper can consecutively rotate the object along the roll and yaw directions, and this was demonstrated using a spherical object. |
Hosang Lee, Yisoo Lee, Jaeheung Park Contact-Consistent Disturbance Observer for Floating-base Robots Conference/Workshop 2018 International Symposium on Experimental Robotics, Buenos Aires, Nov 5-8, 06.11.2018. @misc{Lee2018b, title = {Contact-Consistent Disturbance Observer for Floating-base Robots}, author = {Hosang Lee and Yisoo Lee and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/02/20190209_150828.pdf}, year = {2018}, date = {2018-11-06}, address = {2018 International Symposium on Experimental Robotics, Buenos Aires, Nov 5-8}, abstract = {In this study, a novel disturbance observer for floating-base robots has been developed using contact information to reduce the estimation errors of joint accelerations. Typically, joint angles are measured using motor encoders and joint accelerations are computed by differentiating them twice in discrete time domain. Therefore, the estimated value inherently has a large numerical error, which is an obstacle for precise on-line estimation of the disturbance. The proposed method computes the joint acceleration error, which does not satisfy the kinematic constraints of the contact position such as a supporting foot of the robot, and incorporates it into the dynamics model to remove its effect on the disturbance torque solution. It is shown through experiments using a legged robot that the proposed contact-consistent observer can estimate the disturbance more accurately than an observer that does not consider contact information.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this study, a novel disturbance observer for floating-base robots has been developed using contact information to reduce the estimation errors of joint accelerations. Typically, joint angles are measured using motor encoders and joint accelerations are computed by differentiating them twice in discrete time domain. Therefore, the estimated value inherently has a large numerical error, which is an obstacle for precise on-line estimation of the disturbance. The proposed method computes the joint acceleration error, which does not satisfy the kinematic constraints of the contact position such as a supporting foot of the robot, and incorporates it into the dynamics model to remove its effect on the disturbance torque solution. It is shown through experiments using a legged robot that the proposed contact-consistent observer can estimate the disturbance more accurately than an observer that does not consider contact information. |
Junhyung Kim, Jaehoon Sim, Jaeheung Park Humanoid fall recovery strategy using air thrust Conference/Workshop WorkShop on “Humanoid Robot Falling: Fall Detection, Damage Prevention, and Recovery Actions” at IROS 2018, Madrid, Spain, 01.10.2018. @misc{Kim2018b, title = {Humanoid fall recovery strategy using air thrust}, author = {Junhyung Kim, Jaehoon Sim and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/02/kim_iros_2018_ws_paper-1.pdf}, year = {2018}, date = {2018-10-01}, address = {WorkShop on “Humanoid Robot Falling: Fall Detection, Damage Prevention, and Recovery Actions” at IROS 2018, Madrid, Spain}, abstract = {In this paper, we propose a fall recovery strategy for humanoid robot against an external force applied to lateral direction by using air thrusts. Each air thrust is attached on both sides of the shoulder of humanoid robot to recover from a fall. To distinguish the state of the humanoid robot, stable state (not falling) or unstable state (falling), we use capture point (CP) which is commonly used for fall detection of the humanoid robot before falling. Considering that the humanoid robot is falling when CP leaves the supporting foot, the air thrust generates force opposed to falling direction to recover the humanoid robot from a fall. After CP comes back inside of the supporting foot, the air thrust stops generating force. This paper also shows the real experimental results of the proposed method for fall recovery using DYROS-JET with air thrusts.}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } In this paper, we propose a fall recovery strategy for humanoid robot against an external force applied to lateral direction by using air thrusts. Each air thrust is attached on both sides of the shoulder of humanoid robot to recover from a fall. To distinguish the state of the humanoid robot, stable state (not falling) or unstable state (falling), we use capture point (CP) which is commonly used for fall detection of the humanoid robot before falling. Considering that the humanoid robot is falling when CP leaves the supporting foot, the air thrust generates force opposed to falling direction to recover the humanoid robot from a fall. After CP comes back inside of the supporting foot, the air thrust stops generating force. This paper also shows the real experimental results of the proposed method for fall recovery using DYROS-JET with air thrusts. |
Sumin Park, Jaeheung Park Effect of Heel Area on Utilized Coefficient of Friction During High-Heeled Walking Conference/Workshop The 20th Congress of the International Ergonomics Association (IEA 2018), Florence, Italy, 29.08.2018. @misc{Park2018, title = {Effect of Heel Area on Utilized Coefficient of Friction During High-Heeled Walking}, author = {Sumin Park and Jaeheung Park}, url = {http://dyros.snu.ac.kr/wp-content/uploads/2019/02/Effect-of-heel-area-on-uCOF.pdf}, year = {2018}, date = {2018-08-29}, urldate = {2018-08-29}, journal = {The 20th Congress of the International Ergonomics Association (IEA 2018)}, volume = {819}, pages = {703-709}, publisher = {Springer}, address = {Florence, Italy}, series = {Advances in Intelligent Systems and Computing }, howpublished = {The 20th Congress of the International Ergonomics Association (IEA 2018)}, key = {Italy}, keywords = {}, pubstate = {published}, tppubtype = {presentation} } |
2020 |
수직 주차에서 AVM 을 이용한 동시적 위치추정 및 지도 생성 Conference/Workshop The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea, 06.11.2020. |
클로소이드(Clothoid)를 사용한 주차 경로 후-처리를 통한 경로 추종 오차 및 승차감 개선 Conference/Workshop The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea, 06.11.2020. |
AVM이미지와 ROTATE-YOLOv3를 이용한 주차 영역 인식 Conference/Workshop The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea, 06.11.2020. |
후보 궤적을 이용한 비구조화된 환경에서의 교차로 인식 및 주행 Conference/Workshop The Korean Society of Automotive Engineers Fall Conference, Jeju, Korea, 06.11.2020. |
Implementation of Position-based Walking to Torque-controlled Humanoid Conference/Workshop 2020 The 20th International Conference on Control, Automation and Systems, 13.10.2020. |
Walking Pattern Generation using MPC with minimization of COM Velocity Fluctuation Conference/Workshop 2020 IEEE International Conference on Control, Automation and System, 13.10.2020. |
주어진 힘에 대해 관절 토크를 최소화하는 역기구학 Conference/Workshop The 15th Korea Robotics Society Annual Conference, Pyeongchang, Korea, 18.08.2020. |
Failure Recovery Strategy for Linear Triple Peg-in-Hole Assembly Conference/Workshop The 15th Korea Robotics Society Annual Conference, Kangwon, Pyeongchang, 18.08.2020. |
조립형 가구의 다양한 부품 조립을 위한 동작모듈의 설계 Conference/Workshop The 15th Korea Robotics Society Annual Conference, Kangwon, Pyeongchang, 18.08.2020. |
Rigid Grasp Candidate Generation for Assembly Tasks Conference/Workshop IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM), Boston, USA (Virtual Conference), 06.07.2020. |
Contact States Estimation Algorithm Using Fuzzy Logic in Peg-in-holeAssembly Conference/Workshop 17th International Conference on Ubiquitous Robots (UR), Kyoto, Japan, 22.06.2020. |
Keynote: Compliant Whole-body Control for Real-World Interactions Conference/Workshop IEEE ICRA 2020 Virtual Conference, 2020-06-11, 11.06.2020. |
간접적ZMP최적화를 통한 작업공간 전신 토크제어 Conference/Workshop 11.02.2020. |
ZMP를 고려한 수직 방향의 무게 중심 궤적 생성 Conference/Workshop The 15th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, 10.02.2020. |
조이스틱을 사용한 휴머노이드 로봇 보행 전략(Humanoid robot walking strategy using joystick) Conference/Workshop The 15th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, 10.02.2020. |
2019 |
Global Planning Method for Visiting Roads with Parking Spaces in Priority using Rural Postman Problem Conference/Workshop 2019 IEEE Intelligent Transportation Systems Conference (ITSC), Auckland, New Zealand, 18.11.2019. |
Operational Space Control Framework for Torque Controlled Humanoid Robots with Joint Elasticity Conference/Workshop 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems, 03.11.2019. |
Vertical COM Motion Generation to Reduce Slipping and Mechanical Work during Walking Conference/Workshop IEEE-RAS International Conference on Humanoid Robots (Humanoids), Toronto, Canada, Kanako Miura Award, 17.10.2019. |
Keynote talks: Whole-body-control Conference/Workshop IEEE-RAS International Conference on Humanoid Robots (Humanoids), Toronto, Canada, 16.10.2019. |
Whole-body Control of Non-holonomic Mobile Manipulator Based on Hierarchical Quadratic Programming and Continuous Task Transition Conference/Workshop IEEE International Conference on Advanced Robotics & Mechatronics (ARM), Osaka, Japan, Best Conference Paper Award, 03.07.2019. |
Kinesthetic Teaching of Position-Controlled Robot Manipulators Using Disturbance Observer Conference/Workshop Forthcoming Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. |
End-To-End Driving in Unstructured Environments Using Conditional Imitation Learning Conference/Workshop Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. |
Reinforcement Learning to Drive in Unstructured Environments Conference/Workshop Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. |
Real-Time Segmentation of Drivable Area Using Supervised Learning Conference/Workshop Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. |
Kinematic Parameter Calibration for Humanoid Robot Using Relative Pose Measurement in Walking Motion Conference/Workshop Conference, Ubiquitous Robots 2019 (The 16th International Conference on Ubiquitous Robots), Jeju, Korea, 24.06.2019. |
Continuous Task Transition Approach for Robots with Prioritized Tasks Conference/Workshop IEEE ICRA 2019 Workshop, Montreal, Canada, 2019-05-24, 24.05.2019. |
Online Walking Pattern Generation for Humanoid Robot with Compliant Motion Control Conference/Workshop IEEE International Conference on Robotics and Automation (ICRA), Montreal, Canada, 20.05.2019. |
Design of Configuration-changeable Modular Robot Test Platform Conference/Workshop The 34rd Institute of Control, Robotics and Systems Annual Conference (ICROS 2019), Gyeongju, Korea, 17.05.2019. |
Humanoid Fall Recovery Strategy using Air Thrust Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 23.01.2019. |
A Comparative Study of Tracking Control with Closed Loop Inverse Kinematics Algorithm Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 23.01.2019. |
Self-collision avoidance algorithm for robot manipulator using continuous task transition algorithm Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, , 22.01.2019. |
SOEM을 이용한 실시간 휴머노이드 로봇 제어 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. |
휴머노이드 구동을 위한 딥러닝 음성인식 인터페이스 개발 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. |
로봇을 제어하는 족동식 사용자 인터페이스 개발 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. |
카메라센서와 라이다 센서 융합에 기반한 개선된 주차공간 검출 시스템 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. |
로봇 핸드 제어를 위한 실시간 손 동작 영상 인식 시스템 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. |
Experiential boarding robot system using industrial robot Conference/Workshop 22.01.2019. |
속도 관측기를 이용한 토크 기반 로봇 제어에서의 위치 추종 성능 향상 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. |
비전 센서와 능동적 접촉 인지 제어 방법을 융합한 외부 물체의 접촉 인지 및 모서리 예측 전략 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. |
힘-토크 센서와 전신 동역학 모델을 융합한 보행 로봇의 접촉력 추정 Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea,, 22.01.2019. |
In-hand rolling motion generation using underactuated fingers with Four-bar linkage chains Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, 21.01.2019. |
Design Improvement of Compliant Anti-Gravity Rehabilitation Robot for Lumbar Pain Patients Conference/Workshop The 14th Korea Robotics Society Annual Conference, Pyeongchang ,Korea, 21.01.2019. |
System Design of Humanoid Robot DYROS-JET Conference/Workshop 2019 IEEE/SICE International Symposium on System Integration, Paris, France, 16.01.2019. |
2018 |
Autonomous parking algorithm using automatic selection of forward and backward switching points via model predictive control Conference/Workshop The Korean Society of Automotive Engineers Annual Autumn Conference, Kangwon, Korea, 14.11.2018. |
Development of Autonomous Valet Parking System and Simulation Result Conference/Workshop The Korean Society of Automotive Engineers Annual Autumn Conference, Kangwon , Korea, 14.11.2018. |
V-Rep Simulator for Autonomous Vehicle Research in Parking lot Environment Conference/Workshop The Korean Society of Automotive Engineers Annual Autumn Conference, Kangwon, Korea, 14.11.2018. |
Gripper with Thumb Adduction/Abduction Joint for Enhanced In-hand Orientation Manipulation Conference/Workshop 2018 IEEE-RAS International Conference on Humanoid Robots (Humanoids), Beijing, China, 06.11.2018. |
Contact-Consistent Disturbance Observer for Floating-base Robots Conference/Workshop 2018 International Symposium on Experimental Robotics, Buenos Aires, Nov 5-8, 06.11.2018. |
Humanoid fall recovery strategy using air thrust Conference/Workshop WorkShop on “Humanoid Robot Falling: Fall Detection, Damage Prevention, and Recovery Actions” at IROS 2018, Madrid, Spain, 01.10.2018. |
Effect of Heel Area on Utilized Coefficient of Friction During High-Heeled Walking Conference/Workshop The 20th Congress of the International Ergonomics Association (IEA 2018), Florence, Italy, 29.08.2018. |