publications
2024
-
ConferenceParallel performance benchmark of large sparse matrix using SuperLU_dist and PaScaL_TDMAChanyoung Ahn*, Oh-Kyoung Kwon*, and Ji-Hoon Kang*CDE Aug 2024Sparse matrices find numerous scientific and engineering applications, including simulations, optimizations, and large-scale data analysis. Characterized by a significant number of zero elements, sparse matrix solvers often encounter challenges in computational efficiency due to memory access patterns, becoming a computational bottleneck in many applications. We provide a comparative performance benchmark of two parallel solvers, SuperLU_dist and PaScaL_TDMA, for solving large sparse matrices. While SuperLU_dist, a widely recognized parallel solver developed under the Exascale Computing Project in the U.S., offers versatility and robustness across various sparse matrix structures, PaScaL_TDMA exhibits superior performance for tridiagonal matrices due to its tailored algorithmic design. These findings emphasize the importance of selecting optimized solvers based on the specific characteristics of the matrix structure in scientific and engineering computations. By leveraging the strengths of specialized solvers like PaScaL_TDMA, significant computational efficiencies can be achieved over general solvers, thereby accelerating the solution of large-scale problems in practical applications.
-
ThesisA Reinforcement Learning Testbed for Deformable Object Manipulation using Visuotactile SensingChanyoung Ahn*, and Daehyung ParkKAIST Jan 2024We aim to create a simulated testbed for training and assessing deformable object manipulation skills. This testbed requires tactile sensing to distinguish heterogeneous elasticities of a deformable object to acquire skills. In this work, we introduce a visuotactile testbed, DetactGym, for deformable object manipulation, integrating a novel architecture of tactile sensors leveraging collision cascades. These sensors integrate a diamond-shaped rigid element encased within an external rigid structure. This design overcomes a fundamental limitation of PhysX engine-based simulators: their inability to directly measure contact force on deformable objects. The diamond-shaped element efficiently transmits the contact force to the outer structure during interactions with these objects, ensuring force measurements. Our evaluation focuses on the testbed’s capability, equipped with tactile sensors, to facilitate the learning of lifting heterogeneous deformable objects with minimal deformation through reinforcement learning methods. Our findings highlight the effectiveness of tactile feedback over visual cues in manipulating deformable objects with diverse elasticities, reducing deformation.
2022
-
ConferenceA Telemanipulation Suite for Deformable Object ManipulationBonggyeong Park*, Chanyoung Ahn*, and Daehyung ParkKAIA Jan 2022Our aim is to create a human-expert manipulation dataset of deformable objects toward learning deformable manipulation. However, the data collection is challenging due to the limited sensing capability against to the high number of degree of freedom and complicate deformation/contact of soft bodies. Further, the self occlusion restricts the observation in grasping. In this work, we introduce a haptic-telemanipulation suite by adopting a haptic glove with a state-of-the-art physics simulator, IsaacGym from NVIDIA. The suite enables users to obtain realistic visual-and-tactile feedback as well as collect any part of object states while teleoperating a robotic gripper. We evaluate the suite by building one hundred demonstrations of dataset given five deformable objects.
2020
-
ThesisChanyoung Ahn*, and Haedo JeongPusan National University Dec 2020PNU CAPSTONE Poster Session, 3rd out of 90+ teams Weight reduction of a satellite structure, which is one of the heaviest parts of a cube satellite, is discussed in this paper. The structural strength should be considered when attempting to reduce the structure weight. This paper presents the method for reducing the weight of Cube satellite structure, based on ANSYS analysis and engineer’s intuition for optimal design. In this study, we proposed a structural design concept to reduces cost and improves mechanical durability by using a lightweight technique. The effectiveness of the design has been verified through structural analysis and vibration tests. Numerical simulation results showed that the Margin of Safety of the 12U structure was over 1.4 under the worst-case conditions. Our work allows 34% of lightweight by satisfying the requirement.
2019
-
ConferenceDesign and Validation of Low Noise Airfoil Inspired by Flight of OwlMinsin Kim*, Chanyoung Ahn*, Youjin Im*, YoonSeoung Park*, and Kyung Chun KimKSME Nov 2019Compared to other birds, owls fly achieving remarkably low noise even that human cannot hear. It is well known that the silent flight owes to both huge size of owl’s wing and microscopic morphologies of feathers, which make up wings, composed of leading edge serration, trailing edge fringes, and velvet surface. To validate the noise reduction effect by applying owl’s characteristics to an airfoil model NACA0018. Three models: (1) without and (2) with leading edge serrations and trailing edge fringes and (3) with serrations and feathers were prepared for comparison. Using particle image velocimetry (PIV) and a blow-down low speed wind-tunnel, flow characteristics of boundary layer and wake are measured with the three different models at two angle of attacks. It is found that the hairy wing with serration prevents flow separation followed by noise reduction.