|
Simulator Design Using a General Purpose PC and Off-The-Shelf Interface Boards for GNSS/INS Integrated Navigation System |
|
Jae Hoon Son, Yeong Hye Lee, Junwoo Jung, Yongtaek Hwang, Jiwoo Hwang, Minsu Kim(Chungnam National University, Korea), Gyeongmin Park(Chungbuk National University, Korea), JunMin Park, Hoyoung Yoo(Chungnam National University, Korea), Chansik Park(Chungbuk National University, Korea), Sang Heon Oh(Microinfinity Co., Ltd., Korea), Sang Jeong Lee, Dong-Hwan Hwang(Chungnam National University, Korea) |
| When a navigation systems are developed, lots of tests have to be performed. If a simulator is available which can generate signal and data of real environment, huge amount of time and cost can be reduced for development. In this paper, a simulator design method is proposed for GNSS/INS integrated navigation system using a general-purpose PC and OTS interface boards. In order to verify the proposed design method, a simulator of GPS L1 C/A and HG1700AG58 is implemented using a PC with intel i7-8700 CPU, USRP 2952-R, PCI-1780U-AE, FSCC/4-PCIe and LabVIEW2015. The implementation results show that GNSS signal and IMU data can be used for GNSS/INS integrated navigation system design. |
|
|
Normalizing In-orbit Anomalies and Retaining Stability Using Preferential-Kalman Filtering for Autonomous Small Satellite Constellations in CisLunar Space |
|
Mohammed Irfan Rashed(KAIST, Korea) |
| This paper presents the normalizing technique of on-orbit anomalies specific to small satellite constellations. This is to bring about the concerns that a satellites goes through for executing its functions and daily operations for a specific life span and self-correcting the anomalies and orbit errors as predicted by the assigned autonomous algorithm. |
|
|
Comparing Network RTK and Own RTK Base Station for Lateral Tracking in GNSS-Based Navigation: An Autonomous Golf Cart Study |
|
Surayuth Duenkwang, Chutimon Ketkarn, Sorravee Wanichphol, Benjamas Panomruttanarug(King Mongkut's University of Technology Thonburi, Thailand) |
| This study investigates the impact of utilizing a network RTK and our own RTK base station on the lateral tracking performance of a GNSS-based navigation system. The experiment employs an adaptive PID controller and compares the performance of both methods on an autonomous golf cart. The tracking path, composed of straight and curve segments, is manually generated based on the network RTK. The adaptive PID controller, with the gains tuned based on the slope of the path, is used to perform the tracking based on the real-time positioning observed from the network RTK as well as our own RTK base station. The results demonstrate that both RTKs achieve acceptable tracking performance. |
|
|
PanoNetVLAD: Visual Loop Closure Detection in Continuous Space Represented With Panoramic View Using Multiple Cameras |
|
Sungjae Shin, Yeeun Kim, Byeongho Yu, Eungchang Mason Lee, Dong-Uk Seo, Hyun Myung(KAIST, Korea) |
| In this paper, PanoNetVLAD deals with the problem of visual loop closure detection that occurs when there are limited camera field-of-views and effects of challenging environments. Our proposed method generates a panoramic image using multiple images and creates sub-panorama images for finding the closest query image. Then, a learning-based descriptor is formed for robust feature detection in challenging environments. Finally, find the index of the sub-panama image closest to the query image and bring the image with the same index from the raw database. To prove our method, we acquired the dataset in which visual loop detection was difficult and shared our datasets. |
|
|
Tight Integration of Inertial and Terrain-referenced Navigation under Biased Odometry and Measurements via Factor Graph Optimization |
|
Junwoo Park(KAIST, Korea), Hyochoong Bang(Korea Advanced Institute of Science and Technology, Korea) |
| A factor graph optimization (FGO)-based integration of the inertial navigation and terrain-referenced navigation that can cope with biased input is presented. Due to the slow dynamics of biases included in the inertial navigation system and sensor measurements, nodes designating such biases are sparsely instantiated. We demonstrate that fixed-lag smoothing of FGO, which optimizes a windowed bundle of factors, can perform as well as or better than the particle filter approach. The resultant solution becomes more accurate as the designated factor graph gets more interconnected. Multiple numerical experiments are performed to evaluate the proposed approach. |
|
|
BIG-STEP: Better Initialized State Estimator for Legged Robots with Fast and Robust Ground Segmentation |
|
Seunggyu Song(Kookmin University, Korea), Byeongho Yu, Minho Oh(KAIST, Korea), Hyun Myung(KI Robotics, Korea) |
| We propose a novel state estimator for legged robots that leverages ground plane to mitigate triangulation errors and for better performance. By exploiting the information from the estimated ground plane, which serves as a reliable reference, the proposed estimator refines camera features from input images onto the reference plane along its normal vector. To validate the effectiveness of the proposed state estimator, real-world experiments are conducted on a quadrupedal robot platform. The result demonstrates significant improvements in state estimation accuracy, especially in the z-component when compared with conventional state estimation methods. |
|