Yusheng Wang 교수

A novel method to estimate the missing dimension in 2D acoustic images for 3D reconstruction is proposed in this paper. Acoustic cameras can acquire high resolution 2D images in underwater environment insusceptible to water turbidity and light condition. However, the formulation of acoustic images leads to the missing dimension problem. Estimating the unknown elevation angle dimension is a difficult task which has recently drawn the attention of researchers. The non-bijective characteristic betw

In this article, a novel dense underwater 3-D mapping paradigm based on pose graph simultaneous localization and mapping (SLAM) using an acoustic camera mounted on a rotator is proposed. The demands of underwater tasks, such as unmanned construction using robots, are growing rapidly. In recent years, the acoustic camera, which is a state-of-the-art forward-looking imaging sonar, has been gradually applied in underwater exploration. However, distinctive imaging principles make it difficult to gai

We present a study on the effect of mounting position of IMU to a foot on the navigation accuracy of ZUPT-aided pedestrian inertial navigation. In this study, we compared two IMU mounting positions, forefoot and heel, in terms of the length of the stance phase, velocity uncertainty, and the shock level during the stance phase. Circular Error Probable (CEP) of localization was experimentally demonstrated to be reduced by about 50%, when mounting the IMU on the forefoot rather than behind the heel

In this paper, we present a novel probabilistic three-dimensional (3D) mapping framework that uses acoustic images captured in an underwater environment. Acoustic camera is a forward-looking imaging sonar that is commonly used in underwater inspection recently; however, the loss of elevation angle information makes it difficult to get a better understanding of underwater environment. To cope with this, we apply a probabilistic occupancy mapping framework with a novel inverse sensor model suitabl

ACMarker is an acoustic camera-based fiducial marker system designed for underwater environments. Optical camera-based fiducial marker systems have been widely used in computer vision and robotics applications such as augmented reality (AR), camera calibration, and robot navigation. However, in underwater environments, the performance of optical cameras is limited owing to water turbidity and illumination conditions. Acoustic cameras, which are forward-looking sonars, have been gradually applied

We present a method to identify and compensate systematic errors in the ZUPT-aided pedestrian inertial navigation. We considered two main categories of systematic errors resulting in an underestimate of the length of the trajectory and a drift in the heading of the trajectory. In this study, we identified the dominant factors resulting in the trajectory length and heading errors to be residual velocity during the stance phase and g-sensitivity error of the gyroscopes, respectively. Magnetic moti

In this paper, a three-dimensional (3D) environment reconstruction framework based on graph optimization is proposed that uses acoustic images captured in an underwater environment. Underwater tasks such as unmanned construction using robots are becoming more and more important. In recent years, acoustic cameras which are forward-looking imaging sonars are being commonly used in underwater inspection. However, the loss of elevation angle information makes it difficult to get a better understandi

We present a review of studies that were conducted by the MicroSystems Laboratory at UC Irvine on ZUPT-aided pedestrian inertial navigation. Our most recent results include: adaptive ZUPT detection, IMU mounting position optimization, residual velocity characterization, IMU error calibration, and navigation error prediction. With all the efforts above, a robust and accurate ZUPT-aided pedestrian inertial navigation implementation was demonstrated. The navigation bias was reduced by 10×, and a po

2D forward looking sonar (FLS) has gained widespread application in underwater robotics research, primarily due to its capacity to produce high-resolution images in diverse aquatic environments. This study deals with pose estimation with given 3D positions and corresponding 2D pixels, which is a fundamental problem for computer vision, denoted as acoustic-n-point (AnP) problem. It is the key part for object pose estimation, extrinsic calibration, localization, and structure from motion (SfM). We

2D forward-looking sonar is a crucial sensor for underwater robotic perception. A well-known problem in this field is estimating missing information in the elevation direction during sonar imaging. There are demands to estimate 3D information per image for 3D mapping and robot navigation during fly-through missions. Recent learning-based methods have demonstrated their strengths, but there are still drawbacks. Supervised learning methods have achieved high-quality results but may require further

In this paper, the concept of illuminated area, an important characteristic in acoustic images, is formalized, which can be applied to tasks such as the 3D mapping of underwater environment. Unmanned exploration using underwater robots is gaining attention among the scientific community. A way to sense the underwater environment is to employ the acoustic camera, a next generation forward looking sonar with high resolution even in turbid water. It is more flexible than common underwater sonars; h

In recent years, UAV (Unmanned Aerial Vehicles) are widely used in military reconnaissance and traffic control scenarios. Due to the rapid development of UAV technology, the object detection technology for aerial images has drawn the most attention in computer vision. However, detection in aerial images is a non-trivial task due to low target resolution, large scale variation, and occlusion and illumination variation. Our method is derived from the SSD (Single Shot Detector) which is well- balan

The purpose of this paper is to survey the computation performance of geometric algebra based program in comparison with the conventional analytic vector algebra method. This will be done by solving a classic inverse kinematics problem of a HAL robot arm model in geometric algebra versions (Gaigen 1, Gaigen 2, CLU and MapleTM) as well as in analytical vector algebra version (IKAN). Our research shows that, the with MapleTM optimized geometric algebra approach and Gaigen 2 based approach have the

In unstructured terrain, snakes can move more efficiently by actively pushing and avoiding obstacles, a capability believed to be realized through a decentralized mechanism. Moreover, considering harsh and unpredictable rescue environments where snake robots are expected to be deployed, such as post-earthquake scenarios, some joint sensors are prone to failure. We propose a decentralized adaptation mechanism for snake robots to navigate through random pole environments. This mechanism integrates

2D forward-looking sonar is a crucial sensor for underwater robotic perception. A well-known problem in this field is estimating missing information in the elevation direction during sonar imaging. There are demands to estimate 3D information per image for 3D mapping and robot navigation during fly-through missions. Recent learning-based methods have demonstrated their strengths, but there are still drawbacks. Supervised learning methods have achieved high-quality results but may require further