Ju Huo

Dr. Ju Huo


Research Focus & Metrics

Academic Career Length

21Years (2006-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 July 2023 in IEEE Sensors Journal
Verified
Journal Article
Closed access
HZHanxuan Zhang
Hanxuan Zhang
DWDingyi Wang
Dingyi Wang
JHJu Huo
Ju Huo
Most existing dynamic simultaneous localization and mapping (SLAM) systems simply remove dynamic objects, resulting in the loss of dynamic object motion information that helps the system itself localize and navigate. Aiming at the above problems, a visual-inertial dynamic object tracking SLAM tightly coupled system is proposed. First, a method for classifying dynamic and static feature points is proposed to solve the problem that dynamic objects cannot be extracted by learning-based or vision-only methods. Second, the system accurately extracts and tracks dynamic rigid-body objects in the scene through a combination of visual information, inertial measurement unit (IMU) data, and associated matching point labels, without modeling the shape or geometry of the dynamic objects. Subsequently, a novel pipeline is proposed to jointly optimize the camera poses, trajectories of dynamic objects, the maps of the environment, and IMU information, improving the self-localization performance of the system. Finally, the performance of the system is tested on indoor and outdoor datasets and real dynamic scenes. Experimental results show that the proposed system achieves improved performance in self-localization and dynamic object tracking compared to other state-of-the-art methods. Especially under adverse conditions of partial occlusion and far field of view, the system can still track dynamic objects robustly and accurately to optimize self-localization performance.

27 February 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
HZHanxuan Zhang
Hanxuan Zhang
JHJu Huo
Ju Huo
YHYulong Huang
Yulong Huang
QLQi Liu
Qi Liu
To mitigate the heavy reliance on semantic information and the unreliability of manual feature extraction in dynamic simultaneous localization and mapping(SLAM) and object tracking systems, a novel visual-inertial SLAM with deep feature extraction matching is proposed. A lightweight network is designed for feature extraction and description, replacing the ORB approach, which can address deep learning latency and the shortcomings of manual extraction. A fast object tracking method is developed for extracting and associating dynamic objects based on clustering analyses of feature point pair distances to epipolar lines, which enables real-time tracking of numerous objects without semantic data. An advanced on-line incremental loop closure detector for deep features is designed, which supersedes ORB-based detectors and maintains global pose optimization. The system’s effectiveness and advantages, including its proficiency in real-time embedded platform execution and enhanced self localization and dynamic object tracking, have been demonstrated through extensive evaluations. Notably, the system is capable of consistently tracking dynamic objects with rapid movement or weak texture, furnishing the localization system with robust dynamic constraints.

9 October 2024 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
HZHanxuan Zhang
Hanxuan Zhang
JHJu Huo
Ju Huo
YHYulong Huang
Yulong Huang
JCJiajun Cheng
Jiajun Cheng
XLXiaofeng Li
Xiaofeng Li
The use of unmanned aerial vehicles (UAVs) for Internet of Things applications, like intelligent monitoring and search, is increasingly becoming a popular research focus globally. While various optimization algorithms exist to plan UAV flight paths, they frequently compromise the quality of the planning path to decrease planning time. In view of the above problems, a perception-aware-based UAV trajectory planner via generative adversarial self-imitation learning from demonstration is proposed. First, a progressively growing discriminator is devised to prevent the policy network from being overpowered in early training stages, avoiding potential training failures. Second, the issue of homogenized strategic patterns among optimized expert trajectories is solved by incorporating successful trajectories from the policy network into the expert buffer, which thereby enhances the network’s generalization capabilities. Third, to address the challenges of skewed distribution and considerable performance variation among the strategies learned by the policy network during training, a class-level instance-balancing expert buffer is introduced. Finally, the yaw angle of the UAV in real time during flight is obtained by using the analytical solution of the position trajectory and yaw angle and the position trajectory output from the policy network. Experiments confirm our proposed method achieves comparable flight costs and success rates to those of the reference expert method, while the planning time is reduced. The proposed method is also shown to be well adapted to dynamic environments and obstacle trajectories, which are not involved in training. Additionally, the ablation studies highlight the individual contributions of each component within the proposed method.

21 February 2024 in IEEE Sensors Journal
Verified
Journal Article
Closed access
HZHanxuan Zhang
Hanxuan Zhang
DWDingyi Wang
Dingyi Wang
JHJu Huo
Ju Huo
In dynamic environments, most simultaneous localization and mapping (SLAM) methods face challenges due to potential erroneous data associations caused by the movement of dynamic objects and the possibility of incorrect loop closures when objects move out of the field of view. To address this issue, we proposed RDP-SLAM, a real-time dynamic SLAM method based on the integration of inertial measurement unit (IMU) and segmentation. In this approach, a novel model was developed for calculating the moving probability of feature points using IMU data, allowing for the elimination of feature points associated with dynamic objects. Additionally, a feature point moving probability propagation model was constructed, which combines IMU data and segmentation results. This model enabled the tracking and classification of feature points into three groups: dynamic, static, and potentially-dynamic. In the local and global optimization processes, a weighted objective optimization function was designed specifically for the moving probability of static feature points, thereby enhancing the accuracy of the optimization process. The experimental results conducted on public datasets and real-world scenarios demonstrated that RDP-SLAM was effective in robustly selecting static feature points and improving localization accuracy, particularly achieving a 75% reduction in absolute pose error compared to the ORB-SLAM3 system in complex dynamic environments.

1 February 2022 in Journal of Systems Engineering and Electronics
Verified
Journal Article
Closed access
ZZZijian Zhu
Zijian Zhu
WXWenhao Xiang
Wenhao Xiang
JHJu Huo
Ju Huo
MYMing Yang
Ming Yang
GZGuiyang Zhang
Guiyang Zhang
LWLiang Wei
Liang Wei
For localisation of unknown non-cooperative targets in space, the existence of interference points causes inaccuracy of pose estimation while utilizing point cloud registration. To address this issue, this paper proposes a new iterative closest point (ICP) algorithm combined with distributed weights to intensify the dependability and robustness of the non-cooperative target localisation. As interference points in space have not yet been extensively studied, we classify them into two broad categories, far interference points and near interference points. For the former, the statistical outlier elimination algorithm is employed. For the latter, the Gaussian distributed weights, simultaneously valuing with the variation of the Euclidean distance from each point to the centroid, are commingled to the traditional ICP algorithm. In each iteration, the weight matrix $W$ in connection with the overall localisation is obtained, and the singular value decomposition is adopted to accomplish high-precision estimation of the target pose. Finally, the experiments are implemented by shooting the satellite model and setting the position of interference points. The outcomes suggest that the proposed algorithm can effectively suppress interference points and enhance the accuracy of non-cooperative target pose estimation. When the interference point number reaches about 700, the average error of angle is superior to 0.88°.

7 August 2024 in IEEE Transactions on Aerospace and Electronic Systems
Verified
Journal Article
Closed access
HZHanxuan Zhang
Hanxuan Zhang
JHJu Huo
Ju Huo
YHYulong Huang
Yulong Huang
DWDingyi Wang
Dingyi Wang
We offer a perception-aware quadrotor translation and yaw angle planner approach to solve the navigation problem for a quadrotor with limited field of view (FOV) in unpredictable dynamic situations. Our method is developed based on the static obstacle avoidance method Ego-Planner. Firstly, we establish an environmental perception module based on a depth sensor. This module categorizes obstacles into static and dynamic types, tracking and predicting the trajectories of dynamic obstacles. Secondly, we introduce the minimum volume enclosing polynomial curves (MINVO) basis instead of B-spline basis to solve the problem of conservative convex hull generated by existing planning methods based on B-spline basis. To enable dynamic obstacle avoidance, we design a method that combines modified kinematic path searching with gradient optimization, avoiding the need for maintaining a euclidean signed distance field (ESDF) that adds computational burden in existing approaches. Finally, we jointly optimize translational motion, yaw angle rotation, and the visibility cost of tracking obstacles in the FOV to maximize the UAV's ability to observe unknown obstacles early and evade them promptly. Simulation experiments demonstrate that in unpredictable dynamic situations, our proposed planner can effectively avoid obstacles, achieving a high success rate of up to 93%.