Rongjun Ding

Rongjun Ding

Auto-generated profile, based on publicly available information

Rongjun Ding

State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body, College of Mechanical and Vehicle Engineering, Hunan University, Changsha, China

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

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

Representative Publications

View All Publications
8 May 2023 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
LZLiqin Zhang
Liqin Zhang
MHManjiang Hu
Manjiang Hu
FZFeifei Zhang
Feifei Zhang
YBYougang Bian
Yougang Bian
Anh-Tu NguyenAnh-Tu Nguyen
Anh-Tu Nguyen
RDRongjun Ding
Rongjun Ding
In this paper, a channel-level event-triggered communication scheme is investigated for path tracking control of autonomous ground vehicles via an in-vehicle network to ensure path tracking performance while reducing network resource utilization. First, a path tracking system is established considering the effect of norm bounded uncertainty on lateral vehicle dynamics, and a state feedback path tracking controller is designed within the Takagi–Sugeno fuzzy framework. Next, a decentralized event-triggered communication scheme is adopted between the sensors and controller, which makes it possible to save communication resources at the channel level. Finally, the results in the CarSim-Simulink joint-simulation environment show that the proposed controller can ensure a reliable path tracking performance in both lateral offset and heading error. At the same time, compared with the simple time-triggered scheme, the proposed method can effectively reduce the usage of communication resources by more than 80%.

22 May 2023 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
YBYougang Bian
Yougang Bian
RLRuotian Li
Ruotian Li
GWGuangcai Wang
Guangcai Wang
XQXiaohui Qin
Xiaohui Qin
MHManjiang Hu
Manjiang Hu
RDRongjun Ding
Rongjun Ding
In order to improve the positioning accuracy and reliability of the navigation system in complex underwater environment, a strapdown inertial navigation system (SINS)/doppler velocity log (DVL)/Ultra-short baseline (USBL) tightly-coupled algorithm is proposed based on a centralized Kalman filter. A new 26-dimensional tightly-coupled measurement model is constructed. Different from the traditional loosely-coupled algorithm, the proposed model directly integrates the original measurement information of each sensor, instead of the navigation information such as position and velocity. Simulation and experiment results show that the proposed algorithm can make the best use of the only original information when the sensor cannot output the correct position and velocity in case of poor measurement conditions. Compared with the loosely-coupled algorithm, the proposed algorithm has better positioning accuracy and higher reliability.

11 May 2022 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
JYJun Ye
Jun Ye
YBYougang Bian
Yougang Bian
BLBiao Luo
Biao Luo
MHManjiang Hu
Manjiang Hu
BXBiao Xu
Biao Xu
RDRongjun Ding
Rongjun Ding
In this article, an adaptive dynamic programming (ADP) scheme utilizing a costate function is proposed for optimal control of unknown discrete-time nonlinear systems. The state-action data are obtained by interacting with the environment under the iterative scheme without any model information. In contrast with the traditional ADP scheme, the collected data in the proposed algorithm are generated with different policies, which improves data utilization in the learning process. In order to approximate the cost function more accurately and to achieve a better policy improvement direction in the case of insufficient data, a separate costate network is introduced to approximate the costate function under the actor-critic framework, and the costate is utilized as supplement information to estimate the cost function more precisely. Furthermore, convergence properties of the proposed algorithm are analyzed to demonstrate that the costate function plays a positive role in the convergence process of the cost function based on the alternate iteration mode of the costate function and cost function under a mild assumption. The uniformly ultimately bounded (UUB) property of all the variables is proven by using the Lyapunov approach. Finally, two numerical examples are presented to demonstrate the effectiveness and computation efficiency of the proposed method.

28 June 2023 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
DCDengxiang Chang
Dengxiang Chang
RZRunbang Zhang
Runbang Zhang
SHShengjie Huang
Shengjie Huang
MHManjiang Hu
Manjiang Hu
RDRongjun Ding
Rongjun Ding
XQXiaohui Qin
Xiaohui Qin
Multi-LiDAR systems provide all-round perception of the surrounding environment from different directions simultaneously, significantly enhancing the sensing range and density of measurements. Accurate extrinsic calibration is a key prerequisite for data fusion of multi-LiDAR systems. However, existing multi-LiDAR calibration methods mostly rely on calibration targets, priori environmental information, and initial calibration values, resulting in complex calibration processes and poor adaptability to environments. To this end, this letter proposes a targetless multi-LiDAR automatic calibration method based on global pose graph optimization, which does not rely on field-of-view (FoV) overlaps and can be applied to multiple LiDARs. First, LiDAR odometry is employed to estimate poses of each LiDAR, and relative motion constraints are used to recover initial extrinsic parameters. Then, absolute poses of each LiDAR are estimated by scan-map matching and used as absolute motion constraints for global pose graph to refine calibration parameters, which not only satisfies the observability of the 6 degree-of-freedom (DoF) calibration parameter space, but also avoids the influence of pose estimation error on the calibration accuracy. In addition, global optimization improves the global consistency of multi-LiDAR systems compared to separate pairwise parameter estimation. The proposed multi-LiDAR calibration method has been tested on various datasets. The results of performance evaluation and ablation studies have verified the reliability, accuracy and versatility of our method.

10 April 2024 in IEEE Transactions on Intelligent Transportation Systems
Verified
Journal Article
Closed access
XWXiaowei Wang
Xiaowei Wang
PJPeiwen Jiang
Peiwen Jiang
YLYang Li
Yang Li
MHManjiang Hu
Manjiang Hu
MGMing Gao
Ming Gao
RDRongjun Ding
Rongjun Ding
Well-trained visual object detectors are generally confronted with a severe performance decline when deployed in a novel driving scenario due to the impact of domain shift. Despite excellent improvements in unsupervised domain adaptive object detection achieved by adversarial training, those approaches fail to capture the transfer core underlying the holistic scenes. To solve this problem, we propose a progressive critical region transfer framework for cross-domain visual object detection. Specifically, we exploit a potential foreground mining (PFM) module and a semantic-specific RoI aggregation (SRA) module to improve the robustness of the cross-domain detection framework. Upon the critical regions in the broad sense, the PFM module first highlights the foreground regions by reweighting the hierarchical feature maps in sequence, and then modifies location biases at the downstream position of the backbone network for more accurate upstream predictions. Deep into the critical regions in the narrow sense, the SRA module concentrates on establishing an appropriate matching between batch-wise RoIs and all semantic centers, and further strengthens the aggregation of cross-domain identical semantic with the complement of context references. Together these modules are obligated to transform the adaptation importance from the whole scope to the latent foreground areas, and afterward to the informative regions of interest along the detection pipeline. Experiments show that our progressive critical region transfer framework achieves a state-of-the-art performance in adverse weather, camera configuration, and complicated scene adaptation, which outperforms the baselines by 19.4%, 5.0%, and 6.1%, respectively.

21 December 2023 in IEEE Sensors Journal
Verified
Journal Article
Closed access
DCDengxiang Chang
Dengxiang Chang
SHShengjie Huang
Shengjie Huang
YZYunshui Zhou
Yunshui Zhou
XQXiaohui Qin
Xiaohui Qin
RDRongjun Ding
Rongjun Ding
MHManjiang Hu
Manjiang Hu
Camera and GNSS/IMU are widely used in autonomous vehicles and robots due to their light weight and low cost, in these cases, an accurate extrinsic calibration is essential for sensor fusion. Traditionally, specific vehicle movements or scenarios with known fiducial markers are required for the sophisticated calibration process, however, the former is difficult to eliminate the impact of drift made by the techniques used to estimate the motion of each sensor, and the latter is poorly automated and needs specialist human expertise. To tackle above problems, this paper proposes a target-free stereo camera-GNSS/IMU self-calibration method based on iterative refinement. The initial calibration parameters are calculated with relative pose constraints derived from visual odometry. To eliminate the impact of visual odometry drift, parameters are further refined by iteratively using paired relative pose constraints and absolute pose constraints, which are derived from scan-global map matching. Quantitative calibration results on simulated data show the robustness and accuracy of this approach with RMSEs of 10 −3 deg and 10 −3 m for rotation and translation, respectively. Further experiments in real world confirm the superior of our method in natural environments.