15 July 2025 in IEEE Transactions on Intelligent Transportation Systems
An accurate longitudinal dynamics model is essential for state estimation and control of autonomous vehicles. However, existing physical models suffer from limited working conditions and large errors, while pure data-driven models require massive amounts of driving data to cover working conditions fully. To address these issues, a hybrid architecture composed of a neural network-based traction model, a recursive least square-based parameter estimator, and a physics-based dynamics model is proposed for longitudinal dynamics modeling, in which the parameter estimator is used for mass and modified rolling friction coefficient estimation. Under this architecture, the longitudinal dynamics model can be established using limited driving data collected on a test field with a given load, and achieve precise vehicle dynamics characterization under various roads and loads. To design the neural network for traction description, the dynamics of vehicle powertrain and braking systems are analyzed, and a physics-guided neural network, which fully considers the traction transmission characteristics, is formulated. For model training, a two-stage hybrid model training method is proposed, which can train the hybrid model with the co-existence of unknown network and physical parameters. Results demonstrate that the proposed hybrid model can realize accurate parameter estimation and vehicle longitudinal dynamics modeling using limited driving data collected at a test field under a given load, especially with excellent generalization performance under different loads and roads.
24 April 2024 in IEEE Transactions on Intelligent Vehicles
OpenStreetMap (OSM) is a low-cost promising alternative to provide the environment representation with global consistency when the pre-built LiDAR maps are not available. However, existing descriptor-based perception-OSM matching methods suffered from significant disparities in dimension and precision compared to the perception and pre-built LiDAR map, leading to degradation of matching accuracy. To improve the accuracy and robustness of matching-based global position initialization using OSM, this paper proposes a novel boundary-relative orientation feature descriptor with rotational consistency, facilitating the unified representation of perception and OSM. The proposed scale-free descriptor, derived from the relative changes of boundary trend within the OSM, significantly reduces the reliance on the precision of planar spatial information, thereby promoting the matching accuracy required for global position initialization. Furthermore, the performance of global position initialization with the proposed descriptor is evaluated on KITTI datasets. The results illustrate the proposed descriptor outperforms 3D-3D and 2D-2D descriptor matching-based methods, especially in urban scenarios.
23 October 2025 in IEEE Transactions on Intelligent Transportation Systems
Roadside perception has a wider sensing range than onboard detection, providing enhanced sensory information for intelligent transportation systems, thus gaining increasing attention in recent years. However, directly applying onboard perception algorithms on roadside detectors (RSD) is infeasible due to the challenging requirements for high maneuvering target tracking and discriminating highly similar targets. Therefore, an Interactive Siamese Network (ISN) is proposed in this paper to overcome the roadside perception difficulties. Specifically, an interactive similarity contrast encoder-decoder has been developed within the ISN tracker. The sensitivity of algorithm to rapid changes in vehicle states is enhanced through the adaptive adjustment of weights assigned to critical tracking parameters within the loss function. This strategy enhances the ISN tracking effect for long-term trajectories for high maneuvering driving. Then, a global trajectory optimization unit is integrated into the ISN tracker. A trajectory similarity threshold is established to conduct cross-association analysis on similar trajectories, followed by iterative operations on adjacent trajectories. This approach further enhances the resolution of similar trajectories while ensuring the optimality of global trajectories. The proposed method is evaluated on the DAIR-V2X dataset and compared against current state-of-the-art methods. The experimental results verify that the proposed method provides efficient and accurate estimations and tracks vehicles in the intersection area, affording better accuracy and recall rate in high maneuvering target tracking than existing methods.
2 June 2025 in IEEE Robotics and Automation Letters
Visual odometry needs to be robust against various environmental changes. Although Deep learning (DL) based methods can bring more robust features to visual odometry (VO) than traditional methods, the gap between training and test dataset restricts the performance of DL-based methods when encountering novel scenes. Additionally, due to non-differentiability of visual odometry optimization process, the scene independent geometric constraints of images is hardly used in DL-based methods, which further decreases the accuracy and generalization of odometry. In this letter, we propose a reinforcement learning based framework incorporating the geometric constraints to address the challenge of non-differentiability. Unlike conventional DL-based methods, our VO estimates the pose through traditional direct odometry. To facilitate online training, a weighted direct graph structure is utilized to efficiently organize and simplify local images. This online training scheme enables the keypoint extraction policy to adapt dynamically to the current scene of odometry with less training data. Evaluations against the state-of-the-art visual odometry were performed on the KITTI Odometry, EuRoC MAV and Tum RGBD dataset. The results demonstrate that our method outperform popular approaches in terms of tracking accuracy and generalization ability.
25 December 2022 in Journal of Field Robotics
29 January 2026 in IEEE Internet of Things Journal
Global vehicle positioning through real-time LiDAR perception matched with online digital maps like OpenStreetMap (OSM) offers a promising solution to reduce reliance on high-quality maps in urban scenarios. However, the accuracy of global positioning based on LiDAR–OSM matching suffers from imprecise planar spatial information in OSM and frame-to-frame inconsistencies with the road network, which limits LiDAR–OSM matching to serving only as global position initialization rather than continuous positioning. To extend the application scenarios and performance of global positioning using OSM, a LiDAR–OSM matching framework is proposed to solve the planar association problem between real-time local perception and corresponding global position initialization sequences. Specifically, the motion consistency assumption is used to extract and reconstruct global candidate sequences that satisfy temporal and spatial continuity, which serves as the foundation for sequence-to-sequence matching with the local odometry. Furthermore, an online sequence-to-sequence transformation matrix estimation and verification mechanism is proposed to obtain a local-to-global solution that is subject to the activity range of vehicles and achieves optimal consistency with the road network. The performance of the proposed method is validated under different odometries and global initialization position confidence in the KITTI dataset. The results demonstrate that the proposed method achieves submeter level precision, showing a significant improvement over other LiDAR–OSM positioning methods in most urban sequences.