Qingqing Li

Indoor parking lots represent a critical scenario for the commercialization of intelligent vehicles. The absence of GNSS signals, coupled with complex environmental structures and frequent dynamic interference, makes high-precision localization a core and challenging technology for achieving autonomous parking in such settings. Existing research methods based on multi-sensor fusion often lack strong constraints from high-level behavioral decisions when addressing these challenges, leading to insufficient localization accuracy and robustness in complex dynamic environments. To address this, this paper proposes an innovative approach that deeply integrates the planned path for autonomous parking with multi-sensor data to achieve high-precision localization in underground parking lots. Firstly, the A* and DWA path planning algorithms are improved to generate smoother trajectories that better adapt to dynamic environments. Secondly, in a novel contribution, the real-time optimized velocity and steering angle information fed back from path planning is utilized as key observations. This information is deeply fused with LiDAR and IMU data and embedded into a Bundle Adjustment (BA) optimization framework. This integration compensates for the accuracy degradation associated with velocity estimation relying solely on IMU pre-integration in traditional methods. Experimental results demonstrate the excellent performance of the proposed method across multiple real indoor parking lots. The average localization accuracy stably reaches around 0.2 meters, and notably, achieves 0.2315 meters even in complex commercial underground parking environments, confirming its stability and reliability in complex and dynamic scenarios.

Pavement crack detection is widely spread over road maintenance, ensuring the longevity and safety of infrastructure. Traditional manual inspection methods are time-consuming, labor-intensive, and prone to errors. In response, automated crack detection systems based on deep learning have emerged, offering more efficient and accurate solutions. However, existing models often face challenges such as large model sizes, slow inference speeds, and limited applicability in real-time applications. In this paper, we propose a novel light-weight Crack Regional Segmentation method based on YOLOv11, which introduces attention mechanisms to address challenges in pavement images, such as varying crack sizes, occlusion, and irregular surface textures. By embedding a region-based attention mechanism into the YOLOv11 network, the method enhances the model’s ability to focus on crack features. Specifically, the model network layers are progressively pruned to reduce the number of parameters and floating-point operations, thereby further improving operational efficiency and refining detection in the target regions. Furthermore, to tackle issues with blurred or indistinct crack boundaries, we present a fuzzy information-guided YOLOv11-based model, FIG-YOLO. This model integrates fuzzy logic and fuzzy membership functions to handle uncertainty in crack detection. The fuzzy membership functions are used to quantify the degree of crack features, allowing the model to better distinguish between crack and non-crack regions, especially in cases where crack boundaries are unclear. This approach significantly improves the accuracy of crack detection and segmentation. Extensive experiments demonstrate that our approach effectively addresses challenges such as complex backgrounds and blurred crack edges in pavement images. This research not only provides a novel solution for the automated detection of pavement cracks but also offers insights into the development of intelligent road maintenance systems. With the expansion of large-scale datasets and the advancement of deep learning models, pavement crack detection algorithms are expected to further enhance their accuracy and efficiency, offering significant support for road infrastructure management.