End-to-end autonomous driving systems tightly couple perception and decision-making through latent representations. Consequently, updates to perception models can alter these representations and degrade the performance of downstream policies that remain fixed. Existing solutions typically rely on policy retraining or architectural decoupling, both of which incur substantial computation and validation costs. In this paper, we formulate the model stitching problem for end-to-end autonomous driving and test the hypothesis that policy compatibility can be preserved through lightweight latent-space alignment. We study low-complexity model stitching methods, including linear and convolutional stitchers, for restoring compatibility between updated perception modules and frozen downstream policy modules. Experiments demonstrate that stitching effectively preserves downstream driving behavior under diverse perception updates, including changes in random initialization, sensor configuration, and training domain. In the most challenging cross-domain setting from nuScenes to CARLA, convolutional stitching retains over 91% of the no-shift driving score while reducing adaptation time from 22.18h to 0.91h. These results suggest that model stitching provides an effective and computationally efficient alternative to retraining or fine-tuning for maintaining end-to-end autonomous driving systems. The model will be open-sourced upon paper acceptance at https://github.com/SCP-CN-001/model-stitching to support further research and development in autonomous driving.

Autonomous driving simulation requires diverse and scalable lane-level HD maps to support long-horizon evaluation across complex road networks. Existing approaches either rely on handcrafted or reconstructed real-world maps, which limits scalability, or generate only local road structures rather than complete HD maps. We present RoadWeaver, a coarse-to-fine framework for from-scratch generation of diverse, large-scale HD maps. RoadWeaver first synthesizes a global road layout, expands it into a connected road network, and then constructs lane-level geometry with topologically consistent lane connectivity. Experimental results show that RoadWeaver achieves a 99.8% reachability, a 10.7% dead-end ratio, and an endpoint alignment error of 0.24 m. Compared with SOTA generation methods, it reduces endpoint alignment error by 94.4% while generating complete HD maps in 1.39–3.50 s. The generated maps can be directly deployed in driving simulators, providing scalable simulation environments for future closed-loop evaluation of autonomous driving systems. The training code and an out-of-the-box implementation of RoadWeaver will be released upon acceptance.

Whole-body tracking (WBT) models have become a key foundation for humanoid robots, enabling them to imitate diverse motions with high fidelity. Training such models from scratch requires large-scale data and computation, making rapid deployment on new humanoid platforms costly. This raises a natural question: Can pretrained WBT models transfer across embodiments with minimal adaptation? To answer this question, we propose Any2Any, a paradigm that efficiently transfers an existing WBT specialist to a new humanoid embodiment with only a small amount of data and compute. Any2Any first performs kinematic alignment between source and target humanoids, aligning their input and output spaces so that the pretrained source policy can be meaningfully reused on the target embodiment.Any2Any then performs dynamics adaptation by applying lightweight parameter-efficient fine-tuning (PEFT) components to selected dynamics-sensitive modules, preserving useful behavioral priors while enabling targeted adaptation to the target robot. Extensive experiments on multiple humanoid platforms and pretrained backbones show that Any2Any substantially accelerates convergence and reduces training cost compared with training from scratch, while achieving competitive or superior tracking performance. Notably, using only 1

Reinforcement learning (RL) in autonomous driving employs a trial-and-error mechanism, enhancing robustness in unpredictable environments. However, crafting effective reward functions remains challenging, as conventional approaches rely heavily on manual design and demonstrate limited efficacy in complex scenarios. To address this issue, this study introduces a responsibility-oriented reward function that explicitly incorporates traffic regulations into the RL framework. Specifically, we introduced a Traffic Regulation Knowledge Graph and leveraged Vision-Language Models alongside Retrieval-Augmented Generation techniques to automate reward assignment. This integration guides agents to adhere strictly to traffic laws, thus minimizing rule violations and optimizing decision-making performance in diverse driving conditions. Experimental validations demonstrate that the proposed methodology significantly improves the accuracy of assigning accident responsibilities and effectively reduces the agent's liability in traffic incidents.

Optical Character Recognition (OCR) is a critical research direction in computer vision. To address the inefficiencies and high error rates of manual inkjet code recognition, this study develops an end-to-end deep learning-based inkjet code recognition solution, aiming to improve accuracy while reducing model size for mobile applicability. First, the study summarizes domestic and international OCR research, analyzes mainstream algorithms, and introduces fundamental neural networks (CNN, RNN), including the structure of CNN layers and improvements to RNN limitations. Subsequently, targeting the shortcomings of large size and slow computation in traditional feature extraction networks, the study optimizes the CRNN+CTC model by proposing a lightweight OCR algorithm based on depthwise separable convolution, replacing conventional feature extractors with MobileNetV3 to reduce model size and enhance performance. Experimental results show that compared with the original CRNN model and three mainstream feature extraction algorithms, the proposed lightweight algorithm outperforms in average recognition time, accuracy, and model size. Tests on real inkjet datasets further confirm its robustness and high accuracy in recognizing blurred, deformed, or distorted characters.

Robust obstacle segmentation is essential for the safety of intelligent robots, where LiDAR-based perception systems play a fundamental role in the robot-environment interaction. While extensive LiDAR-based approaches have demonstrated high performance on common obstacles in urban scenarios, their results on tiny obstacles such as curbs, gravel, and potholes remain unsatisfactory due to the significant similarity between tiny obstacles and inherent road undulations. Moreover, their segmentation accuracy even deteriorates sharply when the LiDAR scans suffer from degradation in challenging off-road scenes. To overcome these bottlenecks, we propose LOGOS, a LiDAR-only unified tiny obstacle segmentation system, which models the road surface as a continuous mixture of 2D Gaussian primitives and distinguishes tiny obstacles via high-presicion elevation estimation. Unlike existing Gaussian splatting methods that rely on iterative RGB training, LOGOS is a backpropagation-free LiDAR-only approach. It directly estimates Gaussian parameters via a freespace-aware initialization by incrementally pruning non-road primitives using smoothness constraints. Subsequently, pointwise signed distances are computed via a novel normal-aware elevation splatting function, ensuring robustness to both flat and sloped terrains. We evaluate LOGOS on a highly heterogeneous benchmark of point cloud frames collected from urban mobility scenarios and mining haulage off-road environments. These data are practically acquired using different LiDAR sensors and exhibit large variations in point density, terrain roughness, and obstacle types. Experiments on the road and off-road scenes demonstrate that LOGOS significantly outperforms other state-of-the-art methods, particularly in degraded point cloud regions and challenging off-road scenarios, while maintaining real-time efficiency.

Game-theoretic models provide principled frameworks for modeling vehicle interactions, but their underlying temporal assumptions have not been systematically examined against real-world driving behavior. In particular, it remains unclear how simultaneous, sequential, and asymmetric interaction structures can be measured from vehicle trajectories. This paper develops a trajectory-based interaction measurement framework to identify interaction events and quantify behavioral change onset, temporal organization, post-onset response dynamics, and ordering stability. The framework uses behavioral deviations to verify candidate interactions. We evaluate the framework on six real-world trajectory datasets, including INTERACTION, highD, inD, rounD, Waymo Open Motion, and nuPlan, covering diverse road geometries, traffic environments, and interaction types. The results show that concurrent and sequential behavioral changes both constitute substantial proportions of observed following, merging, and conflicting interactions. Among sequential interactions, stable ordering is more prevalent than alternating ordering, indicating that persistent asymmetric roles are a common interaction structure. Importantly, temporal precedence does not necessarily coincide with a measurable behavioral response, indicating that temporal ordering alone may not be sufficient to characterize behavioral dependence. These findings show that real-world interactions exhibit concurrent, sequential, and persistently ordered temporal structures. Different game-theoretic formulations are therefore better regarded as complementary modeling abstractions for different interaction regimes rather than as a universal structure governing all vehicle interactions.

Traffic signal control at urban intersections inherently introduces stop-and-go behavior, resulting in increased delays and reduced traffic efficiency, especially under high traffic demand. With the emergence of connected and automated vehicles (CAVs), trajectory-level coordination has emerged as a high-potential strategy to augment or transcend conventional phase-based management. This paper proposes DSIP (Diffusion-model-based Signal-free Intersection Planner), a multi-agent motion planning framework driven by a generative diffusion process. DSIP shifts the intersection management paradigm from discrete temporal phasing to continuous multi-vehicle trajectory optimization. This work evaluates the theoretical upper-bound performance of this coordination strategy under idealized communication and execution conditions to isolate the core benefits of the diffusion-driven approach. Using the SUMO platform, we evaluate DSIP across diverse four-leg intersection configurations. Experimental results demonstrate that DSIP significantly reduces average delay and maintains higher average speed compared to both fixed-time signal control and state-of-the-art reinforcement-learning-based controllers, particularly in medium- to high-density traffic. These findings suggest that diffusion-based trajectory planning provides a scalable and robust foundation for future autonomous intersection management. By unlocking latent intersection capacity through software-defined coordination, this approach offers a cost-effective pathway to improve urban traffic flow efficiency without requiring physical infrastructure expansion.

Learning-based 3D reconstruction has emerged as a transformative technique in autonomous driving, enabling precise modeling of environments through advanced neural representations. It has inspired pioneering solutions for vital tasks in autonomous driving, such as dense mapping and closed-loop simulation, as well as comprehensive scene feature for driving scene understanding and reasoning. Given the rapid growth in related research, this survey provides a comprehensive review of both technical evolutions and practical applications in autonomous driving. We begin with an introduction to the preliminaries of learning-based 3D reconstruction to provide a solid technical background foundation, then progress to a rigorous, multi-dimensional examination of cutting-edge methodologies, systematically organized according to the distinctive technical requirements and fundamental challenges of autonomous driving. Through analyzing and summarizing development trends and cutting-edge research, we identify existing technical challenges, along with insufficient disclosure of on-board validation and safety verification details in the current literature, and ultimately suggest potential directions to guide future studies.

The Automated Valet Parking system (AVP), as one of the promising technologies, offers significant benefits in saving maneuver time and parking cost. With the increase number of vehicles using AVP system in large-scale parking lots, the overall efficiency is limited by the selfish decision-making of each individual; therefore, cooperative-AVP (C-AVP) is developed to achieve global optimization by scheduling the vehicles. Existing C-AVP methods focus only on one of the key processes in AVP, such as space allocation, trajectory planning, and electric vehicle(EV) charging. However, these factors interact with each other and are rarely considered in an entire system. Therefore, this paper aims to model the space allocation, trajectory planning, and EV charging problem in a whole framework, named APC-scheduler, and adopt it to improve the overall efficiency for large-scale parking lot with EV. The APC-scheduler comprises two parts: 1) a zone-based parking space allocation module dynamically assigns parking spaces based on real-time conditions using a hierarchical optimization strategy. It contains a charging priority estimation process to determine the EV’s charging requirement urgency; 2) a conflict-based trajectory planning module is developed to reduce trajectory overlaps and vehicle conflicts. It uses spatial-temporal path planner and speed planner to eliminate conflicts. Experiments under reasonable vehicle arrival and departure statistics have been conducted in a large-scale parking lot with 526 parking spaces. The results demonstrate that the proposed method effectively enhances parking efficiency and minimizes conflicts, particularly under high vehicle arrival frequencies and in dense traffic conditions.

In recent years, autonomous parking has made significant advances, yet parking tasks still face challenges in extreme scenarios such as mechanical and dead-end parking slots, often resulting in failures. This is mainly due to traditional parking methods adopting a multistage approach, lacking the ability to optimize the parking problem as a whole. End-to-end methods enable joint optimization across perception and planning modules to eliminate the accumulation of errors, enhancing algorithm performance in extreme scenarios. Although several end-to-end parking methods use imitation or reinforcement learning, the former is limited by data cost and distribution coverage, while the latter suffers from inefficient exploration. To address these challenges, we propose a Reinforcement learning End-to-end Autonomous Parking method (REAP). REAP employs Soft Actor-Critic (SAC) within an asymmetric reinforcement learning framework to improve training efficiency and inference performance. To accelerate model convergence, we distill the capabilities of a rule-based planner into the end-to-end network through behavior cloning. We further introduce a soft predictive collision penalty mechanism to reduce collision rates by penalizing obstacle-approaching actions. To ensure that the trained reinforcement learning network can directly transfer to real-world scenarios, we have established a Real2Sim2Real simulator. In the Real2Sim step, we use 3D Gaussian Splatting (3DGS) to transform real-world scenes into digital scenes. In the Sim2Real step, we deploy the end-to-end model onto the vehicle to bridge the Sim2Real gap. Trained in the 3DGS simulator and deployed on physical vehicles, REAP successfully parks in various types of parking spaces, especially demonstrating the feasibility of end-to-end RL parking in extremely narrow mechanical slots.

Roadside multi-sensor networks, such as RGB-D camera networks, play a crucial role in Intelligent Transportation Systems (ITS). These systems rely on accurate extrinsic parameters (i.e., the relative positions and orientations) of each camera in the network. However, achieving fast and accurate large-scale extrinsic parameter calibration is challenging, especially when the overlap between camera views is limited due to cost constraints. To address this issue, we propose an automated, scalable, and marker-free calibration method that requires no human intervention. Our approach leverages dynamic rigid bodies, such as moving vehicles, as bridges to establish associations among all cameras without manual placement or supervision. The proposed method consists of three main stages: 1) calibrating camera height, roll, and pitch using the ground plane; 2) estimating each camera’s 2D position and yaw angle based on trajectories; and 3) refining these estimates by matching features on the road surface using SuperPoint and SuperGlue. Real-world experiments involving 58 roadside RGB-D cameras deployed in a parking lot demonstrate that our method significantly improves calibration efficiency while maintaining high accuracy, making it well-suited for large-scale RGB-D camera network deployment.

In recent years, with the development of autonomous driving, 3D reconstruction for unbounded large-scale scenes has attracted researchers’ attention. Existing methods have achieved outstanding reconstruction accuracy in autonomous driving scenes, but most of them lack the ability to edit scenes. Although some methods have the capability to edit scenarios, they are highly dependent on manually annotated 3D bounding boxes, leading to their poor scalability. To address the issues, we introduce a new Gaussian representation, called DrivingEditor, which decouples the scene into two parts and handles them by separate branches to individually model the dynamic foreground objects and the static background during the training process. By proposing a framework for decoupled modeling of scenarios, we can achieve accurate editing of any dynamic target, such as dynamic objects removal, adding and etc, meanwhile improving the reconstruction quality of autonomous driving scenes especially the dynamic foreground objects, without resorting to 3D bounding boxes. Extensive experiments on Waymo Open Dataset and KITTI benchmarks demonstrate the performance in 3D reconstruction for both dynamic and static scenes. Besides, we conduct extra experiments on unstructured large-scale scenarios, which can more convincingly demonstrate the performance and robustness of our proposed model when rendering the unstructured scenes. Our code is available at https://github.com/WangXu-xxx/DrivingEditor.

This paper proposes a model pruning named MPSS for semantic image segmentation networks, so that semantic image segmentation models can be deployed into embedded devices. Most existing model pruning methods aim at image classification models. Since semantic segmentation is a fine grained task, the directly use of traditional model pruning methods greatly reduces the model accuracy. The core problems of model pruning in the semantic segmentation task are to determine the appropriate pruning kernels and the pruning structure. We propose a new composite index that defines the similarity between convolution kernels to determine the pruning kernels. Furthermore, we propose a structure mending method based on the neural architecture search to determine the pruning structure. Compared with the method of manually defining the pruning rate, the proposed structure mending method obtains a better pruning structure. We conduct experiments based on two semantic segmentation networks, the FCN and the FASSD-Net. The experimental results show that the proposed model pruning method enables the pruned network to obtain higher accuracy under the same compression rate. In addition, we deploy the compressed models on an embedded platform, and the FASSD Net inference speed is twice as fast as the unpruned model on NVIDIA Xavier NX.

Complete and accurate environment 3D perception is a cornerstone of autonomous driving, and multi-vehicle cooperative perception offers a compelling solution to overcome occlusions and achieve comprehensive environmental awareness by leveraging wireless communication among neighboring agents. However, practical deployment necessitates a careful trade-off between communication efficiency and perceptual accuracy. In this work, we propose an instance-level intermediate fusion framework that achieves performance comparable to dense fusion methods while reducing communication overhead by 100 ×. Our method introduces an alignment module for spatial and temporal domain equipped with learnable position encodings to mitigate misalignments induced by inherent transmission delays and heterogeneous vehicle poses. Furthermore, we decouple the decoding process into anchor-based reference point generation and offset prediction, thereby overcoming the limited perception range constraints in existing intermediate fusion approaches. Evaluated on V2X-Real dataset, our method achieves comparable results with 54.2 mAP (IOU0.3) and 46.7 mAP (IOU0.5), demonstrating robustness under real-world conditions, including time delay and communication constraints, thereby establishing a new paradigm for efficient and scalable cooperative perception.

Intelligent Transportation Systems (ITS) require reliable environmental perception to support safe and efficient transportation. With the rapid development of Vehicle-to-everything (V2X), roadside perception has become an effective means to extend sensing coverage and improve traffic safety. However, the scarcity of large-scale annotated roadside LiDAR datasets poses a major challenge for training high-performance roadside perception models. In this paper, we introduce Vehicle-to-Roadside LiDAR Synthesis (VRS), a data synthesis framework that generates labeled roadside LiDAR datasets from vehicle-side datasets via LiDAR novel view synthesis. To mitigate the vehicle-to-roadside domain gap, VRS employs vehicle point cloud completion to compensate for missing geometry in vehicle-side observations, and introduces an occupancy-based visibility constraint to handle large viewpoint changes during cross-view rendering. The proposed framework enables flexible multi-view rendering for scalable roadside data generation. Extensive experiments on roadside 3D object detection demonstrate that the synthesized data effectively complements real roadside data, mitigates the limitations of limited real-world roadside data, and improves generalization to unseen roadside viewpoints.

3D reconstruction is a critical technology with significant implications for applications such as urban planning, autonomous driving, and virtual reality. Recent advances in 3D Gaussian Splatting (3DGS) have demonstrated impressive results in small-scale scenes, achieving high-quality reconstructions with real-time rendering capabilities. However, when applied to large-scale scenes, existing 3DGS methods face significant challenges due to the exponential growth of model size, often exceeding the memory capacity of consumer-grade GPUs and making training and rendering infeasible. In this paper, we propose a structure-guided memory-efficient 3DGS framework that uses only half the memory of current large-scale 3DGS methods while maintaining state-of-the-art reconstruction accuracy. Specifically, we introduce a structure-guided density control mechanism that uses a heuristic approach to split Gaussian ellipsoids in challenging regions and optimizes their attributes during densification, significantly reducing memory storage requirements while preserving structural details with fewer ellipsoids. Moreover, we propose a novel structure loss to supervise the learning of scene structural information, enabling the model to better capture and preserve geometric details such as straight lines and edges, further enhancing reconstruction accuracy. We also propose the largest known drone dataset for 3D reconstruction, comprising over 10,000 high-resolution images covering more than 2.5 million square meters. Extensive experiments on multiple benchmark datasets and our proposed dataset demonstrate that our new method is highly memory-efficient with high accuracy. We strongly recommend you to watch our demo at https://lvzinan.github.io/STGS.github.io/.

Most existing Uncrewed Aerial Vehicle (UAV) and Uncrewed Ground Vehicle (UGV) cooperative systems adopt a UAV-centric paradigm, in which the UAV carries onboard sensors to search for, localize, and land on the UGV. This UAV-centric architecture increases onboard payload and reduces endurance, and makes it difficult to reliably assess the safety of the entire landing airspace in unstructured environments. To address these limitations, we propose a ground-centric UAV-UGV cooperative system, wherein a UGV-mounted LiDAR serves as the primary perception unit for launch and recovery. The perception architecture operates in two modes, namely an overhead airspace assessment mode in which the UGV evaluates the overhead airspace for UAV landing and relocates to an open area to ensure landing safety, and a tracking mode that supports guided recovery of the returning UAV. This ground-centric design introduces a coordinate alignment challenge between heterogeneous platforms, which is addressed through an online calibration based on a short flight maneuver that aligns the UAV frame with the UGV frame. Real-world experiments validate the complete autonomous launch and recovery process and demonstrate robust system performance with centimeter-level landing accuracy.