Browse Topic: Optics

Items (10,614)

Parametric Screening of Hybrid Powertrain Architectures for High-Energy Laser Vehicles: Application to Foxhound and Boxer

2026-01-7552

9/22/2026

The US Army and several NATO allies have committed funds for directed energy weapons including high energy lasers (HEL), which require substantial electrical power. Silent Mobility and Silent Watch requirements mandate hybridization, which also provides the electrical capacity needed for a HEL. This paper introduces a HEL framework of classes A-E based on target types and fluence physics. It estimates installed HEL mass and power demand by class, and applies rapid powertrain screening to hypothetical hybrid variants of the UK vehicles Foxhound and Boxer. Results show that HELs up to Class B (60 kW) and D (300 kW) laser output can be supported with minimal powertrain modifications by Foxhound and Boxer respectively, and upgrades to support Class C (150 kW) and E (500 kW) are containable within the payload of each vehicle. A rapid methodology is presented to determine what powertrain architecture is needed to support a given HEL.

Salis, Rupert Tull

Multimodal Digital-Twin Simulations for Off-Road Autonomous Vehicles

2026-01-7550

9/22/2026

Research in autonomous driving has largely focused on structured, on-road environments in densely populated urban areas, leaving off-road autonomy comparatively underexplored despite its importance for applications such as search and rescue, agriculture, and defense. Progress in developing solutions for off-road autonomy is heavily constrained by the high cost, safety risks, and limited coverage of real-world off-road data collection, particularly for rare terrain-induced failure cases. To address these challenges, we introduce a large-scale simulation platform for off-road autonomous driving generated in the Unity3D engine. Our simulator provides indefinite data from multiple RGB cameras, LiDAR, GPS, and IMU sensors to support research in long-duration applications. Our platform also facilitates research in risk-aware planning through dynamic weather, lighting, and annotations for sequence-level failures like collisions and weather-induced sensor blockages. By combining long-duration temporal coverage, multimodal sensing, and procedurally generated terrain diversity, our simulator facilitates progress for systematic evaluation of learning-based autonomous driving models in unstructured, safety-critical environments.

Ross, Timothy, Boone, Julia, Afghah, Fatemeh

Clemson Off-Road Dataset (CORD): A Multimodal Benchmark for Autonomy beyond the Flat-World

2026-01-7518

9/22/2026

While autonomous perception has matured within the structured confines of urban roadways, it remains brittle when confronting the chaotic, non-rigid terrain of the natural world. This paper introduces the Clemson Off-Road Dataset, a high-fidelity, multimodal dataset engineered to bridge this gap by challenging standard “flat-world” assumptions. Featuring 2.90 TB of sensor data, the dataset captures a diverse spectrum of unstructured environments, ranging from the transitional trails of CU-ICAR and the day/night lighting dynamics of TN3 to the unstructured wilderness of Camp Daniels and the novel coastal scenery of Edisto Island. Distinguishing itself from existing forest-centric benchmarks, the Clemson Dataset provides a first-of-its-kind focus on coastal data, featuring unique adversarial conditions such as extreme solar glare, loose sand, and shifting tide lines. The data is collected aboard a Polaris RZR Pro R 4, a high-performance platform integrated with a sensor suite designed to perceive physics beyond geometry. Alongside 360° HD camera coverage, 3D LiDAR, and Radar, we integrate Cubert Ultris Hyperspectral imaging and Prophesee EVK4 Event-based vision to enable material classification and high-dynamic-range motion tracking. To overcome the bottleneck in ground truth generation, we used our “AI LabelMate,” a context-aware semi-automated annotation agent that fuses Vision-Language Models (Florence-2) with SAM2 to generate 6331 pixel-perfect annotated frames using a specialized off-road ontology and a human-in-the-loop pipeline. We establish performance baselines using Oneformer for semantic segmentation and used SalsaNext for lidar point clouds labelling. Available in both raw ROS2 bag and extracted standard formats, this Dataset serves as a pivotal testing ground for the next generation of robust autonomous systems.The dataset of this paper is available upon request to the Virtual Prototyping of Autonomy-Enabled Ground Systems (VIPR-GS) Center.

Patil, Ashish, Gupta, Prakhar, Bhosale, Mayuresh, Mukwaya, Arthur, Jegede, Akinbobola, Mikulski, Dariusz, Mwakalonge, Judith, Jia, Yunyi

Bullfrog: AI Powered Counter Unmanned Aerial System

2026-01-7519

9/22/2026

The proliferation of small unmanned aircraft systems (sUAS) presents an asymmetric threat to ground maneuver forces operating in contested and gray-zone environments. The Bullfrog Autonomous Weapon Station (AWS) addresses this operational gap through a passive, AI-powered counter-UAS system employing computer vision and machine learning for autonomous detection, tracking, classification, and engagement. Field testing at Technology Readiness Experimentation (T-REX) 26-1 demonstrated 100% probability of defeat against Group 1 UAS targets with a mean engagement time of 6 seconds and 10 rounds per kill at ranges exceeding 160 meters. Operating in both autonomous and human-in-the-loop modes, Bullfrog achieved 99.45% operational availability while leveraging service-common M240B weapons and Modular Open Systems Architecture for rapid integration with Joint All-Domain Command and Control (JADC2) networks. At $300,000 per unit with $10 cost-per-engagement, Bullfrog demonstrates operational relevance, speed-to-field, and alignment with Army and Marine Corps autonomy priorities.

Cunningham, Jason, Clark, Alex

Development of a UAV Platform for Remote Inspection and Visualization

2026-01-7511

9/22/2026

This paper details the development of an intelligence and inspection platform consisting of an attritable sub-250g UAV, a ground control station, and a visualization interface for users. The UAV architecture combines onboard obstacle detection and avoidance along with simultaneous localization and mapping to have full autonomous navigation inside of complicated GPS-denied environments. The ROS 2-to-Unreal Engine data pipeline allows for sensor fusion, data cleansing, and initial analysis as well as creation of a high-fidelity real-time 3D digital twin. The visualization interface allows users to easily identify critical features and turn data into intelligence to support decision making by soldiers and first responders.

Lee, Yeen K., Bainard, Sean, Shaughnessy, Michael, Bolger, Matt, Koepp, R. Tucker, Salehzadeh, Roya, Mallory, Stephen, Mynderse, James A., Guillen, Pedro, Hernandez, Margarita

View Translation: Geometry-Guided Latent Diffusion for Cross-View Image Synthesis on Ground Vehicles

2026-01-7501

9/22/2026

Synthesizing novel camera views is important for autonomous ground vehicles, with applications in surround-view monitoring, occlusion recovery, and training data augmentation. We present View Translation, a geometry-guided latent diffusion framework that generates a target camera view from a source image, relative camera pose, and an available target-view depth prior. The method combines three components: a Vector Quantized Variational Autoencoder for compact latent encoding, a depth-based warping module that projects the source image into the target view to provide geometric guidance, and a ControlNet-augmented denoising UNet conditioned on source appearance, relative pose, and an auxiliary Image-Depth fusion network. Evaluated on KITTI and a simulated off-road dataset, our method achieves competitive FID while improving LPIPS and PSNR over baseline approaches, supporting cross-view synthesis for ground vehicle perception.

Mayekar, Omkar, Aiyetigbo, Mary, Salvi, Ameya, Samak, Tanmay, Samak, Chinmay, Desjardins, Brendan, Smereka, Jonathon, Brudnak, Mark, Krovi, Venkat, Luo, Feng, Li, Nianyi

Scene-Level Multi-View Instance Consistency Using Multi-Feature Vision Fusion for Synthetic Ground Vehicle Environments

2026-01-7502

9/22/2026

Maintaining consistent object identities across multiple camera viewpoints is a critical challenge in synthetic perception environments used for autonomous ground vehicle evaluation. This paper presents a scene-level multi-view instance consistency framework that integrates OpenUSD scene composition, Omniverse Replicator synthetic-data generation, and a multi-feature vision fusion pipeline. The proposed approach combines semantic embeddings from CLIP, patch-level descriptors from DINOv2, geometric correspondences from LoFTR, mask-derived shape invariants using Hu moments, and relative-position priors to associate object instances across views, including visually identical objects. A compact composite scoring function fuses these complementary cues to achieve robust cross-view identity assignment while preserving OpenUSD asset modularity through grouped-prim support. Synthetic experiments across 120 multi-camera scenes demonstrate improved Top-1 Match Accuracy and Identity Consistency Rate, with reduced ID-switch occurrences compared to single-cue baselines. The framework supports scalable, repeatable, and traceable digital engineering workflows for defense-oriented perception evaluation.

Bhattacharya, Sambit, Nakamoto, Kyle

An Assessment of Biomanufacturing in Forward Deployed Locations

2026-01-7557

9/22/2026

Biomanufacturing uses microorganisms to produce chemicals or materials of interest, much like a brewery uses fermentation by yeast to produce the alcohol in beer. Biomanufacturing relies upon synthetic biology to reprogram yeast or other microorganisms to produce something of greater value, such as fuel, food, or pharmaceuticals. Industrial biomanufacturing has made significant advances and the products it can deliver include reactive coatings and textiles, sensors, optical materials that can bend light, and new therapeutics such as antimicrobials and vaccines. The convergence of synthetic biology, robotics, and artificial intelligence is opening the way to produce materials never before possible in the commercial market. These same technologies create the opportunity for the miniaturization of this technology to fit into ever more compact spaces, bringing forward deployment of these mini-factories closer and closer to the point of need.

Ahern, Brooke, Crumbley, Annie, Walker, Anne, Grodecki, Joseph

Robust Fiber Optics for Ground Combat Vehicles: Unlocking the Bandwidth for the Mission

2026-01-7541

9/22/2026

The U.S. Army’s Modular Open Systems Approach (MOSA) is driving data-centric vehicle architectures that demand higher bandwidth, faster decision loops, and greater cross-platform interoperability. Despite these needs, stakeholders hesitate to adopt fiber optics because of perceived fragility, field-retrofit FOD risk, and soldier-handling concerns. This paper characterizes common fiber failure modes in ground-vehicle environments and demonstrates how system-level ruggedization, connector design, and qualified components mitigate those risks. Drawing on demonstrated experience with hardened optics, connector-integrated transceivers, sealed media converters, and rugged cabling, we summarize practical architecture and field maintenance procedures that enable reliable fiber deployment in vehicles. Results show that, with appropriate component selection, installation practices, and built-in diagnostics, fiber optics can provide a robust, maintainable backbone for future ground vehicle networks.

Kouns, Heath, Simms, Marc, Weiss, Evan, Breedlove, Joe

Change Detection Filtering with Visual Language Models

2026-01-7508

9/22/2026

Intelligence, surveillance and reconnaissance (ISR) often require review of significant quantities of video. While machine vision is used to flag objects for human review, too many flags are generated. Integrating newer methods like Open-Vocabulary Object Detection (OVOD) that support zero shot detection, but with significantly lower accuracy only make the problem worse. This work addresses the utility of OVOD in ISR missions by focusing only what has changed between successive runs through an environment. A Vision Language Model (VLM) compares current observations against a registered “cleared” baseline to focus only on what has changed. Testing across three distinct environments, and using either monocular camera phones or RGB-D equipped vehicles, demonstrates that integrating change detection can automatically remove as much as 80% of unchanged objects without impacting recall.

Martinson, Eric, Fishta, Igri

Real-Time Road Intelligence through a Hazard Authentication Model

2026-24-0027

9/21/2026

Advanced Driver Assistance Systems (ADAS) are evolving beyond onboard perception. The ability to dynamically map and share temporary road hazards is important for connected and autonomous driving, but authenticating the data is critical. An in-vehicle hazard recognition layer performs real-time video analysis and geotagging on embedded platforms. Real-time video is analyzed for road hazards—such as potholes, construction zones, waterlogging, fallen trees, roadside accidents, and debris—using onboard cameras and lightweight computer vision models. This metadata is sent to a cloud-based aggregation layer to validate hazard reports. It is of paramount importance to validate hazard reports originating from diverse sources, regardless of their accuracy. This paper presents a mathematical approach to determine an overall Hazard Confidence Score (HCS) based on data received from diverse sources. A unique hazard authentication model is introduced to quantify the credibility of each hazard report using six validation metrics.

Bose, Souvik

Ultrasonic Frequency-Domain Reverse-Time Migration Imaging Incorporating Element Directivity Correction

2026-99-1419

9/11/2026

The finite width of ultrasonic array elements results in a non-uniform angular radiation pattern of elastic waves in solids, deviating from the ideal point-source assumption commonly adopted in reverse-time migration (RTM). This angular radiation non-uniformity produces a crack tip-dominated imaging amplitude with weak crack flank representation, manifesting as a pronounced depth-dependent amplitude imbalance along vertically oriented defects. As a result, cracks may be misinterpreted as point reflectors, which compromises the reliability of characterization in ultrasonic nondestructive testing (NDT). This study proposes an ultrasonic frequency-domain reverse-time migration (FDRTM) imaging method incorporating element directivity correction. A longitudinal-wave directivity model in solids is formulated in the frequency domain and normalized at each frequency to ensure consistent scaling during multi-frequency stacking. During backward wavefield reconstruction using full matrix capture (FMC) data, the angular-dependent energy distribution associated with the receiving direction is explicitly corrected, rebalancing the angular energy distribution in the reconstructed wavefield. Defect imaging is then performed using a frequency-domain cross-correlation imaging condition, followed by stacking over frequency and normalization. Validation experiments were conducted on an artificially manufactured vertical crack in a 7075 aluminum alloy specimen. The results indicate that, relative to conventional RTM, the proposed method reduces crack tip dominance and enhances the relative visibility and continuity of crack flanks. Compared with conventional RTM, the peak imaging amplitude increases by 22.44%, and the amplitude at a depth of 6.8 mm is enhanced by 24.39%. In addition, the proposed method outperforms the total focusing method (TFM) in crack profile continuity and the relative visibility of crack flanks. The results confirm that incorporating element directivity correction into frequency-domain RTM mitigates depth-dependent amplitude imbalance and restores crack flank visibility, thereby improving the reliability of crack defect characterization in ultrasonic NDT.

Chen, Si, Zhang, Yifeng, Ma, Tengfei, Xu, Zheng, Jiang, Jiansheng, Gao, Jiaqi

An Improved Multi-Object Vehicle Detection Algorithm Based on YOLOv11s

2026-99-1129

9/11/2026

For object detection in complex road situations, such as inadequate detection performance and difficulties caused by vehicle occlusion and cluttered environments, this paper pursues a YOLOv11s-based object detection framework. The algorithm successfully designed a novel PEConv module. This module integrates a partial convolutional network with an efficient multi-head attention mechanism. Through a Split operation, the input image is divided into locally enhanced channels and original channels. The locally enhanced channels undergo partial convolution and feature weight allocation via the efficient multi- head attention mechanism for feature extraction. Finally, these channels are fused with the original channels before undergoing convolution. This approach preserves the original features while minimising feature loss caused by the series of operations. Therefore, the PEConv module is based on a partially convolutional network and efficient multi-head attention. It improves the detection ability by precisely giving more weight to small objects and occluded parts with augmented partial channel attention and original channel fusion. This study further enhances the model’s detection precision and improves its performance in addressing small target vehicles and severe occlusion issues by refining and upgrading the original C3K2 architecture. The LSBlock is integrated into the original model’s bottleneck structure, replacing the traditional 3x3 convolution to create the C3K2 - LSBlock module. Experimental results show that on the UA - DETRAC dataset, compared with the original YOLOv11s, the optimized YOLOv11s has improved the original mAP @ 50 by 3.4%, reaching 61.3%, and improved the original mAP @ 50: 95 by 2%, which verifies the correctness of it.

Chen, Yulin, Wang, Yini, Wang, Jianwei, Zhang, Xin

Analysis of Traffic Capacity at the Intersection of Xiyou Road and Qianshan Road in Hefei City

2026-99-1551

9/11/2026

With the advancement of urbanization and the popularization of automobiles, the traffic load on urban roads is becoming increasingly heavy, resulting in many traffic problems. Road intersections serve as crucial linchpins in the urban transportation grid, wielding considerable influence over the overall traffic capacity of a city’s road network. Enhancing intersection efficiency and cutting down on delays stand at the heart of tackling urban congestion challenges. This study zeroes in on the crossroads where Xiyou Road intersects with Qianshan Road in Hefei City. Employing hands-on observation and photographic documentation, the research examines traffic flow and signal configurations during the peak demand period (7:30-8:30). The analysis evaluates traffic capacity and utilization rates for through, left-turn, and right-turn lanes at this intersection. Findings reveal that the right-turn lane at the southern entrance and the left-turn lanes at both northern and eastern entries show relatively low saturation levels, while the saturation of other lanes is greater than or close to 1. Therefore, this intersection does not have sufficient capacity. The actual traffic operation at the intersection, particularly during peak traffic times, is analyzed to identify the reasons for congestion Finally, improvement plans for optimizing traffic organization at intersections are proposed, such as optimizing signal timing schemes and transforming traffic channelization. Simulation analysis using VISSIM shows a 9.34% reduction in total intersection parking time, a 34.26% decrease in average queue length, and an 8.12% reduction in average vehicle delay. These results provide a reference for future optimization work, including intersection signal timing and channelization.

Wang, Yanmei, Wang, Chen, Fu, Ziyue, Meng, Xianglong

Nc5-PCS Based Heterogeneous Point Cloud Matching for Roads

2026-99-1533

9/11/2026

Point cloud registration represents a fundamental task in geospatial informatics and 3D computer vision, aiming to align heterogeneous point clouds through rigid transformation estimation. While Super-4PCS serves as an efficient coarse registration method, it exhibits limitations when handling large-scale datasets, planar-distributed point clouds, and scenarios with unknown scale differences. To overcome these challenges, this paper proposes the Nc-5PCS (Neighborhood-constrained 5-Point Congruent Sets) algorithm. Nc-5PCS first performs approximate scale estimation through concavity-convexity similarity analysis within coarse overlap regions, addressing the inherent scale limitation in 4PCS-based approaches. Subsequently, the algorithm employs 3D Harris feature point extraction to significantly reduce data volume while preserving critical geometric characteristics. The core innovation lies in designing a non-coplanar 5-point basis with a corresponding hash-based retrieval mechanism, effectively resolving the feature degradation problem caused by coplanar 4-point bases. Furthermore, normal vector angular constraints are incorporated to enhance consensus evaluation during correspondence selection, substantially improving registration accuracy. Experimental validation demonstrates that Nc-5PCS achieves a point-to-point RMS error of ≤ 0.227 m, outperforming Super-4PCS to provide superior initial alignment for subsequent ICP refinement.

Liu, Lei, Yu, Keguang, Li, Xinyi, Sun, Guangde, Zhao, Xinyuan, Zhu, Dongni, Fan, Yabo, Guo, Shihao

Site Selection Method for Solid Waste Processing Plant of Highway Reconstruction and Expansion Project

2026-99-1586

9/11/2026

The highway reconstruction and expansion project is accompanied by the generation of a large amount of construction solid waste. The unreasonable site selection of solid waste processing plants will increase the social, environmental, and economic burden. Taking a highway reconstruction and expansion project in Guangdong Province as an example, this study uses the combination of the analytic hierarchy process and the layer superposition method to extract the influencing factors of site selection, such as geological conditions, natural conditions, hydrological conditions, traffic conditions, and resource conditions, according to relevant specifications, and uses the analytic hierarchy process to quantify each influencing factor. From the relevant research data, official public information, and other channels, we comprehensively collected the data of topography, climate, geology, land use planning, and other aspects of Guangzhou and Dongguan along the project. With the help of buffer analysis tools and overlay analysis tools of GIS software, the optimal decision results were determined. The research results show that using this method to analyze the site selection of the relying project, the factory site selection should be located in Wangniudun Town near the project line, which has comprehensive advantages. The site selection method of a solid waste processing plant for an expressway reconstruction and expansion project proposed in this paper comprehensively considers the influence of 10 sub-factors on the site selection, and has been successfully applied to the site selection decision of an expressway reconstruction and expansion project in Guangdong Province. The final site selection result is more professional and objective than the previous site selection method, which effectively solves the site selection problem of a solid waste processing plant under the influence of economic factors, social factors, municipal factors, and environmental factors.

Zhang, Yuping, Yang, Ming, Zeng, Siqing, Long, Hao, Liu, Yuanqing, Zhao, Qiu

3D Dynamic Object Detection with Multi-Modal Temporal Fusion

2026-99-1117

9/11/2026

This research aims to address the critical challenge of accurately detecting and estimating the state of dynamic objects in autonomous driving. Traditional 3D object detection methods often struggle with motion perception, particularly in velocity estimation, due to the lack of information in single frame perception. We propose a novel framework that enhances the BEV representation with temporal modeling. The core of our method is a two-stage temporal fusion process. First, we align historical BEV features to the current coordinate frame to eliminate the interference of ego-motion. Subsequently, a dedicated temporal fusion encoder, architected with residual connections and a Feature Pyramid Network, refines the aligned multi-frame BEV features to capture complex motion patterns and improve multi-scale object representation. This approach directly tackles the problem of motion decoupling. By aligning features, we disentangle object motion from ego-motion. The temporal fusion encoder then mitigates the positional ambiguity of moving objects in the fused BEV space, a common issue in simple feature concatenation, leading to more robust detection. We built a dataset following the structure of the nuScenes dataset, using data collected from an autonomous driving simulation platform. The evaluation results on our simulation dataset demonstrate that the proposed temporal module achieves a 13.0% improvement in NDS score and a substantial 29.7% reduction in velocity error (mAVE). These results demonstrate that our temporal fusion strategy effectively enhances 3D detection accuracy in dynamic scenarios.

Shao, Mengjia, Li, Wei, Bai, Jie, Zhu, Shaoxiong, Xu, Chenjie

DisReID: Single Domain Generalizable Person Re-ID Method for Highway Construction & Operation

2026-99-1530

9/11/2026

With the advancement of computer vision technologies and the widespread deployment of video surveillance systems, traffic safety and the development of intelligent highways have been significantly enhanced. As a key component of the intelligent video analysis module in smart highways, person re-identification (re-ID) addresses critical challenges, including cross-segment tracking of pedestrians illegally using emergency lanes, multi-camera joint searches for lost persons in service areas, and trajectory tracing of individuals involved in traffic accidents. These functions directly support the core goals of "safety assurance and efficient service" for smart highways. However, due to the complexity of the application scene, its generalization to unseen environments remains a core challenge. This problem is formally studied under the setting of Single-Domain Generalizable Person Re-identification (SDG re-ID), which aims to train a model on a single source domain that can perform well on arbitrary unseen target domains. To handle this issue, this paper proposes a novel Disentangled Augmentation re-ID Framework (DisReID) that disentangles and augments both structure and style. Specifically, DisReID consists of two modules: Structure-aware Viewpoint Simulation (SVS), a novel pre-processing technique that simulates cross-camera perspective changes by perspective transformation, diversifying geometric structure without harming identity semantics; and Style-Dominant Frequency Perturbation (SFP), which selectively focuses on the style-dominant frequencies and applies perturbation to enable controllable style augmentation while preserving structure cues. Furthermore, to alleviate the BN-induced domain bias, we introduce a simple yet effective test-time adaptation strategy, termed Cluster Fine-tuning (CF), that performs unsupervised clustering on target-domain features to assign pseudo-labels and subsequently fine-tunes the model, enhancing adaptability to unseen domains. Extensive experimental results on four public datasets demonstrate that our DisReID achieves superior generalization performance compared to the state-of-the-art methods. This work provides key technical support for the large-scale application of re-ID in smart highways, advancing the goal of "full-domain perception and intelligent collaboration".

Pan, Hong, Yu, Fangying

A Maritime Occluded Ship Detection and Tracking Method Based on Improved YOLOv8 Fused with AIS Trajectory

2026-99-1549

9/11/2026

Typical maritime monitoring scenarios are usually constrained by factors such as multi-scale ship density, frequent motion overlap, and limited viewing angle of shore-based cameras. These challenges often lead to trajectory interruptions and identity mismatches in target detection and multi-target tracking tasks. In order to solve these problems, this study proposes a ship occlusion detection and tracking method based on the improved YOLOv8 model and further integrates an automatic identification system (AIS) trajectory reasoning. The method builds a unified perception framework with enhanced detection architecture, multi-source data fusion, and behavioral reasoning capabilities. First, in the target detection module, the improved SEConv structure is introduced into the YOLOv8 trunk network to address challenges caused by small-scale variations and severe occlusion in maritime scenes. The ReLU activation function in SEConv is replaced by the Swish activation function to enhance the nonlinear feature representation. In addition, the optimized SEConv is embedded in the C2f structure, and the convolutional block attention module (CBAM) attention mechanism is introduced to enhance the sensitivity of the model to the occlusion area. Next, for multi-target tracking, ByteTrack is used as the basic tracking framework. AIS trajectory data is introduced as auxiliary input to compensate for trajectory losses caused by occlusion. Finally, experimental results on the SeaShips public dataset and the self-built occlusion reference dataset show that the improved YOLOv8 detector achieves stable mAP gains in mild, moderate, and severe occlusion scenarios. The AIS enhanced tracking system improves the multi-target tracking accuracy (MOTA) and identification F1 score (IDF1) by about 6.3% and 8.1%, respectively, and the average occlusion reconstruction error is controlled within 1.4 seconds. The proposed method effectively enhances the perception ability of ships in complex occlusion environments and verifies the feasibility and superiority of the strategy of combining visual detection with AIS data assistance.

Guan, Keping, Chen, Miao, Zhou, Yue

Molecular Dynamics Simulations of Microstructural Evolution in Pure Titanium under Laser Shock Peening

2026-99-1285

9/4/2026

Impacts of laser shock peening (LSP) on the evolution characteristics of microstructure in commercially pure α-phase titanium (α-Ti) are explored by molecular dynamics (MD) simulations of high strain-rate compression. The EAM potential (Zhou potential) is selected for its ability to capture the evolution of microstructures. Considering the LSP-induced peak plasma pressure, the strain rate during the simulated shock compression process is set at 10^9 s-1 to replicate the LSP process. The stress-strain curve of the α-Ti under high strain-rate compression is obtained. The maximum equivalent stress reaches 3.6 GPa, consistent with the theoretically calculated value. The simulation results reveal that mechanical twins (MTs) are activated at a strain of 3%. The number of mechanical twins increases and eventually stabilizes, forming a network structure throughout the grains. In the meantime, numerous partial dislocations are generated adjacent to the grain boundaries. The dislocation density also increases with strain and dislocation reactions occur. Moreover, grain refinement is identified. The grain size is refined from the initial ~ 8 nm to ~ 4 nm in the polycrystalline α-Ti. Twinning, together with dislocation-mediated plasticity, drives the refinement of grain size. Gradients of twin density, dislocation density, and grain size density are induced by LSP on the surface of α-Ti. This study comprehensively investigates how LSP influences the evolution of microstructures by MD simulations. It develops an innovative numerical strategy that offers a foundation for elucidating the underlying mechanisms of LSP.

Zhao, Congshan, Zhang, Linbing, Xu, Yidi, He, Jianye, Fang, Jing, Li, Zezhou, Ruestes, Carlos J., Cheng, Xingwang

A Coaxiality Measurement Method Applicable to the Three Holes of an Engine

2026-99-1226

9/4/2026

To address the measurement challenges posed by the large diameters and wide spans of engine crankshaft holes, cylinder holes, and camshaft holes, as well as the abnormal laser measurement data caused by oil film adhesion and local protrusion structures on their inner walls, this paper proposes a non-contact coaxiality measurement method based on a laser displacement sensor (LDS) and a threshold-modified least squares ellipse fitting (TMLSE) algorithm. The method first preprocesses the data through median filtering and then employs an adaptive distance threshold based on maximum inter-class variance to eliminate outliers caused by oil film scattering and protrusions, achieving robust fitting of the cross-sectional centers. Subsequently, the datum axis is established using the least squares midline method, ultimately enabling the evaluation of coaxiality error. In the simulation experiments, a dual-hole model with an inner diameter of 100 mm and a spacing of 700 mm was simulated, where the right hole was translated by 1.5 mm to introduce coaxiality error. Each cross-sectional point cloud included 10% protrusion points (protrusion height of 10 mm) along with random noise and data loss simulating the effects of oil film. To assess the robustness of the method, a systematic analysis was conducted on the influence of laser incidence angles [80°,89°]on measurement accuracy. The results show that across different incidence angles, the maximum deviation between the coaxiality error obtained by the TMLSE method and the theoretical value is only 0.1582 mm, which is significantly better than that of the RANSAC, MZC, and MIC methods. Meanwhile, TMLSE demonstrates stable and efficient computational speed. The simulation verifies that this method offers higher fitting accuracy and robustness under both optical interference and variations in installation angle, making it suitable for the precision measurement of key engine hole systems.

Yin, Jiakuo

Enhancing Laser Processing Head Service and Quality Assurance: How Advanced Laser Diagnostics Transform Service Operations and Product Development for High-Power Laser Applications

26AERP09_02

9/1/2026

In the rapidly evolving arena of high-power laser technology, precision and reliability are becoming increasingly important. The increasing powers in laser applications into the 10's of kW and, in some cases, 100's of kW, present additional challenges when applied to internal laser system components. For Precitec, respected globally for their advanced laser processing heads, delivering consistent performance to customers is both a challenge and a necessity. This case study explores how Precitec's U.S. operation has integrated laser measurement technology into its workflow; a non-contact beam characterization system significantly elevates the company's service, quality assurance, and development processes for high-power laser applications.

Multi-Scale Investigation of Burn-Through Mechanisms during In-Service Welding of Pipeline Steel

2026-99-0409

8/28/2026

The core challenge of in-service welding repair of oil and gas pipelines is the risk control of burn-through. Current research primarily focuses on macroscopic phenomena, lacking a systematic multi-scale analysis of burn-through mechanisms and their dynamic evolution. Existing criteria are primarily based on qualitative experience, and widely accepted quantitative safety assessment standards have yet to be established. Furthermore, insufficient understanding of multi-scale damage failure mechanisms and weak theoretical foundations have become bottlenecks in this field. This study targets X65 pipeline steel and combines in-service welding experiments with in-situ scanning electron microscope tensile tests to elucidate the formation mechanism of burn-through from a multi- scale perspective. The results show that during in-service welding, the remaining wall thickness of the pipeline continuously decreases with the welding process, ultimately resulting in burn-through holes. On one hand, the welding arc drives the expansion of the hole; on the other hand, the internal pressure of the medium further enlarges the hole, leading to the expulsion of water and rapid pressure loss in the pipeline. Notably, the fusion zone behind the maximum melt depth is subject to high temperatures, which reduces strength and degrades plasticity, exhibiting significant plastic strain, making it a high-risk area for burn-through instability. Before instability occurs, this region shows evident grain coalescence, with plastic deformation primarily occurring through dislocation slip; when the difficulty of activating slip systems increases, twinning deformation may be induced, and large twin grains rarely develop cracks. Strain concentration and crack initiation are more likely to occur between grains with significant orientation differences.

Wang, Bangyu, Qiao, YingJie, Li, Dong, Xu, ShiHang

Effect of Copper Valence and Content in Copper Red Glaze on Its Color Appearance

2026-99-0420

8/28/2026

Copper red glazes have received considerable attention due to their perfect decorative effects and vivid coloration. This paper selected four different formula copper red glazes with different colors of bright red, violet blue, dark red, and gray blue as the samples. Based on the analysis of the colorimeter, X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), and scanning electron microscope (SEM), a possible coloration mechanism was proposed to explain the variation of glaze colors. The results indicated that the glaze layers were mainly composed of amorphous phases with few quartz diffraction peaks and mainly presented a granulous structure and phase separation. Increasing the content of CaO could cause color changes of the glaze due to the high ionic potential of calcium ions, which could form a unique feature in the glaze melt. In addition, a small amount of calcium phosphate could greatly change the color of the glaze. The phase separation structure of the blue samples was more obvious than that of the red samples, with a phase separation size of less than 100 nm. The formation of droplet phase separation structure in the glaze could lead to Rayleigh scattering and Mie scattering, which made the color of copper red glazes blue and opacified. Increasing the content of Cu0 and decreasing Cu+ could weaken the structural color, which contributes to a* value of the glaze changing from 26.93 to 22.93. At the same content of Cu0, the higher the ratio of Cu+ /Cu2+ is, the less a* value of the glaze is. Finally, the existence of CuSiO3 in the glaze could also make the blue color.

Ding, Erbao, Yang, Mengli, Liu, Nannan, Li, Yabo, Zheng, Ruimiao, Xu, Yan

Anti-splash System of Sprinkler Based on Millimeter Wave Radar and Machine Vision Fusion

2026-99-0323

8/14/2026

In view of the problems that it is difficult to accurately control the spraying area of the mining sprinkler, and the resource waste caused by the mis-spraying material stacking area, as well as the failure of traditional radar monitoring in the complex electromagnetic environment, this paper proposes an anti-splashing system for the mining sprinkler. By combining millimeter wave radar and visual recognition fusion technology, the overall scheme of the anti-splash system is proposed. Then the control simulation of the whole system is carried out. The results show that the problem of poor control in traditional sprinkler operation can be effectively solved, and the sprinkler area can be adjusted intelligently. Finally, in order to verify the accuracy of the algorithm used in this paper, different algorithms are used for comparative experimental verification. The results show that the Modified YOLOv4 algorithm has a high accuracy of 98.75 %, which has good applicability and provides a theoretical basis for subsequent research.

Hou, Lin

Study on Visible-Light-Driven Photocatalytic Degradation of Tetracycline Hydrochloride by g-C ₃ N ₄ /CoFe ₂ O ₄ Composite Photocatalysts

2026-99-1007

8/7/2026

g-C₃N₄, a metal-free semiconductor photocatalyst, demonstrates remarkable potential, but its practical application in pollutant degradation is significantly limited by the rapid recombination of photogenerated electron-hole pairs and low photocatalytic efficiency. To address this, a series of magnetic recyclable g-C₃N₄/CoFe₂O₄ composite photocatalysts with different CoFe₂O₄ doping ratios were innovatively designed and prepared via thermal polymerization, sol- gel, and combined with ultrasonic and heat treatment processes. The novelty of this composite design lies in the effective integration of magnetic CoFe₂O₄ with g-C₃N₄ through a heterojunction structure. It substantially boosts the absorption of visible light. Concurrently, it effectively fosters the separation and mobility of photo-induced charge carriers. The composite materials were systematically characterized by X-ray diffraction, thermogravimetric analysis, scanning electron microscopy with energy-dispersive X-ray spectroscopy, photoluminescence spectroscopy, and ultraviolet-visible diffuse reflectance spectroscopy. Using tetracycline hydrochloride as the target pollutant, the photocatalytic activity of the composites was evaluated under visible light irradiation, and the effects of initial concentration, catalyst dosage, and the influence of solution pH on degradation efficiency were also examined. The results indicated that the composite with a CoFe₂O₄ to g-C₃N₄ mass ratio of 1:3 (denoted as 3-CN/CFO) exhibited the optimal performance: a TCH degradation rate of 80.29 % within 105 minutes and a total organic carbon removal rate of 61.63 %. After five consecutive cycling experiments, the degradation efficiency remained above 70 %, demonstrating good reusability and stability. The performance improvement is attributed to the formation of heterojunctions in the composite, which effectively facilitates charge separation, inhibits carrier recombination, and enhances visible light absorption. Furthermore, the inherent magnetism of the composite permits efficient recovery, streamlining its integration into practical applications. Toward the purification of antibiotic-contaminated water, this research proposes a viable method for fabricating highly effective and recyclable photocatalysts.

Hua, Longjun, Chai, Tian, Wang, Yiming, Zhang, Jing, He, Ting

Experimental Study on the Combined Effects of Swirl and Inlet Offset on the Flow through an Automotive Catalyst Monolith Substrate

03-19-04-0019

8/7/2026

Maldistributed flow within an automotive catalyst can cause reduced conversion efficiency, high pressure loss, and premature deactivation. However, packaging constraints often result in uneven flow distribution between the monolith channels, thus compromising design and, inevitably, performance of the device. Flow uniformity may be improved by the introduction of swirl upstream of the catalyst assembly, and in turbocharged applications the residual swirl from the turbine can serve that purpose. Indeed, low swirl has been shown to provide favorable flow uniformity in the monolith substrate in an axisymmetric flow setup. However, the automotive exhaust aftertreatment setups are seldom axisymmetric, and the combined effects of inlet swirl and offset on the flow profile through a monolith substrate are unknown. To address this gap, this study provides the first systematic experimental characterization of the coupled influence of inlet swirl and packaging-relevant inlet offset on flow development and uniformity in a sudden expansion catalyst assembly. Particle image velocimetry (PIV), wall pressure measurements, and hot-wire anemometry (HWA) are combined to link the upstream separation and recirculation structures to the velocity distribution downstream of the monolith. The results reveal a previously unreported swirl-dependent sensitivity to geometric asymmetry: under no-swirl and moderate-swirl conditions, flow uniformity is robust to inlet offset, varying by no more than 1.4%, whereas at low swirl the offset reduces uniformity by up to 8% at high mass flow rate. Increasing mass flow rate reduces uniformity by up to 15%, while swirl improves uniformity by up to 19% relative to axial flow. These findings demonstrate that improvements observed for swirl in axisymmetric assemblies cannot be assumed to transfer directly to offset geometries. Swirl intensity and inlet alignment must instead be considered as coupled design variables. The measurements also provide a benchmark dataset for validating computational fluid dynamics simulations before their application to production-type systems.

Rusli, Ijhar, Aleksandrova, Svetlana, Medina, Humberto, Benjamin, Stephen F.

How Airbus is Developing Artificial Intelligence for Future Cockpits

26AERP08_11

8/1/2026

Computer vision, automated landing and embedded AI for tomorrow's cockpits. Airbus, Toulouse, France At the VivaTech forum in June, Airbus showcased a demonstration highlighting the use of computer vision to enhance automated landing procedures and operational efficiency. The “Vision Landing Application” utilizes artificial intelligence to analyze runway features in real-time using onboard cameras. The goal of this research is to create an additional and independent positioning source to guide pilots and/or their aircraft reliably, opening up the perspective of bringing autoland (fully automated landing procedure) capabilities to airports that lack advanced ground infrastructure. While the technology is still in the research phase and far from commercial certification, this technical exploration aligns directly with Airbus' global roadmap for Smart Automation. Airbus already has a head start, since it has already conducted numerous research projects during the last decade, which have led to the demonstrator at Airbus' stand at this year's show.

Microstructure and Mechanical Properties of Vacuum Laser Welded Joints in 42CrMo Thick Plate

2026-99-0132

7/31/2026

Vacuum laser welding trials were carried out on 42CrMo steel, a material widely utilized in the defense sector. By employing a 30 kW fiber laser system, complete penetration welds were successfully produced on 20 mm thick 42CrMo steel plates. The resulting joints displayed satisfactory surface quality on both the top and bottom sides, with no evident defects such as cracks or porosity. A comprehensive analysis of the joint microstructure and mechanical properties was conducted. Findings reveal that the weld zone (WZ) is predominantly composed of lath martensite, accompanied by minor quantities of plate martensite, organized as columnar crystals. The joints demonstrated high tensile strength at ambient temperature, with fracture consistently occurring within the base metal (BM). Microhardness measurements indicated higher values within the weld relative to the base metal, and no pronounced softening was detected in the heat-affected zone (HAZ). Additionally, the joints exhibited commendable impact toughness, suggesting overall superior mechanical performance.

Shi, Haicheng, Zhang, Guoyu, Li, Wuhong, Cao, Dongxu, Liu, Tianlei

Micro-Groove Fabrication and Simulation Analysis on PEEK Surface under UV Nanosecond Laser Irradiation

2026-99-0225

7/31/2026

This study investigates the interaction mechanism between ultraviolet nanosecond pulsed lasers and polyetheretherketone (PEEK). By integrating finite element simulations with experimental validation, the work explores the laser microtexturing characteristics of PEEK surfaces and evaluates the influence of microtextures on the material’s surface biocompatibility. First, the interaction between the laser and the PEEK polymer was analyzed, and a laser ablation model was established using the COMSOL Multiphysics simulation platform. Using finite element simulation, the effects of spot overlap ratio were investigated by adjusting the average laser power, while the influence of single-pulse energy on the ablation characteristics of the PEEK surface was examined by varying the scanning speed. Subsequently, ultraviolet nanosecond laser processing experiments were conducted on planar PEEK microtextures based on the simulation results. Taking surface microgrooves on PEEK as representative structures, the variations in groove depth and width under different combinations of laser parameters were analyzed. The parameters, including average laser power, scanning speed, and repetition frequency, were optimized to identify processing conditions that yield stable depth and width, along with good surface flatness. Finally, experiments have initially verified that the microtextured PEEK surface may improve biocompatibility and regulate surface wettability to a certain extent.

Wu, Yifan, Wang, Xiaohui, Han, Yujie, Jin, Shuo

A Stress Relief Method Based on a Hybrid Process of Shot Peening and Laser Directed Energy Deposition with the Same Materials

2026-99-0232

7/31/2026

Laser directed energy deposition (LDED) is widely used in various fields due to its fine forming structure and superior performance. However, the characteristics of the hot forming process result in significant residual tensile stress in the formed materials, which affects the capability and useful life of the mechanism accessory. The hybrid manufacturing technology of shot peening (SP) and LDED has a significant influence on the elimination of defects and the improvement in microstructure of formed materials and reducing residual stress, but it also has limitations. To solve the problems, such as the introduction of powders and the difficulty in recycling and classification when using heterogeneous materials for shot peening in hybrid processes, this paper proposes a method of strengthening with the same material, establishes a thermal shot peening simulation model for the hybrid process, and conducts experimental verification. The research finds that the average generated stress of SP in hybrid manufacturing technology is -215.6 MPa, and the thickness of the strengthening layer is about 40 μm. The subsequent hot forming process will eliminate part of the induced stress by SP on the previous deposition, but the deposited stress on the surface is reduced compared with that in the single process. The hybrid manufacturing technology of SP and LDED, based on the same material, effectively utilizes the residual heat from the forming process, providing feasibility for engineering applications.

Zhang, Xiaoyu, Zhang, Min, Li, Dichen, Jiang, Yunfeng, Chen, Xinjin, Fei, Ya, Hu, Ying, Liu, Yuyang

Discussion on Methods for Studying Material Fatigue Life: Taking Aluminum Alloy as An Example

2026-99-0121

7/31/2026

Fatigue design is a key common quality technology for improving the quality control capability of China’s automotive products. The fatigue of materials is a multi-scale damage evolution process. Characterizing and processing the large number of three-dimensional defects inside the material, which have different shapes and distributions, and predicting the material’s lifespan based on the cross-scale damage evolution mechanism, is one of the key technologies for fatigue optimization design. This paper discusses the research methods for the fatigue life of aluminum alloy materials. Firstly, based on the staged fatigue damage experiments, the three-dimensional defect features are obtained through CT scanning and reconstruction, and a defect characterization and processing method based on k-d tree and multi-scale feature pyramid is established to accurately represent the topological and geometric relationships of non-uniformly distributed three-dimensional defects. Secondly, a mathematical model for the evolution of micro-damage and macro-cracks is constructed, and the cross-scale transformation of defects is achieved through hierarchical and recursive methods, revealing the cross-scale evolution mechanism of fatigue damage in aluminum alloy materials. Finally, a remaining life prediction model based on defect information and feature weights is established through the support vector regression algorithm (SVR). This research method can provide technical support for the fatigue life optimization design application of lightweight materials such as aluminum alloys.

Zhang, Liangxia, Niu, Zhijun, Cheng, Fangfang, Chen, Hao, Yang, Yali

Design and Optimization of Hollow Rotating Traveling Wave Ultrasonic Motor Based on Laser Beam Expander

2026-99-0154

7/31/2026

To meet the critical need for rapid response and miniaturization in laser beam expander drive systems, this study proposes an innovative actuation solution based on a hollow rotary traveling-wave ultrasonic motor. By thoroughly analyzing the optical adjustment mechanism of laser beam expanders and the electromechanical coupling behavior of ultrasonic motors, the motor structure was systematically optimized. Using a multiphysics coupling approach, the performance of stators fabricated from three distinct materials was compared, and parametric optimization was conducted. Experimental verification confirms that the developed ultrasonic motor precisely matches the load characteristics of beam-expanding optics while fulfilling the stringent requirements for both fast response and compact design. This research provides a reference for the miniaturization drive of high-precision optical systems, with promising applications in space optics and precision instrumentation.

Qiu, Haihui, Niu, Chuanhu, Xiao, Zhong, Xu, Zhangfan, Li, Jialiang, Pan, Song

Study on the Contact Stress Distribution and Calculation Method of Cylindrical Roller Bearings Considering Clearance Changes

2026-99-0124

7/31/2026

The distribution of contact stress in roller bearings has a significant impact on operational performance and safety. Firstly, we established a bearing clearance change model that combines interference fit and thermal expansion effects. Then, we studied the clearance changes of key parameters’ influence under different operating conditions. Using Hertz contact theory, we analyze the nonlinear coupling relationship between clearance changes, load distribution, and contact stress. Through MATLAB analytical calculations, load and stress distribution contour maps were obtained under typical operating conditions, which provided theoretical support for bearing optimization design and reliability analysis. The result depicts that an increase in interference fit and temperature difference leads to clearance decrease, triggering a redistribution of contact stress. As clearance decreases, the maximum contact stress exhibits a nonlinear growth trend. To further enhance engineering practicality, this paper uses the MATLAB platform to develop a visualization of digital image processing software. The software enables interactive analysis throughout the entire process of clearance input, stress calculation, and graphical display.

Pang, Yiqing, Cai, Hongbin, Ren, Siyang

Construction and Verification of Munich EV Driving Cycle Considering Road Slope

2026-99-0869

7/30/2026

The driving cycle is the basic model of certification of vehicle fuel consumption and emissions, or calibration of powertrains. Standard regulatory driving cycles, such as WLTC, in general assume flat roads during their generation and fail to take into account the strong effect that road gradients have on vehicle operation and driving energy consumption. Such a shortcoming, then, leads to gross mismatches in adaptability when used for urban environments with typical hilly topography. To solve this problem, in this paper, we proposed a method for building driving cycles that consider the impact of slope with actual driving data. Initially, high-precision onboard data collectors were used to generate a total sum of 21, 350 km of driving data from the Munich area, thus creating a diversified driving data set with details such as vehicle speed, slope, and environmental information. Subsequently, joint probability distributions of “speed-acceleration” and “slope-slope change rate” are proposed by using the Micro-trip Method, and a novel chi-squared test algorithm is used to obtain a higher fidelity of urban driving cycle representative of typical conditions. Results of the driving cycle results show that the driving cycle built was close to the actual kinematics, indicating a deviation of less than 5%, and can capture the average uphill characteristic of 1.7%, which is quite well represented. Finally, in fact, validation of whole vehicle environmental chamber tests further demonstrates that the energy consumption prediction error of the developed driving cycle is just 2.2%, much lower than 19.1% error of WLTC. It highlights the importance of considering slope parameters in improving the accuracy of energy consumption calibration for an EV operating on complex slope terrains. Furthermore, it underscores that converting the real-world driving data into lab-based driving cycles can reduce the cost and time of actual road tests for Chinese companies going to the overseas markets, thereby offering support for the international marketing strategy of a global database.

Tian, Lichen, Jiang, Ping, Gao, Wang, Liang, Yongkai, Ma, Kunqi, Yu, Hanzhengnan

Design of Telescope Support Truss Based on Topology Optimization and Size Optimization

2026-99-0848

7/30/2026

The primary mirror support truss of large-aperture segmented telescopes, serving as a critical load-bearing component of the optical system, has its structural stability directly determining the optical imaging quality. This paper adopts a collaborative design method integrating topology optimization and size optimization to address issues, including excessive weight and unreasonable stiffness distribution in traditional support truss designs. First, based on the topology optimization theory of the Solid Isotropic Material with Penalization variable density method, topology optimization was performed on the initial truss structure using finite element simulation software, with the volume fraction as a constraint and the objective of maximizing structural stiffness to determine the optimal material distribution model. Subsequently, the truss structure was reconfigured based on the topology optimization results. Finally, the cross-sectional dimensions of the truss members were selected as optimization variables, and size optimization was performed using the NSGA-II multi-objective optimization algorithm with the objectives of minimizing structural weight and minimizing weighted compliance, while considering constraints such as stress and displacement. The results show that the optimized support truss achieves a 3.9% reduction in weight and a 35.47% decrease in elastic strain energy. This effectively meets the high-precision and lightweight design requirements for telescope support structures and provides a feasible technical solution for the design of large-aperture telescope support trusses.

Tan, Deli, Guo, Liquan, Gao, Dedong, Liu, Chuanjie, Dai, Xiaodong, Huang, Lei

Integrating 4D Millimeter-Wave Radar with YOLOv5-Monster for Distance Measurement

2026-99-0853

7/30/2026

To address the challenges of binocular vision ranging under complex environmental conditions—such as illumination variations, occlusion, and textureless regions, which result in unreliable and non-robust performance—this paper proposes a multi-source heterogeneous sensor fusion ranging method integrating 4D millimeter-wave radar with the YOLOv5-Monster framework. This method is capable of overcoming the issue of limited ranging accuracy in monocular or binocular vision algorithms under non-ideal imaging conditions. This study achieves high-precision spatial perception through the following specific pipeline: First, Zhang’s calibration method is used to obtain the intrinsic and extrinsic parameters of the binocular camera, and stereo rectification is performed on the raw images. Next, a lightweight YOLOv5 network is employed for object detection, while a high-performance Monster network is utilized to generate dense disparity maps, thereby accomplishing initial depth estimation. To mitigate the inherent depth estimation errors of vision-only systems, 3D point cloud data from a 4D millimeter-wave radar is further introduced. By applying a Kalman filter algorithm, the millimeter-wave radar point cloud and visual outputs are fused, achieving spatiotemporal synchronization and optimal state estimation across modalities and effectively correcting biases in visual ranging. Experimental results show that within the full range of 4 to 150 meters, the relative error of the proposed method remains below 5%. Specifically, the relative errors are 1.25% (absolute error: 0.05 m) at 4 meters, 1.40% at 5 meters, 2.99% at 75 meters, and 4.91% at 150 meters. Compared with the vision-only Monster-YOLOv5 baseline method, the relative error at 150 meters is reduced from 13.16% to 4.91%, representing an accuracy improvement of over 60%. Meanwhile, in terms of long-distance error control, the proposed method significantly outperforms traditional stereo matching approaches such as SGBM+YOLOv5 and BM+YOLOv5, reducing errors by more than 20 percentage points. These results verify that deep multi-modal fusion can enhance environmental adaptability and measurement reliability, providing a high-precision and highly robust solution for distance estimation in intelligent perception systems, which holds important theoretical and engineering significance.

Li, Fugai, Xie, Yuwen, Su, Hao, Liu, Donglei, Wu, Qiong

Research on Welding Deformation of the Car Body Based on 3D Laser Scanning

2026-99-0821

7/30/2026

This study aims to verify the accuracy and stability of a system used for measuring and analyzing the welding deformation of vehicle bodies under different welding parameters. A 3D laser scanner was employed to capture the surface topography data of the vehicle’s front deck before and after welding. In order to determine the welding deformation, PolyWorks software was utilized for deformation analysis, which processed the 3D scanning data and compared the post-welding data set. A dedicated vehicle body welding deformation measurement system was developed, including hardware configuration and software development. The BP neural network algorithm was adopted to predict the welding deformation, and the results indicated that the deviation between the predicted values and the average experimental measurements was less than 10%. This confirmed the practicality of the BP neural network in predicting welding deformation and highlighted its effectiveness in technical support for the optimization of welding parameters and deformation control in automotive manufacturing.

Li, Lina, Zhang, Yiqi, Sun, Hongchang, Wei, Xiezhen

Three-Dimensional Transient Thermal-Fluid Coupling Numerical Simulation of the Molten Pool in Aluminum Alloy Laser Welding

2026-99-0854

7/30/2026

Laser welding technology for aluminum alloy electrode and busbar connections: addressing challenges in battery module assembly. In this work, a CFD framework was built in ANSYS Fluent using a Gaussian rotating heat-source representation, while a VOF approach was used to capture the transient gas–liquid interface in deep-penetration welding. A three-dimensional, transient, thermal-fluid coupled numerical model of the dual-layer heterogeneous aluminum alloy laser deep penetration weld pool was established concurrently with laser deep penetration welding experiments. Results indicate: Peak flow velocities in the weld pool during welding are concentrated along the weld centerline, with flow vectors predominantly directed axially along the weld. Once a quasi-steady keyhole regime is established, vaporization-induced recoil pressure becomes the primary driver governing melt circulation. The liquid metal first impinges on the pool bottom along the keyhole wall and then recirculates upward near the pool boundary, producing strong vortical motion. These findings are intended to support parameter selection and process optimization for laser welding of layered dissimilar aluminum components used in battery tab–busbar assemblies.

Lv, Wenjun, Wu, Yan

Asteroid Landing Navigation Based on Invariant Extended Kalman Filter

2026-99-1856

7/17/2026

Optical navigation serves as a critical modality for autonomous guidance during small celestial body landing missions. To address the inherent strong nonlinearities in both the lander’s dynamic model and optical observation model, this paper investigates an invariant extended Kalman filter algorithm based on Lie group structures. First, we establish the state model and optical observation model on the special Euclidean group. Subsequently, a linearized right-invariant error dynamics equation is derived using invariance theory, along with the formulation of state prediction models. Furthermore, the feature vector observation model is modified into a right-invariant observation form, enabling state correction through exponential mapping of innovation vectors. Numerical simulations using asteroid Eros 433 demonstrate that the proposed invariant extended Kalman filter (InEKF) outperforms the conventional extended Kalman filter (EKF) in both estimation accuracy and convergence speed. Notably, the algorithm eliminates the need for online Jacobian matrix computations, satisfying the stringent navigation requirements for autonomous landing operations. The results validate the effectiveness of Lie group-based filtering in handling the nonlinear geometry of pose estimation for irregular celestial bodies.

Liu, Zhengdong, ZHU, Shengying

Multi-Frame Centroid Extraction Method for Optical Navigation During the Approach Phase of Small Body

2026-99-1851

7/17/2026

Autonomous optical navigation is one of the important navigation methods for the small bodies approach phase. To improve optical navigation performance during the approach phase to a small body, this paper presents a method for extracting the target centroid from sequential optical images. The process begins with fitting a minimum enclosing ellipse to the detected contours in each frame to obtain an initial estimate of the centroid. Building upon this, edge corner points across adjacent images are matched using normalized cross-correlation, and their displacement is tracked using optical flow techniques. The observed pixel trajectories are analyzed, and a predictive model of pixel motion is formulated based on the geometric relationship between the detector and the small body. By combining the directly extracted centroids with the predicted motion of key pixels, a fusion strategy is developed to improve the reliability of the centroid estimation. Finally, numerical simulation results demonstrate that the method significantly improves the accuracy of centroid extraction, thereby enhancing the overall performance of optical navigation during approach operations.

Liu, Jing, Zhu, Shengying

Landing Site Selection and Visual Scene Generation Based on LiDAR Point Cloud

2026-99-1819

7/17/2026

This paper focuses on autonomous drone landing scenarios. Addressing the core requirements of accurate landing site assessment and intuitive visual presentation, it conducts in-depth research on the application of 3D LiDAR (TOF technology) point cloud data. LiDAR captures point cloud data containing 3D coordinates and reflection intensity values. While sparse, non-uniform, and disordered, its high measurement accuracy and strong anti-interference capabilities make it a key sensor for landing terrain perception. Based on a review of recent research results from related teams, this study designed and implemented a comprehensive technical solution: First, raw point cloud data is acquired via the UDP protocol combined with an SDK interface. Preprocessing is then performed using voxel grid filtering (downsampling) and radius filtering (denoising). The assessment area is then divided into a row-by-column grid. A sliding window method is used to calculate the elevation difference, empty grid ratio, flatness, and slope of each grid. Based on these attributes, the grids are classified into six categories: Risk, Warning, Blank, Unknown, No Landing, and Landing. Finally, a grid attribute coloring method and OpenGL 3D rendering are used to generate the visual scene. Through the development of verification programs and moving obstacle experiments, it has been proven that the solution can efficiently process point cloud data and accurately identify safe landing areas, providing key technical support for the engineering realization of the autonomous landing function of drones, and also laying the foundation for the intelligent development of drone landing decisions in complex environments.

Guo, Hangyu, Shi, Zhe

Research on Clustering Data-Driven Intelligent Perception Algorithm for Micro Mobile Robots

2026-99-1843

7/17/2026

To address the challenges faced by micro flapping-wing flying robots in visual navigation—specifically, the large volume of visual information and the difficulty in transforming it into usable intelligent visual data—this paper proposes a clustering-based data-driven approach for directional and image perception. The aim is to enable intelligent visual navigation for flapping-wing robots. The proposed method performs clustering analysis on gyroscope data from the flapping-wing robot to extract directional features. Simultaneously, it applies clustering techniques to visual images captured by the robot to identify intelligent features such as edges. This approach enables the robot to acquire multiple optimized perceptual data types, thereby enhancing the behavior control system. Through the use of clustering analysis, the method not only improves the effectiveness of visual navigation but also extracts features related to visual targets and environmental information, providing technical support for visual target tracking. The experimental platform consists of a flapping-wing robot equipped with an onboard camera, and the proposed clustering-driven visual image perception approach has been experimentally validated. Experimental results demonstrate the high feasibility and effectiveness of the method in practical applications. The main contributions of this study lie in two aspects: (1) a clustering-driven visual image perception method for flapping-wing robots, and (2) a clustering-based approach for identifying posture and behavioral patterns of flapping-wing flying robots.

Li, Zixuan, Ding, Wei, Zhang, Feng, Song, Min, Liu, Zhaoming, Miao, Lei, Liu, Haotian, Bai, Ning, Tian, Shen, Cui, Long, Wang, Hongwei

Analysis of Quadrotor UAV Position Accuracy Based on EKF Multi-Sensor Data Fusion

2026-99-1859

7/17/2026

When quadrotor unmanned aerial vehicles (UAVs) operate in urban low-altitude airspace, especially within complex environments, their sensor perception signals are highly susceptible to blockages, deviations, and the inclusion of high-frequency noise. These factors, in turn, induce nonlinear variations in the UAVs’ flight mechanical properties, giving rise to abnormal flight stability issues such as attitude jitter, altitude fluctuations, and trajectory deviations. To address these challenges, this paper puts forward a method aimed at enhancing the positional accuracy of quadrotor UAVs, which is based on Extended Kalman Filter (EKF) multi-sensor fusion. In conjunction with the redundant configuration of sensors, a proportional-integral controller is specifically designed to allow optical flow sensors to compensate for the speed data generated by inertial sensors. Building on the EKF method, a comprehensive data fusion model is established, encompassing both position and speed states. Leveraging the MATLAB platform, trajectory flight simulations are conducted, utilizing multi-sensor data fused via EKF, with the sensor suite including GPS, IMU, Optical Flow sensors, and Barometers. The simulation results demonstrate that this proposed method can effectively mitigate the adverse impacts of environmental interference and sensor noise on the positional accuracy of quadrotors. By continuously correcting position information and accurately estimating position states, it significantly improves the UAVs’ flight position accuracy. This research outcome lays a robust and theoretically sound foundation for in-depth investigations on critical issues related to general aviation applications, such as the safe and efficient autonomous flight, adaptive and reliable intelligent navigation, and ultra-precise and mission-critical operations of quadrotor UAVs, thereby significantly contributing to the sustained and innovative advancement of the field.

Cui, Nan, Liu, Wenzhi, Liu, Hanqi, Wang, Jingrui, Wang, Zhizhong, Zhi, Haonan

Few-Shot Semantic Segmentation Method for Crater Detection on Small Bodies

2026-99-1852

7/17/2026

Craters are the primary landmarks used for visual navigation in missions exploring small celestial bodies. However, obtaining high-quality, annotated crater data is often challenging due to limited imaging conditions and strict mission constraints. Conventional semantic segmentation models struggle with limited data and are challenging to train effectively. To overcome this limitation, this study introduces a few-shot segmentation approach for crater detection on small celestial bodies. Our method includes a prototype representation module that constructs class-level prototypes to quickly associate crater regions with their semantic features. This paper also designs an iterative learning module that gradually improves the segmentation output, helping the model better capture detailed edges and structures. Tests on a simulated few-shot dataset demonstrate that our method provides reliable and accurate crater segmentation, achieving a mean intersection-over-union (mIoU) of 88.7, outperforming traditional fully supervised methods.

Li, Shuai, Zhu, Shengying

Aircraft Optoelectronic Payload Transfer Alignment and Dynamic Error Compensation Algorithm Research

2026-99-1849

7/17/2026

This paper proposes a multi-source dynamic error compensation algorithm for the transfer alignment of airborne optoelectronic payloads. This method addresses performance limitations of micro-inertial navigation systems (micro-INS) in complex dynamic environments, specifically those arising from accumulated device noise and the inability to perform static alignment due to installation errors. The algorithm’s core is the Extended Kalman Filter (EKF) technology. By constructing a “velocity + attitude” matching model between the UAV’s master inertial navigation system (MINS) and the optoelectronic payload’s slave inertial navigation system (SINS), it leverages high-precision MINS navigation information to correct SINS errors. Utilizing a 21-dimensional state space equation and measurement equation, the algorithm achieves real-time estimation and compensation of various errors, including attitude misalignment angles, sensor biases, installation errors, and flexure deformation. Simulation results demonstrate significant alignment accuracy improvement. Post-lever arm effect compensation, velocity errors are stably controlled within 0.01 m/s. Concurrently, flexure deformation angle compensation substantially reduces misalignment angle fluctuations across all directions, enhancing system stability and maintaining low misalignment angles. These findings validate the proposed error compensation strategy’s effectiveness.

Zhang, Lu, Li, Mao, Wang, Shiyong, Lei, Chao

Drop Impact Phenomena: An Experimental Study of Isooctane Interaction with Dry Walls and Liquid Films

2026-01-5015

7/13/2026

In this study, high-speed back-illuminated imaging and laser-induced fluorescence (LIF) methods were employed to investigate the impingement behavior of millimeter-sized single isooctane drops on a dry solid wall and various liquid films, including isooctane and glycerol solution films of different concentrations. Various fuel spray impingement scenarios in gasoline direct injection engines were examined. High-speed back-illuminated imaging was primarily used to examine the impact of fuel drops on a dry wall and a fuel film of the same composition as the drops. The LIF method was used to examine the impact of fuel drops on the glycerol solution film, allowing for the distinction between fuel drops and the glycerol solution film. The impingement behavior varied depending on the Weber number of the incident drop and the wall condition. When fuel drops impacted the solid dry wall vertically, they spread into a circular liquid film. The outer edge of the liquid film folded and bulged, and upon reaching the maximum spreading diameter, it maintained equilibrium and did not retract. When isooctane fuel drops impacted the isooctane film, they broke and splashed, with thinner films producing stronger splashes. Additionally, the Weber number of the fuel drops significantly influenced the crown shape and splashing after impact. The impingement behavior of fuel drops on the glycerol solution film was also investigated, focusing on the liquid film morphology after impact. Based on the experimental data, empirical correlations were established between the critical Weber numbers for transitions among different crown morphologies and the dimensionless film thickness under varying film viscosities.

Yang, Tian, Lu, Lili, Guo, Zongwei, Song, Enzhe, Yao, Chong, Ning, Yilin, Ke, Yun

Validation and Exploration of Multimodal Deep-Learning Camera–LiDAR Calibration Models

2026-01-5037

7/13/2026

This paper presents an innovative study in exploring, evaluating, and implementing deep-learning architectures for the calibration of multimodal sensor systems. The aim of this paper is to leverage the use of sensor fusion to achieve dynamic, real-time alignment between 3D LiDAR and 2D camera sensors. Static calibration methods are tedious and time-consuming, which is why we propose utilizing conventional neural networks (CNNs) coupled with geometrically informed learning to solve this issue. We leverage the foundational principles of extrinsic LiDAR–camera calibration tools such as RegNet, CalibNet, and LCCNet by exploring open-source models that are available online and compare our results with their corresponding research papers. Requirements for extracting these visual and measurable outputs involved tweaking source code, fine-tuning, training, validation, and testing of each of these frameworks for equal comparisons. This approach aims to investigate which of these advanced networks produces the most accurate and consistent predictions. Through a series of experiments, we reveal some of their shortcomings and areas for potential improvements. We find that LCCNet yields the best results among all the models that we validated.

Karramreddy, Venkat Sai Raxit, Mitchell, Liam

Missile Pose Measurement Based on ArUco Markers

2026-99-0576

7/10/2026

This paper presents a monocular vision-based system for high-precision missile pose measurement using ArUco markers and Perspective-n-Point (PnP) algorithms. By deploying 6 × 6 ArUco markers on a cylindrical missile mock-up, the system establishes 3D-2D correspondences between structured-light-scanned models and camera images to solve the PnP problem. The proposed approach integrates optimized ArUco marker recognition — leveraging adaptive thresholding, contour simplification, and grid-based validation — with the Efficient PnP (EPnP) algorithm to achieve real-time pose estimation. Experimental validation demonstrates angular accuracy of ± 0.3° in roll/pitch/yaw and positional accuracy of ± 2 mm within a 2 m range under controlled conditions. The system exhibits robustness against partial occlusions and motion blur, with degraded performance (± 1.2°, ± 5 mm) in extreme scenarios. Key innovations include a streamlined marker detection pipeline and adaptive pose refinement using Levenberg-Marquardt optimization. This work provides a cost-effective, non-contact solution for flight tests, with potential applications in weapon separation testing.

Wang, Ruiyang, Zhang, Chaofan

Research on the Design of Lane Line Recognition and Deviation Warning System Based on Machine Vision

2026-99-0566

7/10/2026

In order to reduce traffic accidents caused by cars straying from lanes, a lane line recognition and deviation warning system based on machine vision is designed. It mainly includes image preprocessing, lane line detection, and the design of a deviation warning model. “In this study, an ROS-based intelligent vehicle-mounted camera is adopted for road image collection. To reduce the computational load of data processing while guaranteeing the algorithm’s accuracy and reliability, grayscale conversion and region of interest (ROI) extraction are implemented to finish the image preprocessing stage. Additionally, a fusion strategy of global and local thresholds is introduced to enhance both the operational speed and detection accuracy of the algorithm” use the Canny operator for the edge feature extraction; and complete the fitted lane lines with the improved Hough transform. Finally, based on the Kalman filter and camera viewpoint conversion coefficient algorithm, the lane line offset is detected in real time, and the deviation is judged in combination with the monitoring interface. Simulation experiments show that the system is able to effectively recognize the lane line and judge the deviation status under the condition of setting the offset threshold of 70 pixels, which significantly improves the accuracy and real-time performance of the lane deviation warning and provides effective technical support for reducing traffic accidents.

Wang, Xufeng, Zhang, Chunshu, Wang, Yan, Chen, Yihui, Ji, Rui

Items per page:

1 – 50 of 10614