Browse Topic: Autonomous vehicles

Items (3,215)
Lunar dust consists of extremely fine particles and exhibits electrostatic charging properties and electrostatic adhesion. These characteristics cause lunar dust to be highly susceptible to mobilization during lander touchdowns, rover traversals, and human activities, forming widely distributed dust clouds. Lunar dust contamination not only abrades spacecraft and equipment to impair their performance but also poses a threat to astronauts’ safety. To verify the impact of the lunar dust environment on exploration equipment components, a simulation mechanism adaptable to the thermal vacuum test environment was designed. This mechanism is integrated into the lunar environment simulation system and uses a vacuum stepper motor to drive a ratchet mechanism, enabling precise vibrational injection of simulated lunar dust. It mainly consists of a pretreatment mechanism, a particle sedimentation mechanism, a shielding mechanism, and an ultraviolet (UV) irradiation system. Considering the vacuum operating environment, alternating high and low temperature conditions, as well as the strict requirements for the mechanism’s compact size and high reliability, this paper analyzes in detail a series of problems encountered during the development of the mechanism and their corresponding solutions. Stainless steel and polytetrafluoroethylene (PTFE) were selected as the main materials for the mechanism. Meanwhile, active temperature control measures were adopted to actively regulate the temperature of components such as the motor. Ultimately, the mechanism can withstand alternating high and low temperatures ranging from -150°C to 150°C and a vacuum environment of 5 × 10^–6 Pa. Under this environment, the mechanism can achieve vibration frequency adjustment within the range of 1-5 Hz, and realize the sedimentation of simulated lunar dust particles with a particle size of less than 200 μm over an area of 150 mm × 150 mm. After sedimentation, the simulated lunar dust particles can be charged through the photoelectric effect.
Xu, MenglongLv, ShizengLi, GuohuaGong, Jie
Historically, the demand for advanced technology, efficiency, and safety has been a primary driving force in the evolution of commercial vehicles, particularly with respect to braking systems. More recently, the increasing levels of vehicle autonomy and electrification have emerged as irreversible trends, significantly accelerating the development of new functionalities and innovative electrical/electronic [E/E] architectures. These advancements are essentially focused on performance optimization, risk mitigation, and enhanced system reliability through the application of functional safety and cybersecurity standards, thereby shaping the current landscape of braking system design. From an efficiency standpoint, braking systems with higher levels of electronic content, functional integration – included with regenerative braking systems - and harmonization have been developed to improve energy efficiency and support global scalability. Concurrently, new system configurations are continuously being introduced to enhance vehicle safety and advanced driver assistance capabilities, in alignment with evolving regulatory requirements and market expectations. This paper evaluates the impacts of automation and electrification on commercial vehicle pneumatic braking systems, focusing on Anti-lock Braking Systems [ABS], Electronic Braking Systems [EBS] and air management platforms. It provides a technical overview of both architectures, assessing their capabilities to meet modern requirements such as integration with advanced vehicle architecture, regenerative braking for electrified applications, and Advanced Driver-Assistance Systems [ADAS] support. The study details the evolution of air management systems, with emphasis on electrified vehicles, including key functions such as air compressor charge control, Air Processing Unit [APU] desiccant regeneration, and electronic control strategies. Additionally, it examines key drivers of braking system evolution, braking system selection considering ADAS regulatory developments, Net Zero strategies, and automation trends. The paper further evaluates compliance with functional safety and cybersecurity standards and assesses the readiness of both platforms for emerging mobility concepts. Finally, it highlights the risks of deploying higher levels of autonomy in heavy-duty towing vehicles when operating with non- ABS semi-trailers, identifying this as a critical area for further investigation.
Guarenghi, Vinícius MendesNicora, FabioPizzi, Rafael FortunaResende, Angelo Roberto RodriguesPinto, Gustavo Laranjeira
Trajectory tracking control serves as the core operational component of autonomous vehicles, directly determining driving safety and passenger comfort by ensuring control precision and stability. To enhance the tracking accuracy and stability for autonomous vehicles, this study proposes a coupled lateral–longitudinal trajectory tracking controller based on multi-agent reinforcement learning. The framework first establishes a Model predictive controller (MPC) derived from vehicle dynamics, formulating the lateral control process as a Markov decision process. A reward function incorporating lateral error, heading error, and steering angle is designed, followed by the construction of a Deep Q-Network (DQN) Agent to optimize the prediction horizon of the MPC. Subsequently, a position–velocity dual-loop PID controller is developed for longitudinal control, with its parameter optimization strategy learned through a Deep Deterministic Policy Gradient (DDPG) Agent. The Extended State Observer (ESO) is incorporated to perform steering angle compensation for internal modeling errors and external disturbances. Co-simulation experiments are conducted in CarSim and MATLAB/Simulink, and the results demonstrate that the coupled controller achieves superior tracking accuracy and stability in both overtaking and lane-changing scenarios compared with the decoupled controller.
Kun, FengJinxiang, ZhaiLi, Wenli
Connected and Automated Vehicles (CAVs) represent a transformative innovation poised to revolutionize roadway transportation by leveraging automated driving systems equipped with advanced sensors, high-performance computing, and communication technologies. While urban areas are the primary focus of current CAV developments, rural transportation systems risk being left behind despite the significant benefits that CAVs can bring to these regions. This article, therefore, explores the physical and digital infrastructure requirements for the safe deployment of CAVs in rural areas, drawing insights from standards, recommendations, and guidelines developed by leading standard organizations. The study highlights the specific design of physical infrastructure, including traffic signs, traffic signals, and pavement markings, and digital infrastructure, including communication, sensing, and mapping, to ensure rural communities are effectively prepared to benefit from the potential of CAVs. As its primary contribution, this article provides a comprehensive review of existing standards and guidelines relevant to rural CAV deployment. By synthesizing guidance across multiple standard-setting organizations, this review delivers a structured analytical assessment of existing standards, revealing their limitations and misalignment with rural transportation contexts while highlighting emerging good practices. The article clarifies the applicability of current guidance to rural infrastructure, identifies systemic infrastructure-related failure modes, and informs context-aware planning considerations for efficient and scalable CAV deployment in rural areas.
Zakaria, MohammedGetahun, TesfamichaelTavasoli, MahsaPandey, VenkteshSarrafzadeh, AbdolhosseinKarimoddini, Ali
North American CAV Performance Data StandardWP-00157/22/2026
As the deployment of connected and automated vehicles (CAVs) expands, the need for a consistent, cross-industry approach to performance relevant CAV data exchange is becoming more pressing. Vehicle developers, infrastructure owners and operators (IOOs), and technology providers generate and consume data that support safety, mobility, and operational efficiency, yet much of the data remains fragmented, inconsistently formatted, and difficult to reuse across systems. To address these gaps, the Society of Automotive Engineers (SAE) and the Canadian Standards Association (CSA) convened a multi-stakeholder workshop on November 3, 2025, with participants representing original equipment manufacturers (OEMs), automated driving system (ADS) developers, state and local agencies, standards bodies, and technology partners. The workshop focused on identifying challenges, clarifying needs, and outlining a path toward a North American CAV Performance Data Standard. Key themes from the workshop included: -The need for a shared data language to support safe and interoperable CAV operations. -The lack of consistent formatting, labeling and visibility regarding who produces and consumes data. -A “start small, iterate and scale” approach beginning with well-defined use cases. -The need for technical harmonization and governance structures that build trust and support sustained coordination. This white paper summarizes the key findings and outlines a practical approach to developing a Version 0.1 base-layer data standard that can support measurable progress in 2026 and beyond.
Nesheli, Mahmood
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, ZhengdongZHU, Shengying
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, JingZhu, Shengying
To ensure the successful implementation of the separation, evacuation, and return processes of manned spacecraft after long-term docking at the space station, regular on-orbit health assessments must be conducted. Based on this requirement, a technical method for evaluation through autonomous on-orbit testing is proposed. First, the docking status and characteristics of the manned spacecraft’s systems, such as information management, crew environmental control, thermal control, power management, docking function, attitude, and orbit control function, are described. Then, the functional requirements for the separation, evacuation, and return of the manned spacecraft, such as the relative measurement, the relay communication, TT&C and data transmission, image and voice, instrument display and alarm, and the attitude measurement, are analyzed. Subsequently, the on-orbit testing system, test items, test procedures, and test methods for health assessment are detailed. It also provides the design of TT&C support, the design of energy support, and the main principle explanation for autonomous on-orbit testing of the system.
Cheng, WeiNan, HongtaoTian, YeZhao, Zheng
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, HangyuShi, Zhe
This paper constructs a reinforcement learning framework based on the PPO algorithm for drone air combat to solve 1v1 pursuit-evasion in 2D beyond-visual-range air combat. Firstly, the mission scenario is modeled, defining key roles of ATA and AA. Then, state transition models of pursuer and evader are built based on flight kinematics. To handle reward sparsity in policy network training, a dense reward function combining distance and angle rewards is designed to guide the agent in learning tail-chasing and interception strategies. Using the Actor-Critic architecture, deep neural networks implement the decision-making and evaluation modules. The PPO algorithm trains the pursuing drone in a simulation. Results show that after ~5 million steps, the agent learns a stable strategy, completing tasks promptly and generalizing well in unseen scenarios. This research offers ideas for drone combat and guidance, and supports autonomous decision-making in complex air battles.
Yu, KangjieGong, ZhengHu, RunchangLiu, Huixiang
This study addresses the challenges of communication delays and system stability in autonomous obstacle avoidance (AOA) systems under next-generation vehicular electronic/electrical architectures. A centralized PON-based architecture is proposed, leveraging XGSPON technology to enhance bandwidth capacity and reduce electromagnetic interference, while rigorously analyzing worst-case in-vehicle communication (IVOC) delays. To mitigate latency impacts, a Software-Defined Networking (SDN)-driven dynamic scheduling strategy prioritizes safety-critical data streams (e.g., environmental perception, motion control) through adaptive resource allocation. Further integrated with a robust H-infinity LQR controller, the co-design framework ensures precise trajectory tracking and suppresses steering oscillations under communication uncertainties. Simulation tests validate the framework's efficacy, demonstrating significant reductions in loop delays and improved dynamic stability in complex scenarios. This work bridges communication efficiency and control robustness, offering a scalable solution for advancing safety-critical autonomous driving systems.
Wang, WenweiHan, MuchenCao, Wanke
In the context of the accelerating development of an aging society, the inconvenient mobility of the elderly conflicts with the design of existing vehicles. The promotion and development of autonomous vehicles can provide solutions to this conflict to a certain extent. But existing autonomous vehicles lack a systematic age-friendly design. This study is based on a service design idea and employs the KJKANO hybrid model. The KJ method is used to construct a three-tier demand framework of “safety-function-emotion.” The KANO method is applied to identify the priority classification of each demand within the tiered framework. The study derives an aging-friendly design strategy for autonomous buses that prioritizes safety demands as the foundation, with functionality and emotional demands balanced accordingly. These strategies are then implemented in design practice. This study provides a user-centered systematic solution for the age-friendly design of autonomous buses, offering insights for research on age-friendly smart transportation.
Li, WangyanJi, Yuanyuan
To address the limitations of the traditional A* algorithm in lane-level navigation, we propose an autonomous vehicle path planning algorithm based on high-precision maps and an improved A* algorithm to ensure effective application in complex traffic environments. We construct a hierarchical high-precision map based on the Lanelet2 framework to achieve structured modeling of complex road environments. To address the adaptability issues of the A* algorithm in lane-level navigation, we propose optimization schemes, including heuristic function improvements, path segment division, and target point validity verification, to ensure that vehicles can autonomously change lanes on multi-lane roads. By combining dynamic programming (DP) and quadratic programming (QP), we ensure the safety and smoothness of the path. Simulation results demonstrate that the optimized algorithm enables smooth stopping and starting at traffic lights in structured road environments and autonomous lane changes on multi-lane roads. Compared to using DP alone, QP provides smoother and safer driving paths and exhibits superior obstacle avoidance performance in speed planning. This method effectively ensures the rationality of path planning in complex road environments while strictly adhering to traffic rules, thereby enhancing the safety and reliability of path planning.
Wang, SiyuZhou, RongShi, TianXu, ZhenZhao, Zhiguo
The technology of autonomous vehicles has become the bellwether for the next transportation evolution. Based on the system of level 5 autonomous vehicles (fully autonomous vehicles), there will be space released from the existing urban context, including linear space, nodular space, and intersected space because of the enhancement of transportation efficiency and organization. The study took Beijing as an example to explore the linear space releasing potential under fully Autonomous Vehicles system to provide a reference for future urban planning. Considering saturation flow rate, speed, parallel throughput, vehicle occupancy, and safe headway, we quantitatively analyzed the potential release from various types of urban roads. The results shows that the expressways, arterial roads, secondary arterial roads, and branch roads could release up to 50%, 66% 50%, and 75% of the road space, respectively. The study verified that fully AV system can release great amount of public space, and provided primary modes for the reformation of urban contexts in the future.
Ding, YufeiHou, Shuyu
In recent years, the automotive industry has faced increasing pressure to accelerate development cycles and reduce costs. Simultaneously, ride comfort standards have risen due to the ongoing integration of autonomous driving functionalities. Consequently, it has become essential to ensure that ride comfort attains a high degree of maturity at the very early stages of the automotive development process. This necessitates the establishment of objective criteria that enable the reliable estimation of subjective ride comfort, utilizing simulation-based assessment methods. This study introduces a methodological framework designed to systematically translate the manufacturer specific subjective perception and assessment of ride comfort into objective descriptions using a dynamic driving simulator. The framework is conceived as a generic approach, enabling the comprehensive application to a wide spectrum of subjective ride comfort phenomena, while being specifically optimized for the challenges of the automotive industry. Employing this framework facilitates the derivation of highly detailed, objective descriptions of subjective ride comfort evaluations, which promotes the achievement of advanced ride comfort maturity for new vehicles in early development phases and supports the overall enhancement of ride comfort. The exemplary application of the framework to a transient, one-dimensional ride comfort phenomenon demonstrates its capability to derive robust objective models from subjective evaluations conducted with professional test drivers in a dynamic driving simulator environment.
Stroesser, SimonZwosta, TobiasAngrick, ChristianNeubeck, JensWagner, Andreas
This article presents a data-driven pipeline for autonomous-vehicle (AV) safety testing. The pipeline integrates real-world traffic observations with model-guided scenario expansion and safety-metric evaluation to enable an end-to-end AV safety testing framework, demonstrated on a canonical highway scenario. The framework enhances test diversity, realism, and coverage by generating statistically informed variants of observed driving behaviors. Key parameters such as vehicle speed, trajectories, and headways are extracted from naturalistic data and used to train a probabilistic model of traffic dynamics. Scenario variants are sampled from this model and encoded as behavior trees (BTs) for modular, simulation-ready execution. Each scenario is simulated using a consistent AV control configuration, and safety metrics such as minimum safe distance violation, minimum safe distance factor, time to collision, and aggressive driving are applied to evaluate safety outcomes independently of system-specific tuning. A case study based on the highD dataset (110,000+ trajectories) demonstrates the framework’s ability to generate realistic and safety-relevant scenarios, providing an initial demonstration of pipeline feasibility and metric-based evaluation. This initial study is intentionally scoped to a single scenario class and a simplified parametric model to isolate and validate the end-to-end integration of the pipeline.
Elshenawy, MohamedAboudina, AyaAbdelmotaleb, AnharAmr, MariamEl-darieby, Mohamed
This article presents a cross-layer framework that integrates realistic vehicle-to-network-to-vehicle (V2N2V) delay characterization with a rigorous stability analysis of automated vehicle steering control. Both constant and network-induced time-varying delays modeled via deterministic bounds are addressed. For constant delays, delay-independent stability regions within the controller gain space are analytically derived. For time-varying delays with stochastic network origins, modeled using deterministic bounds, a refined Lyapunov–Krasovskii functional (LKF) incorporating augmented single- and double-integral terms is constructed. To establish delay-dependent linear matrix inequality (LMI) conditions, a reciprocally convex combination approach is employed to handle the delay interval partitioning, and the second-order Bessel–Legendre inequality is applied to tighten the integral quadratic bounds. The resulting LMI conditions explicitly capture the coupled effects of delay magnitude, delay variation rate, and control gains on closed-loop stability. Simulations of a lane-keeping scenario confirm that the predicted stability boundaries accurately match the closed-loop system behavior. Notably, incorporating a realistic time-varying V2N2V delay profile into the controller design reduces the lateral-state root-mean-square error (RMSE) by over 54% and decreases the settling time by a factor of 10 compared to designs relying on an average-delay assumption. However, high packet loss rates are shown to still induce residual oscillations due to information scarcity. Ultimately, these results elucidate delay-induced instability mechanisms and provide practical guidelines for designing delay-robust steering controllers for connected and automated vehicles.
Li, JialinLu, JianweiWei, HengAo, Di
For sustainability reasons, the automotive market is requesting 100% monomaterial noise treatments, particularly for the end-of-life recycling without any part separation operation. But also, OEMs require super light, highly performance insulating noise treatments for electric vehicles in order to extend vehicle autonomy. PP melt-blown fiber felts present good mono-material characteristics with very good absorption, but generally not so good insulation properties behind an airtight barrier due to lack of stiffness. Moreover, these PP melt-blown fiber felts are relatively expensive and not thermoformable, thus forcing them to be used as 2D die-cut parts behind existing hard or soft trims classically. The shown optimization approach proposes to return to 100% thermoformable recycled and recyclable PET formulations blending unusual coarse mechanical specific fibers, in order to optimize the viscothermal exchanges, while maintaining good mechanical properties, with microfibers for best dissipation properties bonded by bi-component fibers. The insulation properties obtained as poroelastic spring behind a barrier allow a weight reduction of -50% compared to cotton felt while being 1 dB better for the Insertion Loss values (2 dB compared to a flexible foam) and perform as well as best PP melt-blown fiber felts while being more competitive as well as thermoformable. It is possible to adjust the sound insulation properties, sound absorption and hardness (static compressibility) using optimal PET fibers formulations but also thanks to felt verticalization processes. These optimization levers will be illustrated in this paper.
Duval, ArnaudLei, LeiWilkinson, AlexandreDelinselle, Eric
Agricultural vehicles operating in rough environments experience increased fatigue damage accumulation, which may decrease machine safety and reliability. Autonomous agricultural machines offer an opportunity to incorporate fatigue damage considerations into path planning. This work investigates whether machine learning can predict fatigue damage to a tractor chassis using light detection and ranging (LiDAR)-based terrain features, vehicle speed, and rotational vehicle state data (e.g., triaxial angle, angular velocity, and angular acceleration). Fatigue damage was estimated using the Rupp filter and the Durability Transfer Concept. Following poor predictive performance of the machine learning models, an exploratory analysis of damage histograms, dominant frequency, and acceleration magnitude was performed. Results indicated that most estimated fatigue damage occurred in the 0–2 Hz band, which coincides with the frequency range of terrain-induced acceleration. On-road driving led to the greatest fatigue damage, potentially due to the harder driving surface and increased vehicle speed. Differences between root mean square (RMS) acceleration magnitude and fatigue damage indicate that isolated high-magnitude events may have contributed to increased estimated fatigue damage. Several suggestions for future development were identified. Identification of the endurance limit of the tractor chassis will permit the removal of nondamaging events, improving label accuracy. Furthermore, the presence of a front-loader implement may have impacted chassis acceleration. Thus, a comprehensive dataset with multiple implement configurations is needed to determine the influence of implement configuration on dynamics and resultant damage.
Govers, Megan EmilyHamilton-Wright, AndrewHassan, MarwanOliver, Michele L.
Large language models (LLMs) have shown remarkable capabilities for perceiving driving environments and making interpretable, logical decisions for autonomous driving. However, their potential for more comprehensive driving strategies, especially concerning energy efficiency, remains underexplored. Most existing studies primarily focus on driving safety, which may inadvertently increase energy consumption. To address this issue, this study explores the use of LLMs as high-level controllers to jointly optimize driving safety and energy efficiency. A textual prompt is designed for the LLM, incorporating few-shot examples that describe scenarios, states, and actions. The LLM processes the scenario and state prompts describing the surrounding traffic environment. It generates a high-level control signal, which is then translated into low-level vehicle motion commands in a high-fidelity traffic simulator with realistic physics, vehicle dynamics, road slopes, and network topology. Experiments in campus-scale digital twin car-following scenarios demonstrate that the proposed LLM-based framework achieves an average reduction of 4.16% in energy consumption compared to the reinforcement learning paradigm, while maintaining driving safety and providing interpretable high-level decision-making. This study highlights the potential of LLMs for longitudinal eco-driving applications under the evaluated simulation settings, extending previous LLM-based autonomous driving research that primarily focused on safety to also consider energy efficiency.
Wang, HaoyuLi, ZhenningWang, SiyingZhou, ZijingZhang, XiangYang, ZhifengOu, Shiqi (Shawn)Qi, Hao
This study examines the involvement of authorities in the development processes of aviation and automotive industries by comparing the depth, frequency, and stages of their engagement. The background of this work is an ongoing research initiative focused on transferring methods from aviation to automotive. The method used in this study is an investigation of best practices across both industries. Based on this investigation, two proposals were developed for managing complex technologies, such as autonomous systems. Both proposals advocate for increased authority involvement, particularly during the early stages of projects. One proposal recommends making this enhanced involvement mandatory, while the other suggests it as a guideline rather than a requirement. To assess the benefits of these proposals, a human-input–based feasibility quantification method was applied. This method assesses feasibility on a scale from 0 to 10, where 0 represents the lowest score, 5 is neutral, and 10 is the highest. The results indicate that the proposal recommending enhanced authority involvement achieved a score of 5.79, whereas the proposal mandating it scored 4.54. The conclusion of this study is that increasing authority involvement offers slight benefits when implemented as a recommendation rather than as a mandatory requirement.
Akkus, YusufAnnighöfer, Björn
Distributed drive electric vehicles (DDEVs) provide enhanced maneuverability through independent wheel torque control, but coordinating precise path tracking with lateral stability remains challenging under aggressive driving conditions. This paper presents a coordinated control strategy that integrates model predictive control (MPC) for path tracking with a proportional gain controller for stability regulation. The proposed framework adopts a hierarchical design. The path tracking control leverages MPC to compute front steering commands while accounting for vehicle dynamics and preview errors. The stability adjustment uses dual proportional gain controllers to generate an additional yaw moment, which is adaptively balanced through a phase plane coordination mechanism, enhancing yaw stability during path tracking. The generated yaw moment is subsequently distributed to individual in-wheel motors with an optimization torque allocation method, respecting tire force limitations. The effectiveness of the proposed strategy is validated with hardware-in-the-loop (HIL) experiments under a double lane change maneuver. Results show that the coordinated approach improves path following and maintains yaw stability more effectively than conventional methods.
He, YangZhu, YuzhengGuo, RuixinZhu, YueyingXing, ChaoLiu, ShuangxiLin, Yier
In the two months since Microvision bought Luminar and acquired key tech and talent, the sensor company has been busy. In that time, they've merged key lidar units from each company and created a perception software stack to run it in a convincing demo of its ADAS and autonomous capabilities. The company is also pushing innovative lidar tech into the defense drone and antidrone markets, already working with a German defense supplier that works with NATO member countries.
Clonts, Chris
Automated aircraft parking systems enhance airport ground operations by enabling precise and autonomous docking of aircraft at gates. These systems reduce turnaround time, minimize human error, and optimize apron space through real-time object detection, obstacle avoidance, and dynamic path planning. Unlike fixed guided-path methods, the proposed system adapts to congestion and environmental conditions such as low visibility, ensuring safety and efficient maneuvering. Validation through simulation demonstrates the system’s potential to improve operational resilience and support scalable automation in future airport infrastructure.
Penugonda, Navya SunainaEdiga, Venkatadiwakar Goud
Aircraft Maintenance, Repair, and Overhaul (MRO) operations are highly complex, involving coordination among multiple stakeholders including airlines, MRO providers, OEMs, and regulatory authorities. A significant challenge in this space is managing unplanned events such as Aircraft on Ground (AOG) conditions, where delays can lead to major financial losses to airlines and safety risks. Engineers must quickly diagnose the damage, evaluate compliance against regulatory limits, coordinate with OEMs, and make critical decisions—all while navigating a fragmented ecosystem of disconnected systems, diverse document types, and time-sensitive processes. This paper presents a real-world, intelligent MRO solution that addresses these challenges through the use of Agentic AI and context engineering. The system is designed to automate and augment key MRO workflows such as damage detection, repair pathway selection, compliance verification, and supplier coordination. At its core, the solution is powered by a set of autonomous agents—each responsible for specific tasks like interpreting repair manuals, evaluating damage severity, or communicating with OEM portals. A key innovation of this system is its use of context engineering, which enables agents to share a unified, real-time view of the aircraft condition, document references, decisions made, and deadlines involved. This shared memory—dynamically updated using modern data stores and retrieval systems—ensures that agents and human experts operate with full situational awareness. The solution facilitates measurable improvements in reducing aircraft downtime, speeding up OEM coordination, and ensuring real-time continuous regulatory compliance. Engineers can take faster, more informed decisions with confidence, while human-in-the-loop oversight was preserved for critical steps such as compliance sign-off and final approvals. Overall, this intelligent MRO system transforms static, document-heavy processes into a dynamic, context-aware workflow. It brings together agent collaboration, regulatory alignment, and real-time information flow to solve one of the most pressing operational problems in aviation today. By improving inspection-to-repair cycles, enhancing SLA adherence, and enabling traceable, data-driven decision-making, this work lays the foundation for the next generation of digital MRO ecosystems that are efficient, safe, and scalable.
Abburu, SunithaG.V.V., Ravi KumarPoovalingam, SundaresanVaderahobli, Devaraja Holla
This paper addresses the critical challenge of fault-tolerant control in autonomous multi-copters, particularly under conditions of one or two rotor failures a scenario that often leads to severe instability and a complete loss of directional control due to unbalanced torque and resultant autorotation. Existing advanced control strategies, including optimal approaches such as LQR, typically require precise system modeling and state estimation, which are difficult to achieve in real-world, dynamic failure scenarios. Alternative methods like fuzzy logic, sliding mode control, and gain-scheduling either lack robust generalization or are impractical for enumerating all possible failure cases. In this work, a hybrid control framework integrating Physics Informed Neural Networks (PINN) with a standard PID controller is proposed for fault-tolerant operation of autonomous multi-copters subject to multiple actuator failures. PINNs incorporate governing physical laws as regularization in their loss functions, allowing them to learn optimal counter-torque actions and thrust balancing necessary to arrest autorotation and stabilize flight, despite limited training data and uncertainty in failure conditions. The calculated moments and thrust commands are executed via a robust PID scheme, enabling reliable real-time implementation and minimizing residual oscillations. This hybrid control architecture demonstrates significant potential to enhance the resilience and operational safety of autonomous multi-copters during unexpected motor failures. By leveraging PINN’s physics-based generalization and PID’s consistent execution, the proposed method offers an adaptive, model-agnostic approach for maintaining stable flight and directional control under severe actuator faults, with implications for next-generation fault-tolerant UAV systems deployed in complex environments.
Charapalle, SamruddhiVenugopalan, NandagopalanNerkundram Muralidharan, ArunSundararaj, Laveen
Researchers from CompPair and the European Space Agency have developed a new composite material for spacecraft with an embedded healing agent. European Space Agency, Paris, France Healable spacecraft structures could soon be possible thanks to cutting-edge composite technology. Swiss companies CompPair and CSEM, and Belgian company Com&Sens have partnered with the European Space Agency (ESA) to modify their self-healing carbon fiber product for use in space transportation. Project Cassandra - an abbreviation for Composite Autonomous Sensing and Repair - includes sensors and a heating element within a composite carbon-fiber material, allowing spacecraft to autonomously repair initial stages of damage.
The uncrewed aerial vehicle (UAV) market is advancing at extraordinary speed, reshaping both commercial and defense aviation. From small tactical systems operating at the edge of the battlefield to high-altitude uncrewed platforms conducting strategic surveillance, UAVs are now critical assets across a wide range of mission environments. Their capabilities continue to expand - carrying more sensors, flying longer missions, and navigating more contested environments. Yet this rapid innovation brings with it growing engineering pressure. UAVs are expected to be lighter, more autonomous, more modular, and more adaptable, all while maintaining near-flawless reliability. This is where heritage becomes decisive. In an era that rewards speed, heritage provides the hard-earned engineering wisdom that ensures systems do not just fly but perform predictably, repeatedly, and safely under real-world conditions.
Trajectory tracking control is a core technology in intelligent vehicle autonomous driving systems, directly influencing both driving safety and control accuracy. To overcome the limitations of traditional model predictive control (MPC) in real-time performance under complex operating conditions, as well as the limited robustness of linear quadratic regulators (LQR) against system uncertainties, this article proposes a hybrid iterative LQR–MPC (ILQR-MPC) control strategy. First, a dynamic model of the intelligent vehicle is developed to capture its behavior during high-speed driving and cornering. Next, an ILQR-MPC hybrid framework is designed. By exploiting the rapid iterative optimization capabilities of the ILQR algorithm, an initial control sequence is generated for the MPC, thereby reducing the computational load during MPC’s online rolling-horizon optimization. This approach preserves MPC’s advantages in handling constraints and maintaining robustness against parameter variations and external disturbances. Finally, joint simulations using MATLAB/Simulink and CarSim are conducted to evaluate the proposed approach against conventional MPC under standard road conditions, curved sections, and sudden changes in road friction. The results show that the ILQR-MPC strategy reduces trajectory tracking errors, shortens computational time, and maintains excellent stability and robustness under complex operating conditions.
Lai, FeiSun, JunhaoHuang, Chaoqun
Autonomous Vehicles (AVs) offer unprecedented opportunities to design control strategies that could be able to simultaneously enhance safety, performance, user experience, time efficiency, and the environmental impact of mobility. However, as automation levels increase, a paradigm shift becomes not only necessary but imperative: the integration of human needs into mobility objectives. This includes not only traditional comfort considerations but also minimizing Motion Sickness (MS), a largely under-explored challenge in control strategy design. In recent literature, several methodologies for modeling and mitigating MS have been proposed, yet their integration into vehicle control logics remains limited, often restricted to isolated and specific case studies, with the research area largely unexplored, particularly with respect to the generalization of the proposed methods. This work introduces a theoretically grounded multi-objective Nonlinear Model Predictive Control (NMPC) framework for coupled vehicle–passenger systems, featuring a novel prediction horizon optimization methodology and adaptive conflict resolution strategies for heterogeneous performance metrics to mitigate motion-induced discomfort while ensuring accurate path tracking. Human-centric control design is pursued by embedding increasingly complex vehicle models and MS metrics, further addressing the trade-off between model fidelity and computational feasibility, and introducing a methodological standpoint for selecting the optimal prediction horizon in the presence of heterogeneous and conflicting control objectives, an aspect often overlooked in current literature. An experimental campaign supports model calibration and validation, while multi-scenario simulations demonstrate the framework’s ability to balance tracking performance, computational efficiency, and passenger comfort.
Ponticelli, LorenzoBottiglione, FrancescoRini, GabrieleTimpone, FrancescoSakhnevych, Aleksandr
In response to the problems of urban traffic congestion and the limited expansion of infrastructure, this paper conducts two core research focusing on the intelligent chassis system of split-type flying vehicle. Firstly, an autonomous navigation strategy for the intelligent chassis module is proposed based on chassis module Navigation 2 architecture, which fuses LIDAR and IMU positioning to plan paths using the A* global planning algorithm on a global cost map, and update the local cost map in real time with sensor data. It is orchestrated by the BT Navigator using a behavior tree, with failures handled by the Recovery Server, to achieve autonomous driving across multiple waypoints. In simulation and closed-field experiments, the system can stably reach the preset target points. The positioning accuracy and trajectory tracking performance can meet the design requirements. Secondly, a mechanical slide rail-type docking structure adapted to the split flying vehicle architecture is designed. Deformation analysis under the representative working conditions are evaluated through finite element software. The test results show that the maximum deformation of this docking structure under typical load is significantly lower than the docking tolerance and positioning repeatability requirements. The structural stiffness and stability meet the design indicators. The above work indicates that the proposed autonomous navigation strategy and the docking structure for the intelligent chassis can effectively support the modular operation of “air trunk & ground terminal” mode, providing a scientific basis for the functional integration and system reliability research of split-type flying vehicles.
Zhao, WenyuShi, QinJiang, CongHe, Zejia
The aging of the population has been a key issue worldwide, with mobility and fall of the elderly an important problem to be solved. In this paper, we propose an elderly mobility assist system based on the intelligent power-assisted device consisting of an assistive cane and an intelligent companion. It has the functions of standing support after falling, daily support and on-site rest. The assistive cane adopts a two-stage expansion mechanism of crank and slider structure, which forms a stable triangular support after unfolding, so that the patient can stand safely. The intelligent companion platform is driven by drive wheels, equipped with pushrod motors and vacuum suction devices, it can automatically approach the user and form an stable support column when the cane is in the out-of reach range; the control system is designed by combining microcontroller, camera object recognition, wristband remote control, to realize automatic steering and autonomous navigation at differential speed. The overall design satisfies the requirements of safety and strength through mechanical verification and stress analysis. The proposed system can help the elderly people to recover from falls better and enhance their independence and safety in their daily walks.
Yu, ChenxiWang, LongyiZhu, HuayunDong, YanMi, RuixueZhu, Lihong
End-to-end autonomous driving in urban environments faces three core challenges. First, camera and LiDAR sensor heterogeneity causes cross-modal perception inconsistencies and sensor fusion instability. Second, diffusion models suffer from training instability due to scale variance and distribution changes, which limits generalization. Third, traditional trajectory decoders lack structured interaction with semantic elements, thereby undermining planning rationality. To address these issues, CMFPNet introduces an integrated framework with three key modules. The HGCF-Backbone integrates LiDAR and camera features using channel focus, deformable cross-focus, and state space modeling to enhance semantic alignment. The NST module maps physical trajectories to normalized space, employing truncated diffusion sampling for stable generation in just 2–4 steps. The NDA models trajectory generation as a semantic narrative, utilizing a six-stage semantic attention flow incorporating BEV context, interactive dynamics, and self-states. Experiments on the NAVSIM dataset demonstrate CMFP Net’s superiority over existing baselines, showing outstanding generalization and trajectory stability in challenging scenarios. Notably, the truncated sampling strategy achieves an 8–10× acceleration during inference while maintaining decision accuracy and reducing computational costs. CMFPNet provides a scalable, semantically consistent solution for diffusion-based autonomous driving with significant potential in both research and practical deployment.
Qu, YanweiMo, Hangjie
Identifying driving heterogeneity is critical for enhancing the strategy learning capabilities of autonomous driving systems, as well as improving their safety and efficiency. This research proposes a novel driving heterogeneity identification framework. The framework consists of three core processes: action phase extraction, action relationship modeling, and behavior heterogeneity identification. First, a rule-based segmentation method is employed to systematically decode and interpret the inherent variations in human driving behavior. Subsequently, an action relationship modeling method is introduced to characterize the temporal relations between the acquired action phases. Finally, to mitigate the inaccurate identification caused by the sparse distribution of critical driving events in long-sequence data, a semantic encoding method is applied to remap the driving behavior space. Experimental results on the Lyft level-5 dataset validate the effectiveness of the proposed framework, which outperforms multiple traditional clustering algorithms. This demonstrates its significant potential to enhance behavior detection and learning in personalized advanced driver-assistance systems (ADAS) and advanced autonomous vehicle (AV) design.
Yin, HuiZhang, QinyaoLi, XiaojianMo, Hangjie
The rapid development of autonomous driving technology has brought emerging opportunities to optimize the omnidirectional vehicle driving performance. However, its compliance with driving habits directly determines its social acceptance. Therefore, how to balance consistency between performance improvement and driving habits has become an important bottleneck restricting the rapid promotion of autonomous driving technology. Manual driving vehicles mostly focus on the safety of both longitudinal and lateral movements, and cannot cope with the vertical movement, let alone the performance of economy, comfort, and efficiency. In this context, this paper proposes an anthropomorphic trajectory optimization method incorporating vehicle omnidirectional dynamic characteristics and corresponding driving habits. Firstly, this paper explores vehicle dynamic characteristics in longitudinal, lateral, and vertical directions, and reveals the coupling effect of motion states during driving. Furthermore, the featured function of anthropomorphic driving is constructed by the driving habit patterns related to the accelerator pedal, deceleration pedal, and steering wheel. Then, a comprehensive trajectory optimization method that considers safety, economy, comfort, and efficiency is constructed to improve the driving performance while ensuring social acceptance. Finally, the sensitivity of method weights and the adaptability toward the real world are verified by the case studies and discussions. The results indicate that the proposed method can fully utilize the optimization potential of autonomous driving technology in the omnidirectional performance, and effectively assist the intelligent and automated development of road traffic systems.
Liao, PengZhang, DefengNing, DonghongLi, SijiaWang, Tao
Nasa Tech Briefs: May 202626AERP055/7/2026
How Machina Labs is Reshaping Defense Manufacturing with AI-Driven 7-Axis Robotics Engineering at the Speed of Conflict: The New Era of Defense Testing MyDefence Expands Production, Validation of Wearable and Mobile Counter-UAS Systems Keeping Pace with Changes in Defense Technology: Why Embedded Systems Must Deliver Agility, Resilience, and Endurance From Autonomous Vehicles to Ship-to-Shore to: Designing 60 GHz Networks for Tactical Applications Defense Research Program Developing Tactical Clocks for GPS-Free Navigation The Drone Equation: Proportional Response to the UAS Threat High-end interceptors. Low-end threats. The answer isn't bigger missiles but smarter integration. Virginia Tech Experts Accelerate Skydio Drone Flights Over People and Vehicles Researchers recently helped Skydio, the leading U.S. drone manufacturer, demonstrate compliance to the Federal Aviation Administration's rules for safe flights over people and vehicles. Teaching Robots to Fly Like Birds Rutgers researchers replace motors with smart materials in an innovative approach to flight. RPI Researchers Harness Agentic AI for Smarter, Faster Aerospace Design Funding from Google and the U.S. Department of Energy helped a team of researchers develop an assortment of agentic AI-enabled tools to help optimize traditional aerospace design processes. Can Multi-Fingered Robots Transform Shipboard Operations and Autonomous Maintenance? USC Viterbi researcher received Office of Naval Research's Young Investigator Program award with Study on dexterous robotics. Can Multi-Fingered Robots Transform Shipboard Operations and Autonomous Maintenance? USC Viterbi researcher received Office of Naval Research's Young Investigator Program award with Study on dexterous robotics.
Autonomous vehicles exhibit extremely strong nonlinearity during drift. However, existing autonomous drift algorithms often neglect previewed path curvature and offer only limited consideration of road surface uncertainty because of the influence of vehicle nonlinear dynamics, which can affect tracking accuracy and robustness of drift control. To solve these problems, this study proposes a robust optimal drift control framework based on curvature preview. First, a preview vehicle kinematic model is constructed, and a preview model predictive control path-tracking controller that considers the forthcoming curvature is designed. Through the analysis of equilibrium points with additional yaw moment, a robust optimal drift controller is developed, which employs a three-degrees-of-freedom vehicle model with an additional yaw moment. This controller adopts integral sliding mode control with a super-twisting algorithm (STA) and exhibits good stability, which is verified through Lyapunov analysis. The proposed control algorithm is validated through hardware-in-the-loop experiments. The experimental results demonstrate that the proposed method significantly improves path-tracking accuracy and robustness under uncertain road surface conditions, thereby providing an effective control solution for drift-based path-tracking maneuvers.
Gan, YurunSong, ZiyuGu, TongtongDing, HaitaoXu, NanZhang, Jianwei
This paper presents enhancements to the supervisory controller developed for the National Research Council Canada's Bell 412 autonomous helicopter. Building on a Discrete Event System Specification (DEVS)-based framework, the updated Supervisor introduces two new operational modes-Knobs Mode and Sticks Mode-and a structured approach for managing transitions between them and the existing modes. Drawing inspiration from NASA's Flight Guidance System philosophy, the proposed design emphasizes consistency, scalability, and flexibility in handling multiple autonomy modes. Implementation results demonstrate the effectiveness of the updated architecture in supporting future expansion of autonomous mission operations in complex and dynamic environments.
Winstanley, CurtisBorshchova, Iryna
Developing a comprehensive autonomy solution for the Army's current and future aircraft fleet requires a robust computational and perception capability for decision-making across the entire flight envelope without a pilot. This also requires a flight control system and infrastructure capable of executing autonomous decisions in complex mission environments. Ongoing development of automation and autonomy, utilizing a wide range of perception sensors, has been conducted on platforms such as Sikorsky's S-70 and the Army's UH-60Mx aircraft. This work builds upon previous efforts and leverages ongoing collaborations with industry, the Department of War (DoW), and the Defense Advanced Research Projects Agency (DARPA) to advance autonomous capabilities for both optionally piloted and uncrewed aircraft.
Arterburn, DavidPolycarpe, CauvinOtt, Carl
Cargo-focused autonomous Vertical Takeoff and Landing (VTOL) operations are advancing toward commercialization significantly faster than passenger missions due to a confluence of regulatory pragmatism, technical readiness, and market economics. This paper examines the commercial potential of integrating Artificial Intelligence (AI) and Beyond Visual Line of Sight (BVLOS) control into an Uncrewed VTOL Air Cargo (AI-UVAC) vehicle for dual use military and commercial logistics applications. The Piasecki KARGO II was designed specifically for these missions and is used as the basis for evaluating this capability. This AI-UVAC concept has useful commercial application in the "mid-weight Less than Truck Load (LTL)" freight market for middle-mile delivery of time-sensitive cargo in infrastructure-constrained markets. To validate the advantages of commercial freight orchestration, a multiphase pilot program is conducted to measure the effectiveness of the LogistiWerx Generative AI-Powered Freight Logistics Orchestration platform as integrated in the KARGO II uncrewed BVLOS VTOL developed by Piasecki Aircraft Corporation (PiAC).
Stanzione, KaydonCavazzoni, MarcoPiasecki, John
The design, testing, and analysis of a Guided Autorotative Delivery System (GADS) for suppression of incipient wildfires is described. The GADS consists of an unpowered 1 m diameter rotor, a control unit, and a payload of 2.2 kg of fire suppressant powder. On release from a fixed-wing UAV, the rotor passively deploys and enters autorotation, decelerating the payload and allowing precise delivery of the suppressant using cyclic pitch control. A numerical model of the system was developed to calculate the trajectory of the GADS during rotor deployment and descent, in the presence of ambient wind and cyclic pitch inputs. A reduced-scale model of the rotor was tested in a wind tunnel, and an uncontrolled full-scale, 1.5 kg prototype of the GADS was fabricated and tested by dropping from a hovering quadcopter as well as a fixed-wing UAV. The full-scale drop experiments validated the deployment and autorotation stability of the system, and demonstrated that the GADS maintains descent velocities suitable for incipient fire suppression (≈ 5 m/s). Numerical predictions indicate that the GADS descent trajectory can be controlled with cyclic pitch in an ambient crosswind of at least 5 m/s (10 kts). Measurements captured during the drop tests using onboard instrumentation show good qualitative agreement with numerical predictions. Future work will include drop tests with remotely controlled cyclic pitch, followed by fully autonomous controlled descent. The study establishes design guidelines for guided autorotative systems and illustrates their potential for scalable UAV-based wildfire suppression or emergency response.
Chadha, JiaJain, RheaSakamuri, SivaThomas, ThomasSirohi, Jayant
This paper presents the results of procedurally generating urban environments, characterising them and simulating a UAS flying missions within it. It shows that varied and practically inifinite new scenarios can be generated for testing UAS. It further shows how these methods can be intergrated into a wider testing framework for the robust testing of UAS.
Harris, IsabellePage, Vincent
This paper presents a spatio-temporal graph neural network (STGNN) centric approach to enable heterogeneous agents to collaborate and cooperate for different types of missions. The STGNN-centric approach and corresponding autonomy are encapsulated in the Advanced Graph-enabled Network Technology for Collaborative Autonomous Agents (AGENTCA) technology. Various decentralized and distributed control architectures are reported in the literature, but in some instances these approaches do not leverage the inherent graph network which can increase scalability to larger teams and algorithmic efficiency. Specifically, in this paper advances in artificial intelligence are leveraged to parameterize and encode optimal, or nearly optimal, swarm control techniques. For this work, the team focused on developing a diffusion-based STGNN swarm controller using imitation learning. An expert, centralized swarm control law was used to guide the STGNN during the learning process. The STGNN controller enables the swarm to follow a leader while avoiding static and dynamic obstacles and maintaining a desired separation distance from neighbors and obstacles. The approach is demonstrated in simulation with hundreds of agents and in flight tests with up to thirteen test vehicles.
Cooper, JaredLu, Chang-TienChen, SijiCarson, AndrewPeters, AndrewOlowin, AaronEnnasr, OsamaLichter, Matthew
This paper develops and tests a feature-based autonomous landing system for vertical lift aircraft on stochastically moving ship decks, under degraded visual conditions. The system is tested with a custom-built quadrotor on a six-degree-of-freedom 1.5-ton Stewart platform reproducing stochastic motions up to Sea State typical of DDG-51-class ships. Experiments began with nominal conditions, followed by a stepwise degradation of deck features through occlusion, low illumination, water distortion, and glare. The vision algorithm tracked the platform and achieved landing across all scenarios, with tracking errors of up to 14% of the vehicle footprint, and up to 2.2° of pitch and roll. Overall, it demonstrated the ability to land in a GPS/Lidar-denied, difficult environment with on-board vision alone, achieving deterministic, repeatable results.
Basak, KumardipDatta, AnubhavChopra, Inderjit
This paper presents the results of a flight test effort examining fully autonomous shipboard operations for small unmanned aerial vehicles (UAVs). Experiments were conducted at the Maneuvering and Seakeeping Basin (MASK) located at the Naval Surface Warfare Center, Carderock Division using custom-built quadrotor UAVs landing on an unmanned surface vessel (USV). These tests build upon previous ship landing algorithm testing in order to expand the envelope of operations and be more representative of a real-world mission. Several new flight modes were implemented, including takeoff and pattern flying, and a finite state machine was developed to allow smooth and autonomous transition between the different flight modes. The results from testing show smoothly executed missions both in still water and in the presence of waves. However, it was found that the initial conditions for the command filters in the position controller needed to be carefully selected. Without the correct initial conditions, discontinuities in the commands were seen when switching between modes that used the command filters and modes that bypassed them. The results of this work will help bridge the gap between ship landing-specific research and real-world applications encompassing multiple flight modes.
Jue, AndrewSydney, AnishLangelaan, JackHorn, JosephArnold, DariusZimmerschied, DarioPrewitt, Jack
Deep learning (DL) models have attained state-of-the-art performance in numerous fields. Nevertheless, for certain real-world applications, existing models encounter diverse challenges, ranging from a lack of generability to new data to issues of scalability and overfitting. In this context, integrating information extracted from different modalities holds promise as a potential solution to alleviate these challenges. This paper introduces MAVEN, a multimodal deep-learning framework for long-range atmospheric visibility estimation. Using multimodal deep learning, MAVEN fuses various modalities to estimate long-range atmospheric visibility. These modalities include RGB imagery, Edge Map, Entropy Map, Depth Map, and Normal Surface Map. Results show that in contrast to single-modality RGB, which achieves only 87.92% accuracy, multimodal deep learning models achieve an accuracy of over 96%. This significant improvement highlights the potential of multimodal approaches to enhance the accuracy and reliability of atmospheric visibility estimation, which is crucial for improving safety in applications such as aviation, maritime navigation, and autonomous vehicles. By addressing challenges such as data variability, environmental factors, and the inherent complexity of atmospheric conditions, MAVEN contributes to more reliable and robust visibility estimation systems, thereby enhancing safety and operational efficiency in critical environments.
Khelifi, AmineJohnson, CharlesBouaynaya, NidhalCarannante, GiuseppinaBouhsine, Taha
Items per page:
1 – 50 of 3215