Papers

Hypersonic glide vehicles (HGVs) operate in challenging flight regimes characterized by strong nonlinearities in actuation and stringent physical constraints. These include state-dependent actuator limitations, asymmetric control bounds, and thermal loads that vary with maneuvering conditions. This paper introduces an iterative control allocation method to address these challenges in real time. The proposed algorithm searches for control inputs that achieve the desired moment commands while respecting constraints on input magnitude and rate. For slender HGV configurations, thermal loads and drag generation are strongly correlated-lower drag typically results in reduced surface heating. By embedding drag-sensitive soft constraints, the method improves energy efficiency and implicitly reduces surface temperatures, lowering the vehicle's infrared signature. These features are particularly advantageous for long-range military operations that require low observability. The approach is demonstrated using the DLR's Generic Hypersonic Glide Vehicle 2 (GHGV-2) simulation model. The results confirm the method's effectiveness in maintaining control authority under realistic, constrained flight conditions.

Increase in the number of space exploration missions has led to the accumulation of space debris, posing risk of collision with the operational satellites. Addressing this challenge is crucial for the sustainability of space operations. To plan a safe trajectory in the presence of moving space debris, an integrated approach of artificial potential field and sliding mode controller is proposed and implemented in this paper. The relative 6-DOF kinematics and dynamics of the spacecraft is modelled in the framework of geometric mechanics with the relative configuration expressed through exponential coordinates. Various collision avoidance guidance algorithms have been proposed in the literature but the Artificial Potential Field guidance algorithm is computationally efficient and enables real-time path adjustments to avoid collision with obstacles. However, it is prone to issues such as local minima. In literature, local minima issue is typically avoided by either redefining the potential function such as adding vorticity or by employing search techniques which are computationally expensive. To address these challenges, a physics-informed APF is proposed in this paper where Hamiltonian mechanics is used instead of the traditional Newtonian mechanics-based approach. In this approach, instead of relying on attractive and repulsive forces for path planning, the Hamiltonian approach uses the potential field to define a path of minimum potential. Additionally, to track the desired trajectory planned by the guidance algorithm within a fixed-time frame, a non-singular fixed-time sliding mode controller (FTSMC) is used. The proposed fixed-time sliding surface not only ensures fixed-time convergence of system states but also guarantees the global stability of the closed-loop system without singularity. The simulation results presented support the claims made.

Audio-based lyrics matching can be an appealing alternative to other content-based retrieval approaches, but existing methods often suffer from limited reproducibility and inconsistent baselines. In this work, we introduce WEALY, a fully reproducible pipeline that leverages Whisper decoder embeddings for lyrics matching tasks. WEALY establishes robust and transparent baselines, while also exploring multimodal extensions that integrate textual and acoustic features. Through extensive experiments on standard datasets, we demonstrate that WEALY achieves a performance comparable to state-of-the-art methods that lack reproducibility. In addition, we provide ablation studies and analyses on language robustness, loss functions, and embedding strategies. This work contributes a reliable benchmark for future research, and underscores the potential of speech technologies for music information retrieval tasks.

Distributed energy resources (DERs) are transforming power networks, challenging traditional operational methods, and requiring new coordination mechanisms. To address this challenge, this paper introduces SecuLEx (Secure Limit Exchange), a new market-based paradigm to allocate power injection and withdrawal limits that guarantee network security during time periods, called dynamic operating envelopes (DOEs). Under this paradigm, distribution system operators (DSOs) assign initial DOEs to customers. These limits can be exchanged afterward through a market, allowing customers to reallocate them according to their needs while ensuring network operational constraints. We formalize SecuLEx and illustrate DOE allocation and market exchanges on a small-scale low-voltage (LV) network, demonstrating that both procedures are computationally tractable. In this example, SecuLEx reduces renewable curtailment and improves grid utilization and social welfare compared to traditional approaches.

Attitude and heading reference systems (AHRS) play a central role in autonomous navigation systems on land, air and maritime platforms. AHRS utilize inertial sensor measurements to estimate platform orientation. In recent years, there has been increasing interest in multiple inertial measurement units (MIMU) arrays to improve navigation accuracy and robustness. A particularly challenging MIMU implementation is the gyro-free (GF) configuration, in which angular velocity is derived solely from accelerometer measurements. While the GF configurations have multiple benefits, including outlier detection and in angular acceleration measurements, their main drawbacks are inherent instability and an increased divergence rate. To address these shortcomings, we introduce a novel symmetrical MIMU formulation, in which the IMUs are arranged in symmetric diagonal pairs to decouple linear and rotational acceleration components. To this end, we derive the theoretical foundations for the symmetrical MIMU formulation of the GF equations, develop a nonlinear least squares estimation process, and integrate statistical hypothesis testing into an AHRS error-state extended Kalman filter. We validate our approach using real-world datasets containing 85 minutes of navigation data recorded on both airborne and land platforms. Our results demonstrated a 30% average reduction in attitude estimation errors, rotation detection accuracy exceeding 95% improvement, and significantly improved stability compared to a standard GF implementation. These results enable reliable GF navigation in applications where gyroscopes are unavailable, unreliable, or energy-constrained. Common examples include miniature platforms, computational-constraint platforms, and long-endurance marine platforms.

In the beyond fifth-generation (B5G) and upcoming sixth-generation (6G) wireless communication systems, millimeter (mmWave) wave technology is a promising solution for offering additional bandwidth resources and mitigating spectrum congestion. Beam tracking is an essential procedure for providing reliable communication services in the mmWave communication system, with the challenge of providing consistent and accurate tracking performance. In this study, we introduce a low-overhead beam tracking algorithm based on channel charting, which significantly reduces beam scanning times during the tracking process. By projecting the beam information to the channel chart, the beam tracking problem is transformed into the acquisition of the beam cluster in the channel chart. Leveraging contrastive learning, the proposed channel chart projects high-dimensional channel state information into a low-dimensional feature space that preserves spatial proximities. Using a dynamic candidate beam acquisition strategy, the complexity of our beam tracking algorithm is significantly reduced. The proposed algorithm significantly reduces scanning complexity while maintaining high prediction accuracy, achieving an accuracy of 98.27% in simulation environments. Compared to existing methods, the proposed method can reduce beam scanning times by up to 55.9%. In addition, we also performed field tests, and the measured results demonstrated excellent communication quality during mobility.

The substantial communication resources consumed by conventional pilot-based channel sounding impose an unsustainable overhead, presenting a critical scalability challenge for the future 6G networks characterized by massive channel dimensions, ultra-wide bandwidth, and dense user deployments. As a generalization of radio map, channel knowledge map (CKM) offers a paradigm shift, enabling access to location-tagged channel information without exhaustive measurements. To fully utilize the power of CKM, this work highlights the necessity of leveraging three-dimensional (3D) environmental information, beyond conventional two-dimensional (2D) visual representations, to construct high-precision CKMs. Specifically, we present a novel framework that integrates 3D point clouds into CKM construction through a hybrid model- and data-driven approach, with extensive case studies in real-world scenarios. The experimental results demonstrate the potential for constructing precise CKMs based on 3D environments enhanced with semantic understanding, together with their applications in the next-generation wireless communications. We also release a real-world dataset of measured channel paired with high-resolution 3D environmental data to support future research and validation.

New-generation space missions require satellites to carry substantial amounts of liquid propellant, making it essential to analyse the coupled control-structure-propellant dynamics in detail. While Computational Fluid Dynamics (CFD) offers high-fidelity predictions, its computational cost limits its use in iterative design. Equivalent Mechanical Models (EMMs) provide a faster alternative, though their predictive performance, especially in closed-loop scenarios, remains largely unexplored. This work presents a comparative analysis of a spacecraft under feedback control, using both CFD and a reduced-order sloshing model. Results show good agreement, validating the simplified model for the manoeuvrer considered. This validation enables efficient sensitivity and stability studies, offering a practical tool for early-stage spacecraft design.

This paper presents a general framework for solving the control allocation problem (CAP) in thrust-vector controlled rigid-bodies with an arbitrary number of thrusters. Two novel solutions are proposed: a closed-form, Lipschitz continuous mapping that ensures smooth actuator orientation references, and a convex optimization formulation capable of handling practical actuator constraints such as thrust saturation and angular rate limits. Both methods leverage the nullspace structure of the allocation mapping to perform singularity avoidance while generating sub-optimal yet practical solutions. The effectiveness and generality of the proposed framework are demonstrated through numerical examples on a marine vessel and an aerial quadcopter.

Stroke is a leading cause of long-term disability and the second most common cause of death worldwide. Although acute treatments have advanced, recovery remains challenging and limited. Brain-computer interfaces (BCIs) have emerged as a promising tool for post-stroke rehabilitation by promoting neuroplasticity. However, clinical outcomes remain variable, and optimal protocols have yet to be established. This study explores strategies to optimize BCI-based rehabilitation by comparing motor imagery of affected hand movement versus rest, instead of the conventional left-versus-right motor imagery. This alternative aims to simplify the task and address the weak contralateral activation commonly observed in stroke patients. Two datasets, one from healthy individuals and one from stroke patients, were used to evaluate the proposed approach. The results showed improved performance using both FBCSP and EEGNet. Additionally, we investigated the impact of session duration and found that shorter training sessions produced better BCI performance than longer sessions.

Robust ASR under domain shift is crucial because real-world systems encounter unseen accents and domains with limited labeled data. Although pseudo-labeling offers a practical workaround, it often introduces systematic, accent-specific errors that filtering fails to fix. We ask: How can we correct these recurring biases without target ground truth? We propose a simple parameter-space correction: in a source domain containing both real and pseudo-labeled data, two ASR models are fine-tuned from the same initialization, one on ground-truth labels and the other on pseudo-labels, and their weight difference forms a correction vector that captures pseudo-label biases. When applied to a pseudo-labeled target model, this vector enhances recognition, achieving up to a 35% relative Word Error Rate (WER) reduction on AfriSpeech-200 across ten African accents with the Whisper tiny model.

To avoid the unpredictable phase deviations of the spaceborne phased array (SPA), this paper considers the over-the-air (OTA) phase calibration of the SPA for the low earth orbit (LEO) satellite communications, where the phase deviations of the SPA and the unknown channel are jointly estimated with multiple transmissions of the pilots. Moreover, the Cramer Rao Bound (CRB) is derived, and the optimization of beam patterns is also presented to lower the root mean squared error (RMSE) of the OTA calibration. The simulation results verify the effectiveness of the proposed OTA phase calibration algorithm as the RMSEs of the phase estimates closely approach the corresponding CRB, and the beam pattern optimization scheme is also validated for more than 4dB gain of SNR over the randomly generated beam patterns.

This paper presents our solution to the Multimodal Personality-aware Depression Detection (MPDD) challenge at ACM MM 2025. We propose a multimodal depression detection model in the Elderly that incorporates personality characteristics. We introduce a multi-feature fusion approach based on a co-attention mechanism to effectively integrate LLDs, MFCCs, and Wav2Vec features in the audio modality. For the video modality, we combine representations extracted from OpenFace, ResNet, and DenseNet to construct a comprehensive visual feature set. Recognizing the critical role of personality in depression detection, we design an interaction module that captures the relationships between personality traits and multimodal features. Experimental results from the MPDD Elderly Depression Detection track demonstrate that our method significantly enhances performance, providing valuable insights for future research in multimodal depression detection among elderly populations.

This paper introduces a new boundary element formulation for transient electromagnetic scattering by homogeneous dielectric objects, based on the time-domain PMCHWT equation. To address dense-mesh breakdown, a multiplicative Calderón preconditioner constructed from a modified static electric field integral operator is employed. Large-timestep breakdown and late-time instability are simultaneously resolved through a rescaling of the Helmholtz components using quasi-Helmholtz projectors, with temporal differentiation and integration serving as the rescaling operators. This rescaling additionally balances the loop and star components in the large-timestep regime, thereby preventing loss of accuracy in the secondary quantities caused by numerical cancellation. The resulting discrete system is solved using a marching-on-in-time scheme in conjunction with iterative solvers. Numerical experiments for simply- and multiply-connected dielectric scatterers, including highly non-smooth geometries, corroborate the stability and efficiency of the proposed approach and demonstrate its ability to produce accurate derived quantities in the large-timestep regime.

A passive radar system uses one or more so-called Illuminators of Opportunity (IO) to detect and localize targets. In such systems, a reference channel is often used at each receiving node to capture the transmitted IO signal, while targets are detected using the main surveillance channel. The purpose of the present contribution is to analyze a method for estimating the target parameters in such a system. Specifically, we quantify the additional error contribution due to not knowing the transmitted IO waveform perfectly. A sufficient condition for this error to be negligible as compared to errors due to clutter and noise in the surveillance channel is then given.

Backdoor data poisoning is a crucial technique for ownership protection and defending against malicious attacks. Embedding hidden triggers in training data can manipulate model outputs, enabling provenance verification, and deterring unauthorized use. However, current audio backdoor methods are suboptimal, as poisoned audio often exhibits degraded perceptual quality, which is noticeable to human listeners. This work explores the intrinsic stealthiness and effectiveness of audio watermarking in achieving successful poisoning. We propose a novel Watermark-as-Trigger concept, integrated into the Bloodroot backdoor framework via adversarial LoRA fine-tuning, which enhances perceptual quality while achieving a much higher trigger success rate and clean-sample accuracy. Experiments on speech recognition (SR) and speaker identification (SID) datasets show that watermark-based poisoning remains effective under acoustic filtering and model pruning. The proposed Bloodroot backdoor framework not only secures data-to-model ownership, but also well reveals the risk of adversarial misuse.

In Music Information Retrieval (MIR), modeling and transforming the tone of musical instruments, particularly electric guitars, has gained increasing attention due to the richness of the instrument tone and the flexibility of expression. Tone morphing enables smooth transitions between different guitar sounds, giving musicians greater freedom to explore new textures and personalize their performances. This study explores learning-based approaches for guitar tone morphing, beginning with LoRA fine-tuning to improve the model performance on limited data. Moreover, we introduce a simpler method, named spherical interpolation using Music2Latent. It yields significantly better results than the more complex fine-tuning approach. Experiments show that the proposed architecture generates smoother and more natural tone transitions, making it a practical and efficient tool for music production and real-time audio effects.

High-quality remote sensing (RS) image acquisition is fundamentally constrained by physical limitations. While Multi-Frame Super-Resolution (MFSR) and Pansharpening address this by exploiting complementary information, they are typically studied in isolation: MFSR lacks high-resolution (HR) structural priors for fine-grained texture recovery, whereas Pansharpening relies on upsampled low-resolution (LR) inputs and is sensitive to noise and misalignment. In this paper, we propose SatFusion, a novel and unified framework that seamlessly bridges multi-frame and multi-source RS image fusion. SatFusion extracts HR semantic features by aggregating complementary information from multiple LR multispectral frames via a Multi-Frame Image Fusion (MFIF) module, and integrates fine-grained structural details from an HR panchromatic image through a Multi-Source Image Fusion (MSIF) module with implicit pixel-level alignment. To further alleviate the lack of structural priors during multi-frame fusion, we introduce an advanced variant, SatFusion*, which integrates a panchromatic-guided mechanism into the MFIF stage. Through structure-aware feature embedding and transformer-based adaptive aggregation, SatFusion* enables spatially adaptive feature selection, strengthening the coupling between multi-frame and multi-source representations. Extensive experiments on four benchmark datasets validate our core insight: synergistically coupling multi-frame and multi-source priors effectively resolves the fragility of existing paradigms, delivering superior reconstruction fidelity, robustness, and generalizability.

Engineering design involves demanding models encompassing many decision variables and uncontrollable parameters. In addition, unavoidable aleatoric and epistemic uncertainties can be very impactful and add further complexity. The state-of-the-art adopts two steps, uncertainty quantification and design optimization, to optimize systems under uncertainty by means of robust or stochastic metrics. However, conventional scenario-based, surrogate-assisted, and mathematical programming methods are not sufficiently scalable to be affordable and precise in large and complex cases. Here, a multi-level approach is proposed to accurately optimize resource-intensive, high-dimensional, and complex engineering problems under uncertainty with minimal resources. A non-intrusive, fast-scaling, Kriging-based surrogate is developed to map the combined design/parameter domain efficiently. Multiple surrogates are adaptively updated by hierarchical and orthogonal decomposition to leverage the fewer and most uncertainty-informed data. The proposed method is statistically compared to the state-of-the-art via an analytical testbed and is shown to be concurrently faster and more accurate by orders of magnitude.

Forced response curves (FRCs) of nonlinear systems can exhibit complex behaviors, including hardening/softening behavior and bifurcations. Although topology optimization holds great potential for tuning these nonlinear dynamic responses, its use in high-dimensional systems is limited by the high cost of repeated response and sensitivity analyses. To address this challenge, we employ the spectral submanifolds (SSMs) reduction theory, which reformulates the periodic response as the equilibria of an associated reduced-order model (ROM). This enables efficient and analytic evaluation of both response amplitudes and their sensitivities. Based on the SSM-based ROM, we formulate optimization problems that optimize the peak amplitude, the hardening/softening behavior, and the distance between two saddle-node bifurcations for an FRC. The proposed method is applied to the design of nonlinear MEMS devices, achieving targeted performance optimization. This framework provides a practical and efficient strategy for incorporating nonlinear dynamic effects into the topology optimization of structures.

Advances in portability and low cost of plenoptic cameras have revived interest in light field imaging. Light-field imaging has evolved into a technology that enables us to capture richer visual information. This high-dimensional representation of visual data provides a powerful way to understand the scene, with remarkable improvement in traditional computer vision problems such as depth sensing , post-capture refocusing , material classification, segmentation, and video stabilization. Capturing light fields with high spatial-angular resolution and capturing light field video at high frame rates remains a major challenge due to the limited resolution of the sensors, with limited processing speed. In this paper, we presented an extensive literature review of light field acquisition techniques, challenges associated with different capturing methodology and algorithms proposed for light-field super-resolution, in order to deal with spatial-angular resolution trade-off issue.

The evolution of radio access networks (RANs) toward virtualization and openness creates new opportunities for flexible, cost-effective, and high-performance deployments. Achieving real-time and energy-efficient baseband processing on commercial off-the-shelf platforms, however, remains a critical challenge. This article explores how single instruction multiple data (SIMD) architectures can accelerate RAN workloads. We first outline why key physical-layer functions, such as channel estimation, multiple-input multiple-output (MIMO) detection, and forward error correction, are well aligned with SIMD's data-level parallelism. We then present practical design guidelines and prototype results, showing significant improvements in throughput and energy efficiency compared to conventional CPU-only processing, while retaining programmability and ease of integration. Finally, we discuss open challenges in workload balancing and hardware heterogeneity, and highlight the role of SIMD as an enabling technology for flexible, efficient, and sustainable 6G-ready RANs.

Most current music source separation (MSS) methods rely on supervised learning, limited by training data quantity and quality. Though web-crawling can bring abundant data, platform-level track labeling often causes metadata mismatches, impeding accurate "audio-label" pair acquisition. To address this, we present ACMID: a dataset for MSS generated through web crawling of extensive raw data, followed by automatic cleaning via an instrument classifier built on a pre-trained audio encoder that filters and aggregates clean segments of target instruments from the crawled tracks, resulting in the refined ACMID-Cleaned dataset. Leveraging abundant data, we expand the conventional classification from 4-stem (Vocal/Bass/Drums/Others) to 7-stem (Piano/Drums/Bass/Acoustic Guitar/Electric Guitar/Strings/Wind-Brass), enabling high granularity MSS systems. Experiments on SOTA MSS model demonstrates two key results: (i) MSS model trained with ACMID-Cleaned achieved a 2.39dB improvement in SDR performance compared to that with ACMID-Uncleaned, demostrating the effectiveness of our data cleaning procedure; (ii) incorporating ACMID-Cleaned to training enhances MSS model's average performance by 1.16dB, confirming the value of our dataset. Our data crawling code, cleaning model code and weights are available at: https://github.com/scottishfold0621/ACMID.

While full-duplex speech agents enable natural, low-latency interaction by speaking and listening simultaneously, their consistency and task performance in multi-turn settings remain underexplored. We introduce Full-Duplex-Bench-v2 (FDB-v2), a streaming framework that integrates with an automated examiner that enforces staged goals under two pacing setups (Fast vs. Slow). FDB-v2 covers four task families: daily, correction, entity tracking, and safety. We report turn-taking fluency, multi-turn instruction following, and task-specific competence. The framework is extensible, supporting both commercial APIs and open source models. When we test full-duplex systems with FDB-v2, they often get confused when people talk at the same time, struggle to handle corrections smoothly, and sometimes lose track of who or what is being talked about. Through an open-sourced, standardized streaming protocol and a task set, FDB-v2 makes it easy to extend to new task families, allowing the community to tailor and accelerate evaluation of multi-turn full-duplex systems.

Dynamic metasurface antennas (DMA) provide low-power beamforming through reconfigurable radiative slots. Each slot has a tunable component that consumes low power compared to typical analog components like phase shifters. This makes DMAs a potential candidate to minimize the power consumption of multiple-input multiple-output (MIMO) antenna arrays. In this paper, we investigate the use of DMAs in a wideband communication setting with practical DMA design characteristics. We develop approximations for the DMA beamforming gain that account for the effects of waveguide attenuation, element frequency-selectivity, and limited reconfigurability of the tunable components as a function of the signal bandwidth. The approximations allow for key insights into the wideband performance of DMAs in terms of different design variables. We develop a simple successive beamforming algorithm to improve the wideband performance of DMAs by sequentially configuring each DMA element. Simulation results for a line-of-sight (LOS) wideband system show the accuracy of the approximations with the simulated DMA model in terms of spectral efficiency. We also find that the proposed successive beamforming algorithm increases the overall spectral efficiency of the DMA-based wideband system compared with a baseline DMA beamforming method.

In this manuscript, we explore the application of model-free reinforcement learning in optimizing secure multiparty computation (SMPC) protocols. SMPC is a crucial tool for performing computations on private data without the need to disclose it, holding significant importance in various domains, including information security and privacy. However, the efficiency of current protocols is often suboptimal due to computational and communicational complexities. Our proposed approach leverages model-free reinforcement learning algorithms to enhance the performance of these protocols. We have designed a reinforcement learning model capable of dynamically learning and adapting optimal strategies for secure computations. Our experimental results demonstrate that employing this method leads to a substantial reduction in execution time and communication costs of the protocols. These achievements highlight the high potential of reinforcement learning in improving the efficiency of secure multiparty computation protocols, providing an effective solution to the existing challenges in this field.

Sampling-based approaches are widely used in systems without analytic models to estimate risk or find optimal control. However, gathering sufficient data in such scenarios can be prohibitively costly. On the other hand, in many situations, low-fidelity models or simulators are available from which samples can be obtained at low cost. In this paper, we propose an efficient approach for risk quantification and path integral control that leverages such data from multiple models with heterogeneous sampling costs. A key technical novelty of our approach is the integration of Multi-level Monte Carlo (MLMC) and Multi-fidelity Monte Carlo (MFMC) that enable data from different time and state representations (system models) to be jointly used to reduce variance and improve sampling efficiency. We also provide theoretical analysis of the proposed method and show that our estimator is unbiased and consistent under mild conditions. Finally, we demonstrate via numerical simulation that the proposed method has improved computation (sampling costs) vs. accuracy trade-offs for risk quantification and path integral control.

As orbital debris continues to become a higher priority for the space industry, there is a need to explore how partnerships between the public and private space sector may aid in addressing this issue. This research develops a space logistics framework for planning orbital debris remediation missions, providing a quantitative basis for partnerships that are mutually beneficial between space operators and debris remediators. By integrating network-based space logistics and game theory, we illuminate the high-level costs of remediating orbital debris, and the surplus that stands to be shared as a result. These findings indicate significant progress toward the continued development of a safe, sustainable, and profitable space economy.

We propose enforcing constraints on Model-Based Diffusion by introducing emerging barrier functions inspired by interior point methods. We demonstrate that the standard Model-Based Diffusion algorithm can lead to catastrophic performance degradation in highly constrained environments, even on simple 2D systems due to sample inefficiency in the Monte Carlo approximation of the score function. We introduce Emerging-Barrier Model-Based Diffusion (EB-MBD) which uses progressively introduced barrier constraints to avoid these problems, significantly improving solution quality, without expensive projection operations such as projections. We analyze the sampling liveliness of samples at each iteration to inform barrier parameter scheduling choice. We demonstrate results for 2D collision avoidance and a 3D underwater manipulator system and show that our method achieves lower cost solutions than Model-Based Diffusion, and requires orders of magnitude less computation time than projection based methods.

This study evaluates whether integrating curriculum learning with diffusion-based synthetic augmentation can enhance the detection of difficult pulmonary nodules in chest radiographs, particularly those with low size, brightness, and contrast, which often challenge conventional AI models due to data imbalance and limited annotation. A Faster R-CNN with a Feature Pyramid Network (FPN) backbone was trained on a hybrid dataset comprising expert-labeled NODE21 (1,213 patients; 52.4 percent male; mean age 63.2 +/- 11.5 years), VinDr-CXR, CheXpert, and 11,206 DDPM-generated synthetic images. Difficulty scores based on size, brightness, and contrast guided curriculum learning. Performance was compared to a non-curriculum baseline using mean average precision (mAP), Dice score, and area under the curve (AUC). Statistical tests included bootstrapped confidence intervals, DeLong tests, and paired t-tests. The curriculum model achieved a mean AUC of 0.95 versus 0.89 for the baseline (p < 0.001), with improvements in sensitivity (70 percent vs. 48 percent) and accuracy (82 percent vs. 70 percent). Stratified analysis demonstrated consistent gains across all difficulty bins (Easy to Very Hard). Grad-CAM visualizations confirmed more anatomically focused attention under curriculum learning. These results suggest that curriculum-guided synthetic augmentation enhances model robustness and generalization for pulmonary nodule detection.