Papers

Discrete Speech Representation Tokens (DSRTs) have become a foundational component in speech generation. While prior work has extensively studied phonetic and speaker information in DSRTs, how accent information is encoded in DSRTs remains largely unexplored. In this paper, we present the first systematic investigation of accent information in DSRTs. We propose a unified evaluation framework that measures both accessibility of accent information via a novel Accent ABX task and recoverability via cross-accent Voice Conversion (VC) resynthesis. Using this framework, we analyse DSRTs derived from several widely used speech representations. Our results reveal that: (1) choice of layers has the most significant impact on retaining accent information, (2) accent information is substantially reduced by ASR supervision; (3) naive codebook size reduction cannot effectively disentangle accent from phonetic and speaker information.

Delay-Doppler multicarrier modulation (DDMC) techniques have been among the central topics of research for high-Doppler channels. However, a complete transition to DDMC-based waveforms is not yet practically feasible. This is because 5G NR based waveforms, orthogonal frequency division multiplexing (OFDM) and discrete Fourier transform-spread OFDM (DFT-s-OFDM), remain as the modulation schemes for the sixth-generation radio (6GR). Hence, in this paper, we demonstrate how we can still benefit from DD-domain processing in high-mobility scenarios using 5G NR sounding reference signals (SRSs). By considering a DFT-s-OFDM receiver, we transform each received OFDM symbol into the delay-Doppler (DD) domain, where the channel is then estimated. With this approach, we estimate the DD channel parameters, allowing us to predict the aged channel over OFDM symbols without pilots. To improve channel prediction, we propose a linear joint channel estimation and equalization technique, where we use the detected data in each OFDM symbol to sequentially update our channel estimates. Our simulation results show that the proposed technique significantly outperforms the conventional frequency-domain estimation technique in terms of bit error rate (BER) and normalized mean squared error (NMSE). Furthermore, we show that using only two slots with SRS for initial channel estimation, our method supports pilot-free detection for more than 25 subsequent OFDM symbols.

Echocardiography is a cornerstone for managing heart failure (HF), with Left Ventricular Ejection Fraction (LVEF) being a critical metric for guiding therapy. However, manual LVEF assessment suffers from high inter-observer variability, while existing Deep Learning (DL) models are often computationally intensive and data-hungry "black boxes" that impede clinical trust and adoption. Here, we propose a backpropagation-free multi-task Green Learning (MTGL) framework that performs simultaneous Left Ventricle (LV) segmentation and LVEF classification. Our framework integrates an unsupervised VoxelHop encoder for hierarchical spatio-temporal feature extraction with a multi-level regression decoder and an XG-Boost classifier. On the EchoNet-Dynamic dataset, our MTGL model achieves state-of-the-art classification and segmentation performance, attaining a classification accuracy of 94.3% and a Dice Similarity Coefficient (DSC) of 0.912, significantly outperforming several advanced 3D DL models. Crucially, our model achieves this with over an order of magnitude fewer parameters, demonstrating exceptional computational efficiency. This work demonstrates that the GL paradigm can deliver highly accurate, efficient, and interpretable solutions for complex medical image analysis, paving the way for more sustainable and trustworthy artificial intelligence in clinical practice.

We present SCOPE, a fast and efficient framework for modeling and manipulating deformable linear objects (DLOs). Unlike conventional energy-based approaches, SCOPE leverages convex approximations to significantly reduce computational cost while maintaining smooth and physically plausible deformations. This trade-off between speed and accuracy makes the method particularly suitable for applications requiring real-time or near-real-time response. The effectiveness of the proposed framework is demonstrated through comprehensive simulation experiments, highlighting its ability to generate smooth shape trajectories under geometric and length constraints.

The performance evaluation remains a complex challenge in audio separation, and existing evaluation metrics are often misaligned with human perception, course-grained, relying on ground truth signals. On the other hand, subjective listening tests remain the gold standard for real-world evaluation, but they are expensive, time-consuming, and difficult to scale. This paper addresses the growing need for automated systems capable of evaluating audio separation without human intervention. The proposed evaluation metric, SAM Audio Judge (SAJ), is a multimodal fine-grained reference-free objective metric, which shows highly alignment with human perceptions. SAJ supports three audio domains (speech, music and general sound events) and three prompt inputs (text, visual and span), covering four different dimensions of evaluation (recall, percision, faithfulness, and overall). SAM Audio Judge also shows potential applications in data filtering, pseudo-labeling large datasets and reranking in audio separation models. We release our code and pre-trained models at: https://github.com/facebookresearch/sam-audio.

Frequency control following a contingency event is of vital concern in power system operations. Leveraging inverter-based resources, it is not hard to shape the center of inertia (COI) frequency nicely. However, under weak grid conditions, it becomes insufficient to solely shape the COI frequency since this aggregate signal fails to reveal the inter-area oscillations. In this manuscript, we advocate for foolproof fine-tuning rules for frequency shaping control (FS) based on a systematic analysis of damping ratio and decay rate of inter-area oscillations to simultaneously meet specified metrics for frequency security and oscillatory stability. To this end, building on a modal decomposition, we simplify the oscillation damping problem into a pole-placement task for a set of scalar subsystems, which can be efficiently solved by only investigating the root locus of a scalar subsystem associated with the main mode, while FS inherently guarantees a Nadir-less COI frequency response. Through our proposed root-locus-based oscillatory stability analysis, we derive closed-form expressions for the minimum damping ratio and decay rate among inter-area oscillations in terms of networked system and control parameters under FS. Moreover, we propose useful tuning guidelines for FS which need only simple calculations or visualized tuning to not only shape the COI frequency into a first-order response that converges to a steady-state value within the allowed range but also ensure a satisfactory damping ratio and decay rate of inter-area oscillations following disturbances. As for the common virtual inertia control (VI), although similar oscillatory stability analysis becomes intractable, one can still glean some insights via the root locus method. Numerical simulations validate the proposed tuning for FS as well as the superiority of FS over VI in exponential convergence rate.

This paper considers an intelligent transmitting surface (ITS) integrated into a base station and develops a low-overhead maximum a posteriori (MAP) probability channel tracking method for the dominant line-of-sight link between the ITS and the user equipment. We cast the per-block channel as a three-parameter model consisting of the channel amplitude, channel phase, and angle-of-arrival at the ITS. We exploit temporal correlation by updating the priors using the estimates from the previous block. Using only two pilots per coherence block alongside a targeted beam alignment strategy, the proposed method achieves precise channel tracking and attains spectral efficiency close to that achievable under perfect channel knowledge.

This paper develops a multi-user downlink communication framework for distributed low Earth orbit satellite networks serving ground users equipped with multiple antennas. Building upon the concept of cell-free multiple-input multiple-output in terrestrial networks, we propose a coordinated transmission scheme where multiple satellites jointly transmit spatially multiplexed data streams to each user. Using a new approximate achievable rate expression, we formulate a sum rate maximization problem under per-satellite and per-antenna power constraints and use the classical equivalence between sum rate maximization and mean square error minimization to optimize the satellites' precoding matrices using statistical channel state information. We numerically examine the performance of the proposed scheme in different settings and validate its effectiveness by comparing it against traditional precoding designs.

We study damping of inter-area oscillations in transmission grids using voltage-source-converter-based high-voltage direct-current (VSC-HVDC) links. Conventional power oscillation damping controllers rely on system models that are difficult to obtain in practice. Data-driven Predictive Control (DPC) addresses this limitation by replacing explicit models with data. We apply AutoRegressive with eXogenous inputs (ARX)-based predictive control and its Transient Predictive Control (TPC) variant, and compare them with Data-enabled Predictive Control (DeePC) and two standard model-based controllers. The methods are evaluated in simulation on a system exhibiting both inter-area and local oscillation modes. ARX-based predictive control and DeePC both achieve effective damping, while the ARX-based methods require less online computation. Using warm-started, pre-factorized operator-splitting solvers, ARX/TPC control actions are computed in less than 1ms. These results demonstrate that DPC is a viable approach for power-system oscillation damping for the given test case.

We investigate robust direction-of-arrival (DoA) estimation for sensor arrays operating in adverse weather conditions, where weather-induced distortions degrade estimation accuracy. Building on a physics-based $S$-matrix model established in prior work, we adopt a statistical characterization of random phase and amplitude distortions caused by multiple scattering in rain. Based on this model, we develop a measurement framework for uniform linear arrays (ULAs) that explicitly incorporates such distortions. To mitigate their impact, we exploit the Hermitian Toeplitz (HT) structure of the covariance matrix to reduce the number of parameters to be estimated. We then apply a generalized least squares (GLS) approach for calibration. Simulation results show that the proposed method effectively suppresses rain-induced distortions, improves DoA estimation accuracy, and enhances radar sensing performance in challenging weather conditions.

Port-Hamiltonian neural networks have shown promising results in the identification of nonlinear dynamics of complex systems, as their combination of physical principles with data-driven learning allows for accurate modelling. However, due to the non-convex optimization problem inherent in learning the correct network parameters, the training procedure is prone to converging to local minima, potentially leading to poor performance. In order to avoid this issue, this paper proposes an improved initialization for port-Hamiltonian neural networks. The core idea is to first estimate a linear port-Hamiltonian system to be used as an initialization for the network, after which the neural network adapts to the system nonlinearities, reducing the training times and improving convergence. The effectiveness of this method is tested on a chained mass-spring-damper setup for varying noise levels and compared to the original approach.

We study rocket-driven Penrose extraction in the test-particle limit on a fixed Kerr background for equatorial prograde flybys under explicit steering prescriptions. A spacecraft ejects exhaust inside the ergosphere; when the exhaust attains negative Killing energy, the remaining spacecraft gains energy by 4-momentum conservation. Across 320{,}000 simulated trajectories spanning black-hole spin, exhaust velocity, and orbital parameters, extraction with escape is rare in broad parameter scans (at most ${\sim}1\%$) and requires high spin ($a/M\gtrsim 0.89$), highly relativistic exhaust ($v_e\gtrsim 0.91c$), and finely tuned initial conditions. Under optimal tuning the success rate reaches ${\sim}70\%$ at $a/M = 0.95$. For representative escape trajectories, a single periapsis impulse is more propellant-efficient than the continuous-thrust controllers studied here. All quoted thresholds are empirical and specific to the orbit family, prior, and steering protocol studied.

Recent advances in video-audio (V-A) understanding and generation have increasingly relied on joint V-A embeddings, which serve as the foundation for tasks such as cross-modal retrieval and generation. While prior methods like CAVP effectively model semantic and temporal correspondences between modalities using contrastive objectives, their performance remains suboptimal. A key limitation is the insufficient modeling of the dense, multi-scale nature of both video and audio signals, correspondences often span fine- to coarse-grained spatial-temporal structures, which are underutilized in existing frameworks. To this end, we propose GMS-CAVP, a novel framework that combines Multi-Scale Video-Audio Alignment and Multi-Scale Spatial-Temporal Diffusion-based pretraining objectives to enhance V-A correspondence modeling. First, GMS-CAVP introduces a multi-scale contrastive learning strategy that captures semantic and temporal relations across varying granularities. Second, we go beyond traditional contrastive learning by incorporating a diffusion-based generative objective, enabling modality translation and synthesis between video and audio. This unified discriminative-generative formulation facilitates deeper cross-modal understanding and paves the way for high-fidelity generation. Extensive experiments on VGGSound, AudioSet, and Panda70M demonstrate that GMS-CAVP outperforms previous methods in generation and retrieval.

The dielectric characterization of human tissues can play a crucial role in the development of new medical diagnostic tools. In particular, the characterization of healthy and pathological tissues can provide vital information for diagnosis. In this paper, preliminary results from a small-scale measurement campaign conducted in 0.5-26.5GHz during real surgeries on healthy and malignant human colon tissues are presented. Those measurements were carried out externally to the colon, without direct contact to the tumor growing inside the colon. Furthermore, different tumor stages are taken into account. Initial findings reveal that advanced tumor stages are related with increased higher values of dielectric properties in malignant tumor tissues compared to the healthy ones.

This paper introduces a reconfigurable intelligent surface (RIS) to support parameter estimation in machine-type communications (MTC). We focus on a network where single-antenna sensors transmit spatially correlated measurements to a multiple-antenna collector node (CN) via non-orthogonal multiple access. We propose an estimation scheme based on the minimum mean square error (MMSE) criterion. We also integrate successive interference cancelation (SIC) at the receiver to mitigate communication failures in noisy and interference-prone channels under the finite blocklength (FBL) regime. Moreover, recognizing the importance of channel state information (CSI), we explore various methodologies for its acquisition at the CN. We statistically design the RIS configuration and SIC decoding order to minimize estimation error while accounting for channel temporal variations and short packet lengths. To mirror practical systems, we incorporate the detrimental effects of FBL communication and imperfect CSI errors in our analysis. Simulations demonstrate that larger reflecting surfaces lead to smaller MSEs and underscore the importance of selecting an appropriate decoding order for accuracy and ultimate performance.

Electromagnetic (EM) exposure compliance has long been recognized as a crucial aspect of communications terminal designs. However, accurately assessing the impact of EM exposure for proper design strategies remains challenging. In this paper, we develop a long-term thermal EM exposure constraint model and propose a novel adaptive exposure-aware beamforming design for an mmWave uplink system. Specifically, we first establish an equivalent channel model based on Maxwell's radiation equations, which accurately captures the EM physical effects. Then, we derive a closed-form thermal impulse response model from the Pennes bioheat transfer equation (BHTE), characterizing the thermal inertia of tissue. Inspired by this model, we formulate a beamforming optimization problem that translates rigid instantaneous exposure limits into a flexible long-term thermal budget constraint. Furthermore, we develop a low-complexity online beamforming algorithm based on Lyapunov optimization theory, obtaining a closed-form near-optimal solution. Simulation results demonstrate that the proposed algorithm effectively stabilizes tissue temperature near a predefined safety threshold and significantly outperforms the conventional scheme with instantaneous exposure constraints.

This paper focuses on audio deepfake detection under real-world communication degradations, with an emphasis on ultra-short inputs (0.5-2.0s), targeting the capability to detect synthetic speech at a conversation opening, e.g., when a scammer says "Hi." We propose Short-MGAA (S-MGAA), a novel lightweight extension of Multi-Granularity Adaptive Time-Frequency Attention, designed to enhance discriminative representation learning for short, degraded inputs subjected to communication processing and perturbations. The S-MGAA integrates two tailored modules: a Pixel-Channel Enhanced Module (PCEM) that amplifies fine-grained time-frequency saliency, and a Frequency Compensation Enhanced Module (FCEM) to supplement limited temporal evidence via multi-scale frequency modeling and adaptive frequency-temporal interaction. Extensive experiments demonstrate that S-MGAA consistently surpasses nine state-of-the-art baselines while achieving strong robustness to degradations and favorable efficiency-accuracy trade-offs, including low RTF, competitive GFLOPs, compact parameters, and reduced training cost, highlighting its strong potential for real-time deployment in communication systems and edge devices.

This paper presents a Cramer-Rao bound (CRB) for the estimation of parameters confined to an arbitrary set. Unlike existing results that rely on equality or inequality constraints, manifold structures, or the nonsingularity of the Fisher information matrix, the derived CRB applies to any constrained set and holds for any estimation bias and any Fisher information matrix. The key geometric object governing the new CRB is the tangent cone to the constraint set, whose span determines how the constraints affect the estimation accuracy. This CRB subsumes, unifies, and generalizes known special cases, offering an intuitive and broadly applicable framework to characterize the minimum mean-square error of constrained estimators.

Within the context of massive machine-type communications+, reconfigurable intelligent surfaces (RISs) represent a promising technology to boost system performance in scenarios with poor channel conditions. Considering single-antenna sensors transmitting short data packets to a multiple-antenna collector node, we introduce and design an RIS to maximize the weighted sum rate (WSR) of the system working in the finite blocklength regime. Due to the large number of reflecting elements and their passive nature, channel estimation errors may occur. In this letter, we then propose a robust RIS optimization to combat such a detrimental issue. Based on concave bounds and approximations, the nonconvex WSR problem for the RIS response is addressed via successive convex optimization (SCO). Numerical experiments validate the performance and complexity of the SCO solutions.

Cell-free massive multiple-input-multiple-output is considered a promising technology for the next generation of wireless communication networks. The main idea is to distribute a large number of access points (APs) in a geographical region to serve the user equipments (UEs) cooperatively. In the uplink, one of two types of operations is often adopted: centralized or distributed. In centralized operation, channel estimation and data decoding are performed at the central processing unit (CPU), whereas in distributed operation, channel estimation occurs at the APs and data detection at the CPU. In this paper, we propose a novel uplink operation, termed Master-Assisted Distributed Uplink Operation (MADUO), where each UE is assigned a master AP, which receives soft data estimates from the other APs and decodes the data using its local signals and the received data estimates. Numerical experiments demonstrate that the proposed operation performs comparably to the centralized operation and balances fronthaul signaling and computational complexity.

Most existing sound field reconstruction methods target point-to-region reconstruction, interpolating the Acoustic Transfer Functions (ATFs) between a fixed-position sound source and a receiver region. The applicability of these methods is limited because real-world ATFs tend to varying continuously with respect to the positions of sound sources and receiver regions. This paper presents a permutation-invariant physics-informed neural network for region-to-region sound field reconstruction, which aims to interpolate the ATFs across continuously varying sound sources and measurement regions. The proposed method employs a deep set architecture to process the receiver and sound source positions as an unordered set, preserving acoustic reciprocity. Furthermore, it incorporates the Helmholtz equation as a physical constraint to guide network training, ensuring physically consistent predictions.

This paper aims to provide a clear and rigorous understanding of commonly recognized safety constraints in physical human-robot interaction, particularly regarding ISO/TS 15066. We investigate the derivation of these constraints, critically examine the underlying assumptions, and evaluate their practical implications for system-level safety and performance in industrially relevant scenarios. Key design parameters within safety-critical control architectures are identified, and numerical examples are provided to quantify performance degradation arising from typical approximations and design decisions in manufacturing environments. Within this analysis, the fundamental role of energy in safety assessment is emphasized, providing focused insights into energy-based safety methodologies for collaborative industrial robot systems.

Autonomous UAV forestry operations require robust depth estimation with strong cross-domain generalization, yet existing evaluations focus on urban and indoor scenarios, leaving a critical gap for vegetation-dense environments. We present the first systematic zero-shot evaluation of eight stereo methods spanning iterative refinement, foundation model, diffusion-based, and 3D CNN paradigms. All methods use officially released pretrained weights (trained on Scene Flow) and are evaluated on four standard benchmarks (ETH3D, KITTI 2012/2015, Middlebury) plus a novel 5,313-pair Canterbury Tree Branches dataset ($1920 \times 1080$). Results reveal scene-dependent patterns: foundation models excel on structured scenes (BridgeDepth: 0.23 px on ETH3D; DEFOM: 4.65 px on Middlebury), while iterative methods show variable cross-benchmark performance (IGEV++: 0.36 px on ETH3D but 6.77 px on Middlebury; IGEV: 0.33 px on ETH3D but 4.99 px on Middlebury). Qualitative evaluation on the Tree Branches dataset establishes DEFOM as the gold-standard baseline for vegetation depth estimation, with superior cross-domain consistency (consistently ranking 1st-2nd across benchmarks, average rank 1.75). DEFOM predictions will serve as pseudo-ground-truth for future benchmarking.

We propose a digital backpropagation method that employs machine-learning-aided joint optimization of dispersion step lengths and nonlinear phase rotation filters within an FFT-based enhanced split-step Fourier structure, achieving improved accuracy at low computational complexity.

Queue management and resource allocation play a critical role in enabling cooperative status awareness in vehicular networks. This paper investigates the problem of age of information (AoI)-aware status updates in vehicle-to-vehicle (V2V) communication, where each vehicle's status is represented by multiple interdependent packets. To enable fine-grained queue management at the packet level under resource constraints, we formulate a joint optimization problem that simultaneously learns active packet dropping and transmit power control strategies. A hybrid action space is designed to support both discrete dropping decisions and continuous power control. To exploit the graph-structured interference inherent in V2V topology, a graph neural network (GNN) is introduced to aggregate slowly varying large-scale fading, allowing agents to capture topological dependencies implicitly without frequent message exchange. The overall framework is built upon multi-agent proximal policy optimization (MAPPO), with centralized training and decentralized execution (CTDE). Simulations demonstrate that the proposed method significantly reduces average AoI across a wide range of network densities, channel conditions, and traffic loads, consistently outperforming several baselines.

Cooperative double intelligent reflecting surface (double-IRS) has emerged as a promising approach for enhancing physical layer security (PLS) in MIMO systems. However, existing studies are limited to narrowband scenarios and fail to address wideband MIMO-OFDM. In this regime, frequency-flat IRS phases and cascaded IRS links cause severe coupling, rendering narrowband designs inapplicable. To overcome this challenge, we introduce cooperative double-IRS-assisted wideband MIMO-OFDM and propose an efficient manifold-based solution. By regarding the power and constant modulus constraints as Riemannian manifolds, we reformulate the non-convex secrecy sum rate maximization as an unconstrained optimization on a product manifold. Building on this formulation, we further develop a product Riemannian gradient descent (PRGD) algorithm with guaranteed stationary convergence. Simulation results demonstrate that the proposed scheme effectively resolves the OFDM coupling issue and achieves significant secrecy rate gains, outperforming single-IRS and distributed multi-IRS benchmarks by 32.0% and 22.3%, respectively.

Deploying deep learning agents for autonomous navigation in unstructured environments faces critical challenges regarding safety, data scarcity, and limited computational resources. Traditional solvers often suffer from high latency, while emerging learning-based approaches struggle to ensure deterministic feasibility. To bridge the gap from embodied to embedded intelligence, we propose a self-supervised framework incorporating a differentiable hard constraint projection layer for runtime assurance. To mitigate data scarcity, we construct a Global-Guided Artificial Potential Field (G-APF), which provides dense supervision signals without manual labeling. To enforce actuator limitations and geometric constraints efficiently, we employ an adaptive neural projection layer, which iteratively rectifies the coarse network output onto the feasible manifold. Extensive benchmarks on a test set of 20,000 scenarios demonstrate an 88.75% success rate, substantiating the enhanced operational safety. Closed-loop experiments in CARLA further validate the physical realizability of the planned paths under dynamic constraints. Furthermore, deployment verification on an NVIDIA Jetson Orin NX confirms an inference latency of 94 ms, showing real-time feasibility on resource-constrained embedded hardware. This framework offers a generalized paradigm for embedding physical laws into neural architectures, providing a viable direction for solving constrained optimization in mechatronics. Source code is available at: https://github.com/wzq-13/SSHC.git.

Multimodal MRI is essential for brain tumor segmentation, yet missing modalities in clinical practice cause existing methods to exhibit >40% performance variance across modality combinations, rendering them clinically unreliable. We propose AMGFormer, achieving significantly improved stability through three synergistic modules: (1) QuadIntegrator Bridge (QIB) enabling spatially adaptive fusion maintaining consistent predictions regardless of available modalities, (2) Multi-Granular Attention Orchestrator (MGAO) focusing on pathological regions to reduce background sensitivity, and (3) Modality Quality-Aware Enhancement (MQAE) preventing error propagation from corrupted sequences. On BraTS 2018, our method achieves 89.33% WT, 82.70% TC, 67.23% ET Dice scores with <0.5% variance across 15 modality combinations, solving the stability crisis. Single-modality ET segmentation shows 40-81% relative improvements over state-of-the-art methods. The method generalizes to BraTS 2020/2021, achieving up to 92.44% WT, 89.91% TC, 84.57% ET. The model demonstrates potential for clinical deployment with 1.2s inference. Code: https://github.com/guochengxiangives/AMGFormer.

This article presents an eco-driving algorithm for electric vehicles featuring multi-speed transmissions. The proposed controller is formulated as a co-optimization problem, simultaneously optimizing both vehicle longitudinal speed and powertrain operation to maximize energy efficiency. Constraints derived from a connected vehicle based traffic prediction algorithm are used to ensure traffic safety and smooth traffic flow in dynamic environments with multiple signalized intersections and mixed traffic. By simplifying the complex, nonlinear mixed integer problem, the proposed controller achieves computational efficiency, enabling real-time implementation. To evaluate its performance, traffic scenarios from both Simulation of Urban MObility (SUMO) and real-world road tests are employed. The results demonstrate a notable reduction in energy consumption by up to 11.36% over an \SI{18}{\km} drive.

This paper introduces an interference-free multi-stream transmission architecture leveraging stacked intelligent metasurfaces (SIMs), from a new perspective of interference exploitation. Unlike traditional interference exploitation precoding (IEP) which relies on computational hardware circuitry, we perform the precoding operations within the analog wave domain provided by SIMs. However, the benefits of SIM-enabled IEP are limited by the nonlinear distortion (NLD) caused by power amplifiers. A hardware-efficient interference-free transmitter architecture is developed to exploit SIM's high and flexible degree of freedom (DoF), where the NLD on modulated symbols can be directly compensated in the wave domain. Moreover, we design a frame-level SIM configuration scheme and formulate a maxmin problem on the safety margin function. With respect to the optimization of SIM phase shifts, we propose a recursive oblique manifold (ROM) algorithm to tackle the complex coupling among phase shifts across multiple layers. A flexible DoF-driven antenna selection (AS) scheme is explored in the SIM-enabled IEP system. Using an ROM-based alternating optimization (ROM-AO) framework, our approach jointly optimizes transmit AS, SIM phase shift design, and power allocation (PA), and develops a greedy safety margin-based AS algorithm. Simulations show that the proposed SIM-enabled frame-level IEP scheme significantly outperforms benchmarks. Specifically, the strategy with AS and PA can achieve a 20 dB performance gain compared to the case without any strategy under the 12 dB signal-to-noise ratio, which confirms the superiority of the NLD-aware IEP scheme and the effectiveness of the proposed algorithm.