Efficient Frequency Selective Surface Analysis via End-to-End Model-Based Learning

This paper introduces an innovative end-to-end model-based deep learning approach for efficient electromagnetic analysis of high-dimensional frequency selective surfaces (FSS). Unlike traditional data-driven methods that require large datasets, this approach combines physical insights from equivalent circuit models with deep learning techniques to significantly reduce model complexity and enhance prediction accuracy. Compared to previously introduced model-based learning approaches, the proposed method is trained end-to-end from the physical structure of the FSS (geometric parameters) to its electromagnetic response (S-parameters). Additionally, an improvement in phase prediction accuracy through a modified loss function is presented. Comparisons with direct models, including deep neural networks (DNN) and radial basis function networks (RBFN), demonstrate the superiority of the model-based approach in terms of computational efficiency, model size, and generalization capability.

Unsupervised Learning for Gain-Phase Impairment Calibration in ISAC Systems

Gain-phase impairments (GPIs) affect both communication and sensing in 6G integrated sensing and communication (ISAC). We study the effect of GPIs in a single-input, multiple-output orthogonal frequency-division multiplexing ISAC system and develop a model-based unsupervised learning approach to simultaneously (i) estimate the gain-phase errors and (ii) localize sensing targets. The proposed method is based on the optimal maximum a-posteriori ratio test for a single target. Results show that the proposed approach can effectively estimate the gain-phase errors and yield similar position estimation performance as the case when the impairments are fully known.

Optimal blind focusing on perturbation-inducing targets in sub-unitary complex media

The scattering of waves in a complex medium is perturbed by polarizability changes or motion of embedded targets. These perturbations could serve as perfectly non-invasive guidestars for focusing on the targets. In this Letter, we theoretically derive a fundamental difference between these two perturbation types (the change of the scattering matrix is of rank one [two] for target polarizability changes [motion]) and identify accordingly optimal strategies to perfectly focus on the target in both cases. For target motion, at least two displacements are necessary. Furthermore, for the case of dynamic complex media additionally featuring parasitic perturbers, we establish a non-invasive scheme to achieve optimal time-averaged power delivery to a perturbation-inducing target. In all cases, no assumptions about the unitarity of the system's scattering matrix or the size of the perturbation are necessary. We experimentally demonstrate all results in the microwave regime using a strongly sub-unitary lossy chaotic cavity as complex medium. Our experiments highlight that the target's "structural scattering" is irrelevant [must be negligible] in the case of target polarizability changes [motion]. We expect our results to find applications in communications, cybersecurity, bioelectronics, flow-cytometry and self-propelled nano-swimmers.

Model-based Deep Learning for Beam Prediction based on a Channel Chart

Channel charting builds a map of the radio environment in an unsupervised way. The obtained chart locations can be seen as low-dimensional compressed versions of channel state information that can be used for a wide variety of applications, including beam prediction. In non-standalone or cell-free systems, chart locations computed at a given base station can be transmitted to several other base stations (possibly operating at different frequency bands) for them to predict which beams to use. This potentially yields a dramatic reduction of the overhead due to channel estimation or beam management, since only the base station performing charting requires channel state information, the others directly predicting the beam from the chart location. In this paper, advanced model-based neural network architectures are proposed for both channel charting and beam prediction. The proposed methods are assessed on realistic synthetic channels, yielding promising results.

Semi-Supervised End-to-End Learning for Integrated Sensing and Communications

Integrated sensing and communications (ISAC) is envisioned as one of the key enablers of next-generation wireless systems, offering improved hardware, spectral, and energy efficiencies. In this paper, we consider an ISAC transceiver with an impaired uniform linear array that performs single-target detection and position estimation, and multiple-input single-output communications. A differentiable model-based learning approach is considered, which optimizes both the transmitter and the sensing receiver in an end-to-end manner. An unsupervised loss function that enables impairment compensation without the need for labeled data is proposed. Semi-supervised learning strategies are also proposed, which use a combination of small amounts of labeled data and unlabeled data. Our results show that semi-supervised learning can achieve similar performance to supervised learning with 98.8% less required labeled data.

Optimizing Multicarrier Multiantenna Systems for LoS Channel Charting

Channel charting (CC) consists in learning a mapping between the space of raw channel observations, made available from pilot-based channel estimation in multicarrier multiantenna system, and a low-dimensional space where close points correspond to channels of user equipments (UEs) close spatially. Among the different methods of learning this mapping, some rely on a distance measure between channel vectors. Such a distance should reliably reflect the local spatial neighborhoods of the UEs. The recently proposed phase-insensitive (PI) distance exhibits good properties in this regards, but suffers from ambiguities due to both its periodic and oscillatory aspects, making users far away from each other appear closer in some cases. In this paper, a thorough theoretical analysis of the said distance and its limitations is provided, giving insights on how they can be mitigated. Guidelines for designing systems capable of learning quality charts are consequently derived. Experimental validation is then conducted on synthetic and realistic data in different scenarios.

Model-based Deep Learning for High-Dimensional Periodic Structures

This work presents a deep learning surrogate model for the fast simulation of high-dimensional frequency selective surfaces. We consider unit-cells which are built as multiple concatenated stacks of screens and their design requires the control over many geometrical degrees of freedom. Thanks to the introduction of physical insight into the model, it can produce accurate predictions of the S-parameters of a certain structure after training with a reduced dataset.The proposed model is highly versatile and it can be used with any kind of frequency selective surface, based on either perforations or patches of any arbitrary geometry. Numeric examples are presented here for the case of frequency selective surfaces composed of screens with rectangular perforations, showing an excellent agreement between the predicted performance and such obtained with a full-wave simulator.

Model-based learning for location-to-channel mapping

Modern communication systems rely on accurate channel estimation to achieve efficient and reliable transmission of information. As the communication channel response is highly related to the user's location, one can use a neural network to map the user's spatial coordinates to the channel coefficients. However, these latter are rapidly varying as a function of the location, on the order of the wavelength. Classical neural architectures being biased towards learning low frequency functions (spectral bias), such mapping is therefore notably difficult to learn. In order to overcome this limitation, this paper presents a frugal, model-based network that separates the low frequency from the high frequency components of the target mapping function. This yields an hypernetwork architecture where the neural network only learns low frequency sparse coefficients in a dictionary of high frequency components. Simulation results show that the proposed neural network outperforms standard approaches on realistic synthetic data.

Experimentally realized physical-model-based wave control in metasurface-programmable complex media

The reconfigurability of radio environments with programmable metasurfaces is considered a key feature of next-generation wireless networks. Identifying suitable metasurface configurations for desired wireless functionalities requires a precise setting-specific understanding of the intricate impact of the metasurface configuration on the wireless channels. Yet, to date, the relevant short and long-range correlations between the meta-atoms due to proximity and reverberation are largely ignored rather than precisely captured. Here, we experimentally demonstrate that a compact model derived from first physical principles can precisely predict how wireless channels in complex scattering environments depend on the programmable-metasurface configuration. The model is calibrated using a very small random subset of all possible metasurface configurations and without knowing the setup's geometry. Our approach achieves two orders of magnitude higher precision than a deep learning-based digital-twin benchmark while involving hundred times fewer parameters. Strikingly, when only phaseless calibration data is available, our model can nonetheless retrieve the precise phase relations of the scattering matrix as well as their dependencies on the metasurface configuration. Thereby, we achieve coherent wave control (focusing or enhancing absorption) and phase-shift-keying backscatter communications without ever having measured phase information. Finally, our model is also capable of retrieving the essential properties of scattering coefficients for which no calibration data was ever provided. These unique generalization capabilities of our pure-physics model significantly alleviate the measurement complexity. Our approach is also directly relevant to dynamic metasurface antennas, microwave-based signal processors as well as emerging in situ reconfigurable nanophotonic, optical and room-acoustical systems.

Model-Based End-to-End Learning for Multi-Target Integrated Sensing and Communication

We study model-based end-to-end learning in the context of integrated sensing and communication (ISAC) under hardware impairments. A monostatic orthogonal frequency-division multiplexing (OFDM) sensing and multiple-input single-output (MISO) communication scenario is considered, incorporating hardware imperfections at the ISAC transceiver antenna array. To enable end-to-end learning of the ISAC transmitter and sensing receiver, we propose a novel differentiable version of the orthogonal matching pursuit (OMP) algorithm that is suitable for multi-target sensing. Based on the differentiable OMP, we devise two model-based parameterization strategies to account for hardware impairments: (i) learning a dictionary of steering vectors for different angles, and (ii) learning the parameterized hardware impairments. For the single-target case, we carry out a comprehensive performance analysis of the proposed model-based learning approaches, a neural-network-based learning approach and a strong baseline consisting of least-squares beamforming, conventional OMP, and maximum-likelihood symbol detection for communication. Results show that learning the parameterized hardware impairments offers higher detection probability, better angle and range estimation accuracy, lower communication symbol error rate (SER), and exhibits the lowest complexity among all learning methods. Lastly, we demonstrate that learning the parameterized hardware impairments is scalable also to multiple targets, revealing significant improvements in terms of ISAC performance over the baseline.