This paper proposes a radio tomography system for objects using a camera and sub-terahertz (THz) measurement setup. Unlike conventional reflection-based radar imaging or penetration-based tomography, our system analyzes variations in the radio signal attenuated by a freely and randomly moving object within the line-of-sight (LoS) region of the radio link. The tomography images are reconstructed by synchronizing the object movements captured from the camera with the recorded sub-THz radio measurement attenuation values and performing interpolation. When tested with a set-up at 300 G Hz frequency, 2 dBm transmit power, 30 cm radio link distance, and sweep time of 400 s, the proposed system successfully demonstrates tomography construction of various objects in the LoS region with greater than 99% high spatial resolution accuracy. Also, the proposed approach effectively sees through the objects to identify the internal objects, their composition, and thickness.

Integrating sensing, enabling network intelligence, and improving spectral efficiency are key drivers of the vision for sixth-generation (6G) wireless systems. Conventional pilot-based channel characterization methods incur significant radio-resource overhead, particularly in indoor traffic scenarios, whereas sensing-aided channel characterization offers a viable alternative. Digital twins (DTs) are seen as a promising enabler providing high-fidelity virtual representations of physical environments, which can enhance sensing-aided channel characterization. This paper proposes a DT framework for channel characterization in radio access networks (RANs), leveraging multimodal sensing. Ray tracing is performed on a fused and segmented multimodal point cloud that supports arbitrary placements of both transmitters and receivers, and the DT is validated using a bistatic radio link. In this work, we use received signal strength (RSS) and power delay profile (PDP) as the KPIs. Our results show that multimodal sensing significantly improves environmental awareness, achieving a relative mean absolute error of 2.5 % for RSS and below 15 % for PDP. Therefore, the DT facilitates accurate sensing-aided channel characterization in indoor environments, enabling closed-loop optimization in RAN.

Accurate indoor positioning is vital for applications such as augmented reality and autonomous robotics. Channel state information (CSI)-based methods, particularly when combined with beamforming, massive multiple input multiple output (mMIMO) techniques, and artificial intelligence (AI) algorithms, offer enhanced indoor user equipment (UE) positioning accuracy and robustness in complex indoor environments. In this paper, we present an AI-driven CSI-based indoor positioning method for mMIMO systems, where channel impulse, channel frequency, and angular response domain features are extracted from the CSI data and combined to form both uni-domain and multi-domain feature sets. We introduce a deep attention network (DAN), an AI algorithm that leverages attention mechanisms to effectively integrate and process multi-domain CSI data, thereby enhancing UE positioning performance. We evaluate DAN using a publicly available mMIMO dataset and compare its performance against the baseline and multi-domain convolutional neural network (CNN) models. Our results show that multi-domain DAN outperforms CNN approaches in positioning accuracy, though at the cost of increased inference complexity-highlighting a trade-off between performance and computational overhead. These findings demonstrate the potential of attention mechanisms and multi-domain CSI features for accurate indoor UE positioning systems.

In the domain of non-contact biometrics and human activity recognition, the lack of a versatile, multimodal dataset poses a significant bottleneck. To address this, we introduce the Oulu Multi Sensing (OMuSense-23) dataset that includes biosignals obtained from a mmWave radar, and an RGB-D camera. The dataset features data from 50 individuals in three distinct poses – standing, sitting, and lying down – each featuring four specific breathing pattern activities: regular breathing, reading, guided breathing, and apnea, encompassing both typical situations (e.g., sitting with normal breathing) and critical conditions (e.g., lying down without breathing). In our work, we present a detailed overview of the OMuSense-23 dataset, detailing the data acquisition protocol, describing the process for each participant. In addition, we provide, a baseline evaluation of several data analysis tasks related to biometrics, breathing pattern recognition and pose identification. Our results achieve a pose identification accuracy of 87% and breathing pattern activity recognition of 83% using features extracted from biosignals. The OMuSense-23 dataset is publicly available as resource for other researchers and practitioners in the field.

This paper presents a real-time multimodal radio imaging framework operating in the 110-170 GHz sub-terahertz band, integrating a communications link, an Intel RealSense stereo camera, and a seven-degree-of-freedom robotic arm. The system enables synchronized acquisition of spatial and radio signal data, guided by a YOLOv10-based visual tracker and processed through a software-defined pipeline using inverse distance weighting interpolation. Experimental demonstrations highlight the system's capability to generate high-resolution radio tomography images, even through visual occlusions, showcasing its potential for integrated sensing and communication in 6G applications.

Next-generation wireless systems demand highly efficient power amplifiers (PAs), which introduce complex non-linearities and memory effects. Traditional digital pre-distortion (DPD) methods, such as the Generalized Memory Polynomial (GMP), struggle with these impairments. This paper proposes a Modular Domain-Aware Time-Delay Neural Network (MDA-TDNN), a hybrid physically-inspired architecture that combines established signal processing components into a unified TDNN-based structure. Rather than introducing fundamentally new building blocks, the contribution lies in the systematic integration and evaluation of these components, providing a structured characterization of their individual and joint effects on DPD performance. It includes a Finite Impulse Response (FIR) layer for memory modeling, a phase normalization layer for complex signal behavior, an envelope layer for amplitude non-linearities, and a residual path for stable training. MDA-TDNN was trained on an ideal dataset generated by an Iterative Learning Control scheme and experimentally validated on an RF testbed using a 75 MHz 5G NR signal. The results demonstrate that MDA-TDNN outperforms GMP by 7.95 dB in Adjacent Channel Leakage Ratio (ACLR), achieving performance within 1 dB of ideal DPD. An ablation study highlights the importance of phase normalization and FIR layers. These findings validate the effectiveness of domain-aware neural architectures for PA linearization.

This paper proposes using a cascade of a static nonlinear system and a neural network for power amplifier behavioral modeling and compensation. The results demonstrate that the proposed model provides a better performance-to-complexity trade-off than the current state-of-the-art augmented real-valued time-delay neural network (ARVTDNN) method. When tested with a 64 quadrature amplitude modulated long-term evolution signal with 20 MHz bandwidth and 10.5 dB peak-to-average power ratio at 9.4 dB output back-off on a Doherty-like power amplifier operated at 2.35 GHz, the proposed model achieves an error vector magnitude (EVM) 36.1 dB and adjacent channel leakage ratio (ACLR) of -45.1 dBc. Compared to ARVTDNN, it improves EVM by 2.2 dB and ACLR by 0.4 dB, with a 22% reduction in running complexity.

Accurate reconstruction of radio images can enable the use of communication signals in medical and industrial applications. In this work, we propose a laser-assisted approach for reliable construction of radio frequency (RF) tomographic images using sub-terahertz (THz) communication links and a camera. The experimental setup integrates a 300 GHz radio channel measurement system with a camera and a laser pointer to generate images of freely moving objects within the radio illumination area. We introduce a synchronization strategy based on the ratio of radio and camera timestamp data to ensure temporal alignment. We evaluate the proposed method under varying link distances and measurement durations. We compare the proposed approach against a non-laser baseline using nine hidden objects of different materials and sizes, under an optimal link distance of 50 cm and measurement duration of 400 s. The results demonstrate that the proposed laser-assisted approach achieves consistent and reliable RF imaging performance across multiple objects, with an average imaging accuracy of 88.56%.

Integrated sensing and communications (ISAC) is essential for future 6 G, bridging physical and digital worlds by enabling wireless systems to sense and respond to their environment. Digital twins (DTs) enhance ISAC by providing real-time, data-driven models for applications like localization and autonomous navigation. However, existing DT frameworks lack real-time modeling and multimodal sensing, limiting their use for high-resolution ISAC. This demo paper extends our recently proposed real-time DT framework for indoor ISACenabled robotics and introduces an LLM-driven speech interface for control and planning, a continuous 3D reconstruction scheme using RGB-D data, and a novel ray tracing approach for wireless channel modeling from point clouds. These innovations address key limitations, supporting advanced ISAC applications and immersive human-computer interaction.

Indoor localization is essential for future 6G systems that combine communication and sensing. Accurate positioning is vital for applications like augmented reality and autonomous robotics. In this context, we propose a multi-domain channel state information (CSI)-based localization approach using deep attention networks (DAN) in a massive multiple input multiple output-based (maMIMO) system. We introduce the extraction of features from CSI information in multiple domains, including time, frequency, and Doppler, and design uni-domain and multidomain feature sets. We implement the proposed DAN approach leveraging attention mechanisms to integrate and effectively process the multi-domain CSI data. We evaluate the performance of our model using a publicly available maMIMO dataset and compare it with baseline convolutional neural network (CNN) models. Our results indicate that the DAN-based approach enhances localization performance more than uni-domain, multidomain CNN models and also existing multi-domain-based CNN benchmarks. These findings highlight the benefits of using multidomain features, especially from the Doppler domain along with attention mechanisms for reliable indoor localization.

Millimeter (mmWave) beamforming is an integral component of fifth-generation (5G) and beyond radio communications. 5G beamforming involves the initial beam selection procedure using a codebook with multiple radio beam directions. Conventional codebook-based alignment schemes involve exhaustive sweeping over the pre-defined beam directions, the number of which increases significantly with large numbers of antennas resulting in undesirable latency and communications signal overhead. In this paper, we propose a novel algebraic-based codebook using Grobner basis polynomial solvers to reduce the signal overhead during beam alignment. We also analyze the complexity-performance trade-off between the proposed algebraic-based codebook and the exhaustive-based beam alignment across different monomial thresholds, multiple antenna configurations and radio contextual location information. Our results show that the proposed approach reduces the beam-search overhead at an average complexity reduction ratio of 73.95% with a performance trade-off error of 32.25%.

To enable a wide range of joint communications and sensing applications, future channel models must support dynamic variations of the propagation environment. Moreover, when discovering sensing applications, it is essential to connect physical objects and actions to the radio channel characteristics. In this paper, we perform computer vision (CV) aided automatic mapping and showcase that channel measurements can be linked to the dynamic actions and objects in the environment, thus supporting sensing applications. The proposed method employs a vector network analyzer-based channel measurement system operating at 300 GHz and a camera to enable CV for tracking the object in the radio link. Using human hand as the blockage object, the CV-extracted coordinates and the measured channel responses are combined to generate radio channel footprints for different hand movements.

In recent years, there has been a growing interest in the development of digital technologies for monitoring and treating movement disorders. Nonwearable stationary systems offer an alternative to traditional wearable devices, which can be uncomfortable and inconvenient for patients. This chapter presents an overview of the latest advancements in nonwearable stationary systems for the assessment of movement disorders, including a discussion of their advantages and limitations. We also provide an overview of the current state of the art in this field and present some promising future directions for research and development.

With the rapid development of millimeter wave (mmWave) wireless systems, there is an increasing demand for joint communications and sensing solutions. Indoor human position estimation using wireless fidelity (WiFi) radios may be helpful to provide beyond communication applications such as e-Health, smart buildings, human tracking, etc. under sixth generation (6G) wireless systems. Moreover, WiFi sensing information has the potential to improve communication itself. In the literature, the received signal strength indicator (RSSI)-based methods have been studied but with coarse results due to the usage of omnidirectional antennas. Recent WiFi sensing approaches employ vendor-specific channel state information (CSI) to obtain reliable indoor positioning. In this work, we propose a feed-forward neural network (FNN)-based indoor position estimation framework using RSSI measurements from indoor radio beamforming communication procedure. The acquired RSSI characteristic information from the exhaustive mmWave beam selection process serves as distinctive fingerprints to estimate indoor static human positions. We construct a dataset with obtained RSSI fingerprints for subject positions along LoS, nLoS, and the empty room environment. We obtain a position estimation model using FNN and the dataset. Our results show that the FNN-based framework predicts indoor static human positions using RSSI measurements at an F1-score of 0.86 and accuracy of 0.95. Moreover, the model from such a framework is also robust to distinguish symmetric static positions with respect to the LoS link during mmWave communication.

Indoor human monitoring systems are integral in various applications. They leverage a wide range of sensors, including cameras, radio devices, and inertial measurement units, to collect extensive data from users and the environment. These sensors contribute distinct data modalities, encompassing video feeds from cameras, received signal strength indicators and channel state information from WiFi devices, and three-axis acceleration data from accelerometers. In this context, we present a comprehensive survey of multimodal approaches applied to indoor human monitoring systems, with a specific focus on their relevance in elderly care. Our survey primarily highlights non-contact technologies, particularly cameras and radio devices, as key components in the development of indoor human monitoring systems. Throughout this article, we explore well-established techniques for extracting features from multimodal data sources. Our exploration extends to methodologies for fusing these features and harnessing multiple modalities to improve the accuracy and robustness of machine learning models. Furthermore, we conduct comparative analysis across different data modalities in diverse human monitoring tasks and undertake a comprehensive examination of existing multimodal datasets. This extensive survey not only highlights the significance of indoor human monitoring systems but also emphasizes their versatile applications. In particular, we emphasize their critical role in enhancing the quality of elderly care, offering valuable insights into the development of non-contact monitoring solutions tailored to the needs of aging populations.

Piecewise linearization techniques require dividing the signal into multiple pieces each linearized individually. Machine learning (ML) is one of the useful tools to perform the automatic division of these pieces. Complexity reduction in the classification of piecewise digital predistortion is possible through carefully constructing features from both the signal statistics and the power amplifier (PA) characteristics. Our paper introduces two low-complex classical ML-based methods that facilitate the classification of baseband input data into distinct segments. These methods effectively linearize PA behavior by employing tailored Volterra models corresponding to each segment. Moreover, we perform an in-depth analysis of the proposed schemes to further optimize their classification and regression complexities. The two proposed low-complexity approaches are validated by laboratory experiments and show up to 4 dB error vector magnitude (EVM) improvement over the conventional approach for a class A PA at 28 GHz. Similarly, the EVM improvement is up to 2 dB over the vector-switched general memory polynomial scheme. With only one indirect learning architecture iteration, the two proposed schemes obey the 5G new radio standard up to 6.5 dB and 7 dB output backoff, respectively.

It is essential to mitigate power amplifier (PA) nonlinear (NL) effects to achieve energy-efficient radio communications. To restore the transmit signal quality, digital pre-distortion (DPD) is widely used. Recently, fast convergence DPD (FC-DPD) which offers good PA linearization has been proposed for next-generation broadcasting systems. However, it suffers from complexity issues and this paper addresses that drawback and we propose a low-complex version to make it hardware-friendly. We achieve significant complexity reduction by simplifying the Jacobian computation needed in the FC-DPD algorithm. This scheme can be extended to any memory-less PA model or a measured PA fitted to a polynomial model. We have provided proof that the proposed technique has linear complexity and the simulation results indicate that the proposed scheme achieves performance close to the FC-DPD algorithm. This method can be applied to other transmission systems as well.

Unmanned aerial vehicles (UAVs) are the emerging vital components of millimeter wave (mmWave) wireless systems. Accurate beam alignment is essential for efficient beam based mmWave communications of UAVs with base stations (BSs). Conventional beam sweeping approaches often have large overhead due to the high mobility and autonomous operation of UAVs. Learning-based approaches greatly reduce the overhead by leveraging UAV data, like position to identify optimal beam directions. In this paper, we propose a deep Q-Network(DQN)-based framework for uplink UAV-BS beam alignment where the UAV hovers around 5G new radio (NR) BS coverage area, with varying channel conditions. The proposed learning framework uses the location information and maximize the beamforming gain upon every communication request from UAV inside the multi-location environment. We compare the proposed framework against multi-armed bandit (MAB)-based and exhaustive approaches, respectively and then analyse its training performance over different coverage area requirements, antenna configurations and channel conditions. Our results show that the proposed framework converge faster than the MAB-based approach and comparable to traditional exhaustive approach in an online manner under real-time conditions. Moreover, this approach can be further enhanced to predict the optimal beams for unvisited UAV locations inside the coverage using correlation from neighbouring grid locations.

The vast majority of designs on peak-to-average power ratio (PAPR) reduction and PA linearization schemes in broadcasting systems can be found in literature dealing with both of them in a separate manner on problem formulation, optimization objectives, and implementation issues without considering their mutual influence. Their overall performance might be suboptimal even if each of them has been optimized independently due to possible conflicts as both techniques are interdependent. This paper proposes an adding signal method that jointly achieves PAPR reduction and PA linearization simultaneously, and no extra processing is required at the receiver. The simulation results show that the proposed scheme offers a good performance/complexity trade-off requiring fewer iterations than recent methods.

Millimeter wave (mmWave) beamforming is a vital component of the fifth generation (5G) new radio (NR) and beyond wireless communication systems. The usage of mmWave narrow beams encounters frequent signal attenuation due to random human blockages in indoor environments. Human blockage predictions can jointly improve the signal quality as well as passively sense human activities during mmWave communication. Human sensing using wireless fidelity (WiFi) systems has earlier been studied using receiver signal strength indicator (RSSI) signal level fluctuations based on distance measurements. Other conventional approaches using cameras, lidars, radars, etc. require additional hardware deployments. Current device-free WiFi sensing approaches use vendor-specific channel state information to obtain fine-grained human blockage predictions. Our novelty in this work is to obtain fine-grained human blockage direction predictions in mmWave spectrum, using a time series of RSSI measurements and build fingerprints. We perform experiments to construct a Human Millimetre-wave Radio Blockage Detection (HuMRaBD) dataset and observe human influence in different radio beam directions during each radio initial access procedure. We design a multi layer perceptron (MLP) framework to analyze the HuMRaBD dataset over coarse-grained and fine-grained mmWave blockage directions from static and dynamic human movements. The results show that our trained MLP-trained models can simultaneously sense multiple indoor human radio-blockage directions at an average F1 score of 0.84 and area under curve (AUC) score of 0.95 during mmWave communication.