Browsing by Department "Department of Information and Communications Engineering"

Filter results by typing the first few letters
Now showing 1 - 20 of 552
  • 3D Radio Map-based GPS spoofing Detection and Mitigation for Cellular-Connected UAVs
    (2023) Dang, Yongchao; Karakoc, Alp; Saba, Norshahida; Jäntti, Riku
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    With the upcoming 5G and beyond wireless communication system, cellular-connected Unmanned Aerial Vehicles (UAVs) are emerging as a new pattern to give assistance for target searching, emergency rescue, and network recovery. Such cellular-connected UAV systems highly rely on accurate and secure navigation systems, e.g. the Globe Navigation System (GPS). However, civil GPS services are unencrypted and vulnerable to spoofing attacks that can manipulate UAVs’ location and abort the UAVs’ mission. This paper leverage 3D radio map and machine learning methods to detect and mitigate GPS spoofing attacks for cellular-connected UAVs. Precisely, the edge UAV flight controller uses ray tracing tools deterministic channel models, and Kriging methods to construct a theoretical 3D radio map. Then the machine learning methods, such as Multi-Layer Perceptrons (MLP), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN), are employed to detect GPS spoofing by analyzing the UAV/base station reported Received Signal Strength (RSS) values and the theoretical radio map RSS values. Once spoofing is detected, the particle filter is applied to relocate the UAV and mitigate GPS deviation. The experiment results indicate that the Universal Kriging (UK) with exponential covariance function has the lowest standard errors for radio map construction. Moreover, the MLP achieves the highest spoofing detection accuracy with different spoofing margins because of the statistic prepossessing relieving environmental impacts, while the CNN has a comparable detection accuracy with less training time than MLP since CNN inputs are raw RSS data. Furthermore, the particle filter-based GPS spoofing mitigation can relocate the UAV to the real position within an error of 10 meters using 100 particles.
  • 5G edge for power system applications
    (2023-06-15) Kokkoniemi-Tarkkanen, H.; Raussi, P.; Horsmanheimo, S.; Hovila, P.; Kulmala, A.; Borenius, S.
     A4 Artikkeli konferenssijulkaisussa
    Constantly evolving wireless technologies accelerate the trend to move from application on hardware devices to the edge. 5G standalone (SA) and beyond networks enable large-scale virtualization, which could benefit smart grids. This paper investigates the potential of 5G SA and edge technology to provide a platform for smart grid applications. We study how operating on the edge could benefit applications and provide a concrete perspective based on measurements and the process of building an edge communication pilot for smart grids. Results of preliminary QoS measurements with a commercial 5G SA network are compared with prior 5G non-standalone (NSA).
  • Accelerating XR Innovation through a pan-European Lab Network: An overview of the EMIL project
    (2024-06-12) Blönnigen, Justus; Clarke, Christopher; Dahn, Andreas; Forelli, Lisa; Gowrishankar, Ramyah; Heikura, Tuija; Helzle, Volker; Hine, Paul; Jicol, Crescent; Kreische, Alexander; Lutteroth, Christof; MacÍa, Francisco; Moesgen, Tim; Pares, Narcis; Plichta, Leszek; Potts, Dominic; Schäfer, Eduard; Sharma, Adwait; Spielmann, Simon; Tenhunen, Juhani; Trottnow, Jonas; Tseng, Yu Han; Vikberg, Esa; Xiao, Yu
     A4 Artikkeli konferenssijulkaisussa
    European Media and Immersion Lab, or EMIL, is a pan-European network of extended reality (XR) labs consisting of 4 European academic institutions, with a mission to accelerate development of virtual, augmented and mixed reality technologies, content, services and applications. The 30-month project, which started in September 2022, has been funded by the European Union and co-funded by Innovate UK. This paper gives an overview of the project's goals, its organization, and selected results that have been achieved.
  • Accurate RF-sensing of complex gestures using RFID with variable phase-profiles
    (2023) Golipoor, Sahar; Sigg, Stephan
     A4 Artikkeli konferenssijulkaisussa
    We propose the use of clothing-integrated conductive textile-based Radio Frequency Identification (RFID) tags featuring variable-phase profiles for RF-based human sensing. This approach enables the distinction and interpretation of movements from various body parts independently. We propose a scheme based on varying phase profiles in order to isolate reflections from distinct tags. The feasibility of the approach is demonstrated analytically in this work. Our instrumentation in a laboratory environment involves tag-groups attached to a solid board. The next step is the evaluation of the system when tags are mounted to a moving person.
  • Acoustic Coatings—A Discreet Way to Control Acoustic Environment
    (2021-05-11) Cucharero, Jose; Hänninen, Tuomas; Makkonen, Marko; Lokki, Tapio
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Acoustic comfort is directly related to enhanced well-being and performance of people. A typical challenge faced by architects and acousticians is to achieve adequate acoustics while maintaining the aesthetics of the space and reducing the visual aspects of acoustic materials and elements. In this study, we present a biofiber-based acoustic coating as a feasible solution to improve acoustic environments while preserving the aesthetics of spaces. An acoustic coating is a thin layer of absorption material, but the coating can be sprayed on other sound absorbing structures to make it more effective on a wide frequency range. In addition, this biofiber-based coating acts as a carbon sink during its operating life, thus reducing the carbon footprint of the building. Therefore, the coating is sustainable and is an environmental friendly solution. The absorption properties of the biofiber-based coating are demonstrated in the present study with three case studies, which all had demanding requirements to conceal the acoustic structures.
  • Acoustic Properties of Aerogels: Current Status and Prospects
    (2023-03) Budtova, Tatiana; Lokki, Tapio; Malakooti, Sadeq; Rege, Ameya; Lu, Hongbing; Milow, Barbara; Vapaavuori, Jaana; Vivod, Stephanie
     A2 Katsausartikkeli tieteellisessä aikakauslehdessä
    Noise reduction remains an important priority in the modern society, in particular, for urban areas and highly populated cities. Insulation of buildings and transport systems such as cars, trains, and airplanes has accelerated the need to develop advanced materials. Various porous materials, such as commercially available foams and granular and fibrous materials, are commonly used for sound mitigating applications. In this review, a special class of advanced porous materials, aerogels, is examined, and an overview of the current experimental and theoretical status of their acoustic properties is provided. Aerogels can be composed of inorganic matter, synthetic or natural polymers, as well as organic/inorganic composites and hybrids. Aerogels are highly porous nanostructured materials with a large number of meso- and small macropores; the mechanisms of sound absorption partly differ from those of traditional porous absorbers possessing large macropores. The understanding of the acoustic properties of aerogels is far from being complete, and experimental results remain scattered. It is demonstrated that the structure of the aerogel provides a complex three-dimensional architecture ideally suited for promising high-performance materials for acoustic mitigation systems. This is in addition to the numerous other desirable properties that include low density, low thermal conductivity, and low refractive index.
  • Acoustics and the well-being of children and personnel in early childhood education and care
    (2023-12-07) Martikainen, Silja; Prawda, Karolina; Ståhlberg-Aalto, Freja; Lautanala, Ida; Kostilainen, Kaisamari; Välimäki, Vesa; Tervaniemi, Mari
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Studies implementing a multimethod perspective in evaluating the acoustics of early childhood education and care (ECEC) spaces both quantitatively and qualitatively are still scarce. In this study the acoustic environments (noise levels and reverberation times) of seven Finnish ECEC group’s premises were examined in association with personnel’s (N = 22) and children’s (N = 71) well-being. Personnel’s well-being and vocal health and children’s well-being were assessed with questionnaires. The findings were further elaborated by documentation of the ECEC spaces and semi-structured interviews with the ECEC personnel detailing their views on the acoustic environment of the daycare buildings and how and if the acoustics should be improved. The results showed that noise exceeding 70 dB affected personnel’s vocal health negatively, whereas no associations were found regarding acoustics and children’s or personnel’s well-being. Based on the interviews, sound spreading, poor insulation, and hard surfaces add to negative experiences of noisiness. ECEC groups need spaces that can be closed and acoustically separated from each other and from other groups. The possibility to close a space supports the perceived well-being of the users and provides a more varied and individualized use of the spaces.
  • Active Acoustics With a Phase Cancelling Modal Reverberator
    (2024-10-15) De Bortoli, Gian; Prawda, Karolina; Schlecht, Sebastian
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Active acoustics (AA) systems are used to electronically modify the acoustics of a room (e.g., in live music venues). AA systems have an inherent feedback component and can suffer from instability and coloration artifacts resulting from too high feedback gains. State-of-the-art methods can improve system stability and coloration, usually at the cost of complex implementations and long parameter-tuning sessions. They can also cause sound artifacts due to time-varying components, limiting the enhancement at low frequencies. This work proposes a time-invariant feedback attenuation method for low frequencies based on a modal reverberator. The attenuation is achieved through destructive acoustic interference, obtained via phase shifts between the input and output signals. The analyzed frequency range is 0–500 Hz, where the room transfer functions are considered highly invariant over time. The results show a gain-before-instability increase of more than 5 dB for a modal reverberator with high mode density in this frequency range. The improvement is also stable for low-magnitude changes in the room transfer functions over time. The proposed method provides a robust AA system with artificial reverberation for the low-frequency range and can be used alongside other established methods.
  • AdaBoost-Based Efficient Channel Estimation and Data Detection in One-Bit Massive MIMO
    (2024) Esfandiari, Majdoddin; Vorobyov, Sergiy A.; Heath, Robert W.
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    The use of one-bit analog-to-digital converter (ADC) has been considered as a viable alternative to high resolution counterparts in realizing and commercializing massive multiple-input multiple-output (MIMO) systems. However, the issue of discarding the amplitude information by one-bit quantizers has to be compensated. Thus, carefully tailored methods need to be developed for one-bit channel estimation and data detection as the conventional ones cannot be used. To address these issues, the problems of one-bit channel estimation and data detection for MIMO orthogonal frequency division multiplexing (OFDM) system that operates over uncorrelated frequency selective channels are investigated here. We first develop channel estimators that exploit Gaussian discriminant analysis (GDA) classifier and approximate versions of it as the so-called weak classifiers in an adaptive boosting (AdaBoost) approach. Particularly, the combination of the approximate GDA classifiers with AdaBoost offers the benefit of scalability with the linear order of computations, which is critical in massive MIMO-OFDM systems. We then take advantage of the same idea for proposing the data detectors. Numerical results validate the efficiency of the proposed channel estimators and data detectors compared to other methods. They show comparable/better performance to that of the state-of-the-art methods, but require dramatically lower computational complexities and run times.
  • Adaptive Cache Policy Optimization Through Deep Reinforcement Learning in Dynamic Cellular Networks
    (2024) Srinivasan, Ashvin; Amidzade, Mohsen; Zhang, Junshan; Tirkkonen, Olav
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    We explore the use of caching both at the network edge and within User Equipment (UE) to alleviate traffic load of wireless networks. We develop a joint cache placement and delivery policy that maximizes the Quality of Service (QoS) while simultaneously minimizing backhaul load and UE power consumption, in the presence of an unknown time-variant file popularity. With file requests in a time slot being affected by download success in the previous slot, the caching system becomes a non-stationary Partial Observable Markov Decision Process (POMDP). We solve the problem in a deep reinforcement learning framework based on the Advantageous Actor-Critic (A2C) algorithm, comparing Feed Forward Neural Networks (FFNN) with a Long Short-Term Memory (LSTM) approach specifically designed to exploit the correlation of file popularity distribution across time slots. Simulation results show that using LSTM-based A2C outperforms FFNN-based A2C in terms of sample efficiency and optimality, demonstrating superior performance for the non-stationary POMDP problem. For caching at the UEs, we provide a distributed algorithm that reaches the objectives dictated by the agent controlling the network, with minimum energy consumption at the UEs, and minimum communication overhead.
  • ADMM-Based Solution for mmWave UL Channel Estimation with One-Bit ADCs via Sparsity Enforcing and Toeplitz Matrix Reconstruction
    (2023) Esfandiari, Majdoddin; Vorobyov, Sergiy A.; Heath, Robert W.
     A4 Artikkeli konferenssijulkaisussa
    Low-power millimeter wave (mmWave) multi-input multi-output communication systems can be enabled with the use of one-bit analog-to-digital converters. Owing to the extreme quantization, conventional signal processing tasks such as channel estimation are challenging, making uplink (UL) multiuser receivers difficult to implement. To address this issue, we first reformulate the UL channel estimation problem, and then combine the idea of ℓ1 regularized logistic regression classification and Toeplitz matrix reconstruction in a properly designed optimization problem. Our new method is referred to as ℓ1 regularized logistic regression with Toeplitz matrix reconstruction (L1-RLR-TMR). In addition, we develop a computationally efficient alternating direction method of multi-pliers (ADMM)-based implementation for the L1-RLR-TMR method. Numerical results demonstrate the performance of the L1-RLR-TMR method in comparison with other existing methods.
  • Advances and New Applications of Spectral Analysis
    (2023) Esfandiari, Majdoddin
     School of Electrical Engineering | Doctoral dissertation (article-based)
    Spectral analysis is a mathematical tool for modeling signals and extracting information from signals. Among the areas where it finds applications are radar, sonar, speech processing, and communication systems. One of applications of spectral analysis is to model random signals with the so-called rational models like autoregressive (AR) model. Spectral analysis is used for estimating direction-of-arrival (DOA) in array processing as well. In addition, spectral analysis is used for channel estimation in communication systems such as massive/mmWave MIMO systems. It is because channels in these systems can be modeled with multipaths' gains and angles. This thesis proposes methods for the problems of noisy AR parameter estimation, DOA estimation in unknown noise fields, and massive/mmWave MIMO channel estimation and data detection with one-bit ADCs. The AR model offers a flexible yet simple tool for modeling complex signals. In practical scenarios, the observation noise may contaminate the AR signal. In this thesis, five methods for estimating noisy AR parameter estimation are proposed. In developing the methods, the concepts such as eigendecomposition (ED) and constrained minimization are used. The most common assumption about the structure of the observation noise in DOA estimation problem is the uniform noise assumption. According to it, the noise covariance matrix is a scaled identity matrix. However, other noise covariance matrix structures such as nonuniform or block-diagonal are more accurate in some practical situations. A generalized least-squares (GLS)-based DOA estimator that takes into consideration the signal subspace perturbation and also enjoys a properly designed DOA selection strategy is proposed for the case of uniform sensor noise. For the case of nonuniform noise, a non-iterative subspace-based (NISB) method is developed which is computationally efficient compared to state-of-the-art competitors. Moreover, a unified approach to DOA estimation in uniform, nonuniform, and block-diagonal sensor noise is presented. The use of one-bit ADCs instead of high-resolution ADCs is considered as an elegant solution for reducing power consumption of large-scale systems such as massive/mmWave MIMO and radar systems. In this thesis, we use the analogy between binary classification problem and one-bit parameter estimation to develop algorithms for massive MIMO and mmWave systems. In this regard, a method called SE-TMR is developed for one-bit mmWave UL channel estimation. Another method called L1-RLR-TMR is also offered for one-bit mmWave UL channel estimation. At last, the concept of AdaBoost is combined with Gaussian discriminant analysis (GDA) for developing computationally very efficient channel estimators and data detectors. It is shown that the proposed methods which use approximate versions of GDA as weak classifiers in iterations of the AdaBoost-based algorithms are exceptionally efficient, specifically in large-scale systems.
  • Advancing Audio Emotion and Intent Recognition with Large Pre-Trained Models and Bayesian Inference
    (2023-10-27) Porjazovski, Dejan; Getman, Yaroslav; Grósz, Tamás; Kurimo, Mikko
     A4 Artikkeli konferenssijulkaisussa
    Large pre-trained models are essential in paralinguistic systems, demonstrating effectiveness in tasks like emotion recognition and stuttering detection. In this paper, we employ large pre-trained models for the ACM Multimedia Computational Paralinguistics Challenge, addressing the Requests and Emotion Share tasks. We explore audio-only and hybrid solutions leveraging audio and text modalities. Our empirical results consistently show the superiority of the hybrid approaches over the audio-only models. Moreover, we introduce a Bayesian layer as an alternative to the standard linear output layer. The multimodal fusion approach achieves an 85.4% UAR on HC-Requests and 60.2% on HC-Complaints. The ensemble model for the Emotion Share task yields the best value of .614. The Bayesian wav2vec2 approach, explored in this study, allows us to easily build ensembles, at the cost of fine-tuning only one model. Moreover, we can have usable confidence values instead of the usual overconfident posterior probabilities.
  • Adversarial Guitar Amplifier Modelling with Unpaired Data
    (2023-06-10) Wright, Alec; Välimäki, Vesa; Juvela, Lauri
     A4 Artikkeli konferenssijulkaisussa
    We propose an audio effects processing framework that learns to emulate a target electric guitar tone from a recording. We train a deep neural network using an adversarial approach, with the goal of trans-forming the timbre of a guitar, into the timbre of another guitar after audio effects processing has been applied, for example, by a guitar amplifier. The model training requires no paired data, and the resulting model emulates the target timbre well whilst being capable of real-time processing on a modern personal computer. To verify our approach we present two experiments, one which carries out un-paired training using paired data, allowing us to monitor training via objective metrics, and another that uses fully unpaired data, corresponding to a realistic scenario where a user wants to emulate a guitar timbre only using audio data from a recording. Our listening test results confirm that the models are perceptually convincing.
  • Affine Equivariant Tyler's M-Estimator Applied to Tail Parameter Learning of Elliptical Distributions
    (2023-08-03) Ollila, Esa; Palomar, Daniel P.; Pascal, Frederic
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    We propose estimating the scale parameter (mean of the eigenvalues) of the scatter matrix of an unspecified elliptically symmetric distribution using weights obtained by solving Tyler's M-estimator of the scatter matrix. The proposed Tyler's weights-based estimate (TWE) of scale is then used to construct an affine equivariant Tyler's M-estimator as a weighted sample covariance matrix using normalized Tyler's weights. We then develop a unified framework for estimating the unknown tail parameter of the elliptical distribution (such as the degrees of freedom (d.o.f.) ν of the multivariate t (MVT) distribution). Using the proposed TWE of scale, a new robust estimate of the d.o.f. parameter of MVT distribution is proposed with excellent performance in heavy-tailed scenarios, outperforming other competing methods. R-package is available that implements the proposed method.
  • Ahead-Me Coverage (AMC): On Maintaining Enhanced Mobile Network Coverage for UAVs
    (2023-01-11) Hellaoui, Hamed; Yang, Bin; Taleb, Tarik; Manner, Jukka
     A4 Artikkeli konferenssijulkaisussa
    This paper proposes the concept of Ahead-Me Cov-erage (AMC) aiming to get the coverage of a cellular network ahead of the mobile users for maintaining enhanced Quality- of-Service (QoS) in cellular-connected unmanned aerial vehicle (UAV) networks. In such networks, each base station (BS) with an intelligent logic can automatically tilt the direction of its radio antennas based on the trajectory of UAV s. For this purpose, we first formulate AMC as an integer optimization problem for maximizing the minimum transmission rate of UAVs by jointly optimizing the angles of the different radio antenna, the resource allocation and the selection of the appropriate serving BS for the UAVs throughout their path. For this complex optimization problem, we then propose a solution based on Deep Reinforcement Learning (DRL) to solve it. Under this solution, we adopt a multi-heterogeneous agent-based approach (MHA-DRL) including two types of agents, namely the UAV agents and the BS agents. Each agent implements an Advantage Actor Critic (A2C) to learn optimal policies. Specifically, the BS agents aim to tilt their antennas to get ahead of the UAV s throughout their mobility, and the UAV agents target selecting the appropriate serving BSs along with resource allocation. Performance evaluations are presented to validate the effectiveness of the proposed approach.
  • An AI-based Framework to Optimize Mobile Services
    (2023) Naas, Si-Ahmed
     School of Electrical Engineering | Doctoral dissertation (article-based)
    In current digital networks, novel verticals require diversity and extreme efficiency from network infrastructures. Recent studies have shown the key role of Artificial intelligence (AI) to improve resource management and to reduce operational costs of mobile services and networks. Furthermore, the rapid development of connected Internet of Things devices has created a variety of services. Since the complexity of mobile services has increased, efficient designs for these services are required to satisfy increasing power and processing requirements. Mobile services, however, face numerous challenges, including a shortage of data, inefficient data processing, and a lack of multi-service collaboration mechanisms. These setbacks have delayed the development of commercial applications that rely on mobile networks. Therefore, this thesis presents a framework, based on the "global brain" concept, to orchestrate various mobile services. The framework enhances mobile services for two scenarios: 1) low data rate (e.g., ambient intelligence) and 2) high data rate (e.g., virtual reality (VR) streaming). The distinction between low and high data rates is based upon bandwidth requirement. For low data rate scenarios, this thesis investigates the usefulness of emotion-based data for static (e.g., customer satisfaction) and dynamic (e.g., location-based recommendations) environments. We begin by presenting a practical multimodal emotion recognition system based on audio and video information, which improves emotion recognition accuracy. The results indicate that the integration of emotions is vital for customer satisfaction assessment and recommendation systems. For high data rate scenarios, we address VR streaming due to its high computational cost and latency. Notably, improved viewport (VP) and gaze prediction schemes are critical for enabling smoother VR streaming, as these features are the foundation of the user's experience while wearing the headset. Thus, the global brain proposes novel VP and gaze prediction schemes which produce high prediction accuracy and reduced resource consumption. Finally, this thesis proposes a novel knowledge sharing mechanism that enables mobile services to improve their models by learning from others and incorporating global knowledge, which significantly increases inference accuracy. To prevent sensitive information leakage during knowledge sharing among service providers, we propose a lightweight multi-key homomorphic encryption scheme that allows mobile services to protect their knowledge (i.e., deep neural network weights).
  • AI/ML Service Enablers & Model Maintenance for Beyond 5G Networks
    (2023-09) Samdanis, Konstantinos; Nait Abbou, Aiman; Song, JaeSeung; Taleb, Tarik
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    Artificial Intelligence and Machine Learning (AI/ML) can transform mobile communications, enable new applications and services, and pave the way beyond 5G. The adoption of AI/ML may also advance network optimizations and service life-cycle management. This article provides an overview of AI/ML analytics exploring the main concepts in relation to network automation and service-based architectures. Furthermore, it sheds light on AI/ML analytics service enablers by considering a system-level approach elaborating on the key technologies related to service discovery, request, control, and reporting of AI/ML analytics. Finally, it provides an analysis for maintaining the services of AI/ML analytics up-to-date, by considering modifications of the AI/ML model in order to detect, interpret, and compensate for potential performance drifts.
  • Amortized Inference with User Simulations
    (2023-04-19) Moon, Hee Seung; Oulasvirta, Antti; Lee, Byungjoo
     A4 Artikkeli konferenssijulkaisussa
    There have been significant advances in simulation models predicting human behavior across various interactive tasks. One issue remains, however: identifying the parameter values that best describe an individual user. These parameters often express personal cognitive and physiological characteristics, and inferring their exact values has significant effects on individual-level predictions. Still, the high complexity of simulation models usually causes parameter inference to consume prohibitively large amounts of time, as much as days per user. We investigated amortized inference for its potential to reduce inference time dramatically, to mere tens of milliseconds. Its principle is to pre-train a neural proxy model for probabilistic inference, using synthetic data simulated from a range of parameter combinations. From examining the efficiency and prediction performance of amortized inference in three challenging cases that involve real-world data (menu search, point-and-click, and touchscreen typing), the paper demonstrates that an amortized-inference approach permits analyzing large-scale datasets by means of simulation models. It also addresses emerging opportunities and challenges in applying amortized inference in HCI.
  • Angle-Agnostic Radio Frequency Sensing Integrated into 5G-NR
    (2024) Salami, Dariush; Hasibi, Ramin; Savazzi, Stefano; Michoel, Tom; Sigg, Stephan
     A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
    The fusion of radio frequency (RF) sensing with cellular communication networks presents a revolutionary paradigm, enabling networks to seamlessly integrate communication and perception capabilities. Leveraging electromagnetic radiation, this technology facilitates the detection and interpretation of human movements, activities, and environmental changes. This article proposes a novel implementation of RF sensing within the allocated resources for new radio (NR) sidelink direct device-to-device (D2D) communication, showcasing the synergy between RF sensing and machine-learning (ML) techniques. The article addresses the inherent challenge of angle dependency in the sidelink-enabled sensing scheme, and introduces innovative solutions to achieve angle-agnostic environmental perception. The proposed approach incorporates a graph-based encoding of movement and gesture sequences, capturing spatio-temporal relations, and integrates orientation tracking to enhance human gesture recognition. The proposed model surpasses state-of-the-art algorithms, demonstrating a remarkable 100% accuracy in RF sensing when all the angles are available. Although the performance of our proposed method does decline with fewer available angles, it demonstrates exceptional resilience to missing data. Specifically, our model significantly outperforms existing models by approximately 70% in scenarios where seven out of eight angles are unavailable. To further advance sensing capabilities in RF sensing systems, a comprehensive dataset comprising 15 subjects performing 21 gestures, recorded from eight different angles, is openly shared. This contribution aims to enhance the performance and reliability of RF sensing systems by providing a robust and efficient ML-driven solution for human gesture recognition within NR sidelink D2D communication networks, aligning with the latest advancements in ML for RF sensing applications.