AI4C3 – Multi-connectivity Wieless Communications
Thursday, 5 June 2025, 11:00-12:30, room 1.G
Session Chair: Jordi Mongay Batalla (Warsaw Univ. Technology, PL)
Multi-User Beamforming with Deep Reinforcement Learning in Sensing-Aided Communication
Xiyu Wang and Gilberto Berardinelli (Aalborg University, Denmark); Hei Victor Cheng (Aarhus University, Denmark); Petar Popovski and Ramoni O Adeogun (Aalborg University, Denmark)
Mobile users are prone to experience beam failure due to beam drifting in millimeter wave (mmWave) communications. Sensing can help alleviate beam drifting with timely beam changes and low overhead since it does not need user feedback. This work studies the problem of optimizing sensing-aided communication by dynamically managing beams allocated to mobile users. A multi-beam scheme is introduced, which allocates multiple beams to the users that need an update on the angle of departure (AoD) estimates and a single beam to the users that have satisfied AoD estimation precision. A deep reinforcement learning (DRL) assisted method is developed to optimize the beam allocation policy, relying only upon the sensing echoes. For comparison, a heuristic AoD-based method using approximated Cramér-Rao lower bound (CRLB) for allocation is also presented. Both methods require neither user feedback nor prior state evolution information. Results show that the DRL-assisted method achieves a considerable gain in throughput than the conventional beam sweeping method and the AoD-based method, and it is robust to different user speeds.
Dual-Task Supervised Learning for Distributed Spectrum Monitoring Applications
Victor Shatov (Friedrich-Alexander University Erlangen-Nürnberg, Germany); Nikita Shanin (Friedrich-Alexander University of Erlangen-Nuremberg, Germany); Tobias Veihelmann and Bastian Perner (Friedrich-Alexander Universität Erlangen-Nürnberg, Germany); Norman Franchi (Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany); Maximilian Lübke (Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany)
This paper examines a novel data-driven framework for spectrum monitoring in the context of non-public networks. To this end, we model an indoor scenario with a single non-cooperative transmitter (TX) and a spatially distributed network of sensor units (SUs) capable of recording the broadcasted signal as I/Q samples. Then, a dual-task deep neural network (DNN) is designed to jointly classify wireless signal waveform and estimate the TX coordinates, using raw data from the SUs as input. The numerical simulations focus on the model’s scalability, generalizability, and robustness. In this context, we study the effect of the number of SUs and varying signal-to-noise ratio (SNR) on the accuracy of the developed DNN-based joint signal classification and TX localization algorithm. In particular, we show that an increased number of SUs improves the localization accuracy, whereas the performance of waveform classification remains nearly unchanged. Finally, we assess the model’s complexity regarding the number of learnable parameters compared to single-task DNNs, demonstrating that extra functionality can be obtained at a low additional complexity cost.
Enhancing URLLC Availability in Multi-Connectivity Scenarios Using Deep Reinforcement Learning
Sheikh Tawsiful Islam (KTH Royal Institute of Technology, Sweden); Milad Ganjalizadeh (Ericsson AB, Sweden); Hossein Shokri Ghadikolaei (Ericsson Research, Sweden); Mustafa Ozger (Aalborg University, Denmark & KTH Royal Institute of Technology, Sweden)
Ultra-reliable low-latency communication (URLLC) services are essential for real-time control applications, such as industrial automation and autonomous vehicles, where stringent performance and reliability are paramount. Traditional diversity techniques-employing time, frequency, or spatial domains-enhance communication service availability. In bandwidth-constrained systems, these techniques often result in redundant transmissions and excessive resource consumption, limiting the efficient utilization of available resources. This paper investigates the potential of dynamic spatial diversity to enhance the availability of URLLC services in multi-connectivity scenarios. To this end, we propose to employ an entropy-based deep reinforcement learning framework. This framework leverages the soft actor-critic algorithm to dynamically optimize spatial diversity by selecting transmission paths and determining the optimal number of packet instances for transmission. The proposed approach, implemented in a 3GPP-compliant simulator, is evaluated in a factory automation scenario employing dual connectivity and packet duplication. The experiments demonstrate that our framework significantly outperforms conventional single-path and static packet duplication strategies, achieving superior efficiency in packet duplication and load balancing.
Learning-Based Multiuser Scheduling in MIMO-OFDM Systems with Hybrid Beamforming
Pouya Agheli (EURECOM, France); Tugce Kobal (Nokia Bell Labs, United Kingdom (Great Britain)); François Durand (Nokia Bell Labs France, France); Matthew Andrews (Nokia Bell Labs, USA)
We investigate the multiuser scheduling problem in multiple-input multiple-output (MIMO) systems using orthogonal frequency division multiplexing (OFDM) and hybrid beamforming in which a base station (BS) communicates with multiple users over millimeter wave (mmWave) channels in the downlink. Improved scheduling is critical for enhancing spectral efficiency and the long-term performance of the system from the perspective of proportional fairness (PF) metric in hybrid beamforming systems due to its limited multiplexing gain. Our objective is to maximize PF by properly designing the analog and digital precoders within the hybrid beamforming and selecting the users subject to the number of radio frequency (RF) chains. Leveraging the characteristics of mmWave channels, we apply a two-timescale protocol. On a long timescale, we assign an analog beam to each user. Scheduling the users and designing the digital precoder are done accordingly on a short timescale. To conduct scheduling, we propose combinatorial solutions, such as greedy and sorting algorithms, followed by a machine learning (ML) approach. Our numerical results highlight the trade-off between the performance and complexity of the proposed approaches. Consequently, we show that the choice of approach depends on the specific criteria within a given scenario.
Towards Smarter Sensing: 2D Clutter Mitigation in RL-Driven Cognitive MIMO Radar
Adam Umra (Ruhr University Bochum, Germany); Aya Ahmed (Ruhr University Bochum, USA); Aydin Sezgin (RUB, Germany)
Motivated by the growing interest in integrated sensing and communication for 6th generation (6G) networks, this paper presents a cognitive Multiple-Input Multiple-Output (MIMO) radar system enhanced by reinforcement learning (RL) for robust multitarget detection in dynamic environments. The system employs a planar array configuration and adapts its transmitted waveforms and beamforming patterns to optimize detection performance in the presence of unknown two-dimensional (2D) disturbances. A robust Wald-type detector is integrated with a SARSA-based RL algorithm, enabling the radar to learn and adapt to complex clutter environments modeled by a 2D autoregressive process. Simulation results demonstrate significant improvements in detection probability compared to omnidirectional methods, particularly for low Signal-to-Noise Ratio (SNR) targets masked by clutter.
