Tracking beam and location in wireless networks

Abstract

To meet the stringent requirements in manufacturing, the fifth generation (5G) is integrated with the production process since it supports massive machine type communication (mMTC), ultra-reliable low-latency communications (URLLC), and enhanced mobile broadband (eMBB). Multi-antenna millimeter-wave (mmWave) systems in 5G, promising for higher throughput, struggle with blockage issues in narrow beams, affecting service continuity for URLLC. To mitigate this, we introduce a dual-path beam tracking framework, employing a Recurrent Neural Network-based Constrained Deep Reinforcement Learning algorithm, which efficiently allocates time and frequency resources for beam sweeping, tracking, and data transmission while maintaining URLLC service interruption probability constraints.

Next, we further explore beam prediction in multi-antenna systems for mission-critical applications in IIoT using millimeter-wave bands. Continuous service requires periodic beam sweeping by the base station, which leads to resource inefficiency and potential beam misalignment. To address these issues, we propose a beam prediction architecture attention mechanism, with which the next beam direction and signal-to-noise ratio (SNR) are predicted based on different observations of channel state information.

Finally, this thesis introduces a novel generative fusion framework for indoor localization, which is crucial for the growing need for precise positioning in IIoT. Utilizing generative neural networks, our approach effectively combines different measurements, making it ideal for new applications. The framework separates round-trip time (RTT) and inertial measurement unit (IMU) measurements into individual generative models. This allows for creating multiple potential positions while maintaining diversity in location predictions. Additionally, we incorporate an attention-based fusion model to merge positions generated from various measurements efficiently, enhancing overall localization precision.