This research paper delves into leveraging Machine Learning (ML) for precise localization in GPS-challenged environments like urban canyons, addressing the complexities of time-varying signal propagation types, where transient obstructions, such as vehicles, can modify the channel state information (CSI) over time. It presents a novel approach termed semantic localization, which recognizes signal propagation conditions as semantic elements, incorporating them into the localization framework to enhance both accuracy and resilience. To tackle the issue of diverse CSIs at each location and the extensive need for labeled data, the paper proposes a multi-task deep domain adaptation (DA) strategy. This approach trains neural networks using a limited set of labeled data complemented by a vast array of unlabeled samples, coupled with innovative scenario adaptive learning techniques for optimal representation learning and knowledge transfer. Employing Bayesian theory for the efficient management of task importance weights minimizes the necessity for laborious parameter tuning. By making certain assumptions, the study introduces a deep homoscedastic DA method for enhanced joint task efficacy. Through detailed simulations using a 3D ray tracing dataset, the paper evidences that the integration of environmental semantics and the advanced DA localization techniques markedly elevates the precision of localization in various demanding settings.

This paper investigates the use of machine learning to a particularly challenging wireless localization problem, namely localization in the presence of high environmental dynamics. We first introduce the semantic localization scheme to achieve this goal, adding time-varying environmental semantics recognition as a joint task with localization. The environmental semantics is characterized by the propagation conditions between the base station (BS) and the receiver(s). To reduce the negative impact of high dynamics, our method is based on the deep unsupervised domain adaptation (UDA) techniques, which jointly train the network with labeled and unlabeled samples. This also greatly reduces the effort during data collection since the acquisition of unlabeled data requires much smaller effort than labeled data. We further propose a deep multi-task UDA method with a novel scenario adaptive learning strategy, enhancing the optimization process of the neural network with the environmental semantics. Lastly, we evaluate the proposed method over the widely used deepMIMO dataset that is generated from detailed ray tracing. The experimental results demonstrate the superior performance of semantic localization over standalone localization; the proposed method outperforms the competing ones over multiple experimental conditions.

Estimating vehicle locations is important for realizing Intelligent Transportation Systems (ITS). This paper considers utilizing vehicle-to-vehicle (V2V) communication for relative vehicular localization. In particular, we develop a machine learning (ML) solution that uses the Channel State Information (CSI) from multiple-antenna transceivers for vehicular localization. We develop suitable pre-processing to obtain a compact CSI representation as an input feature to the ML solution. The proposed solution is then based on feed-forward neural networks. Training and evaluation are done on measured real-world data in the 5.9 GHz band. The performance on two routes shows that the proposed feature may improve the performance while reducing the number of trainable parameters. Furthermore, the paper raises a number of interesting observations regarding the learnability in V2V ML-based localization solutions.

Many wireless networks, including 5G NR (New Radio) and future beyond 5G cellular systems, are expected to operate on multiple frequency bands. This paper considers the band assignment (BA) problem in dual-band systems, where the basestation (BS) chooses one of the two available frequency bands (centimeter-wave and millimeter-wave bands) to communicate with the user equipment (UE). While the millimeter-wave band might offer higher data rate, there is a significant probability of outage during which the communication should be carried on the (more reliable) centimeter-wave band. With mobility, the BA can be perceived as a sequential problem, where the BS uses previously observed information to predict the best band for a future time step. We formulate the BA as a binary classification problem and propose supervised Machine Learning (ML) solutions. We study the problem when both the BS and the UE use (i) omnidirectional antennas and (ii) both use directional antennas. In the omnidirectional case, we derive analytical benchmark solutions based on the Gaussian Process (GP) assumption for the inter-band shadow fading. In the directional case, where the labeling is shown to be complex, we propose an efficient labeling approach based on the Viterbi Algorithm (VA). We compare the performances for two channel models: (i) a stochastic channel and (ii) a ray-tracing based channel.