Owing to their controlled mobility, extensive coverage and low cost, unmanned aerial vehicles (UAVs) have significant promise in the field of military, civilian and commercial applications, which has sparked a lot of research interest. This paper explores the application of the UAV in the symbiotic radio (SR) system. Specifically, we study the multi-user SR system assisted by the UAV and intelligent reflecting surface (IRS), and the modulation scheme with high flexibility is adopted to embed the IRS symbols into the UAV signals. First, we focus on minimizing the maximum bit error rate (BER) among all ground devices (GDs) while meeting the minimum communication rate requirements by jointly designing the active transmit beamforming and flight trajectory of the UAV as well as IRS reflective phase shifts. Then, we develop an iterative algorithm based on block coordinate descent (BCD) technology to address this optimization problem. Specifically, the optimization problem is decoupled into three sub-problems to be alternately solved, which are solved by invoking the semidefinite relaxation (SDR) and successive convex approximation (SCA) techniques. Moreover, the convergence performance and computational complexity of the proposed algorithm are analyzed. Finally, simulation results show the proposed algorithm can significantly improve the quality of IRS transmission.

Due to the high propagation path loss in the millimeter wave (mm-wave) system, it is necessary to use the large antenna array for obtaining the high beamforming gain in mm-wave systems. In order to reduce the cost and complexity of hardware, an architecture called hybrid precoding is proposed, which divides the precoding operation into the analog domain and the digital domain. However, this poses a challenge for channel estimation in mm-wave systems. In addition, due to the great amount of users and antennas, the feedback overhead of channel estimation is also an issue that needs to be considered. In the letter, we consider the mm-wave massive multiple-input-multiple-output (MIMO) system and propose a federated learning (FL)-base channel state information (CSI) estimation and feedback (FCEF) scheme, where each user trains the local model through the local dataset and exchanges the model parameters with the base station (BS). Thanks to upload model parameters rather than huge channel data, the FCEF scheme has lower transmission overhead than centralized learning (CL). The architecture of the proposed FCEF model is described in the letter, which consists of a CSI estimation and compression (CEC) network and a CSI recovery (CR) network. Numerical results demonstrate that the proposed FCEF scheme has good performance for CSI estimation and feedback in mm-wave MIMO systems, which is close to that of the CL-based scheme while ensuring lower transmission overhead.

In this letter, we investigate a reconfigurable intelligent surface (RIS)-aided integrated sensing and communication system enhanced by rate-splitting multiple access. The RIS is deployed to establish a virtual line-of-sight link for target detection, which also act as a potential eavesdropper, and to assist in user communication. To ensure secure communication and detection, we design the communication and radar precoders, allocate a common stream to serve as artificial noise for eavesdroppers at the base station, and adjust the phase shifts of the RIS to maximize the minimum secrecy rate while being subject to constraints on transmit and sensing power. To this end, an efficient alternating optimization algorithm, which leverages semidefinite relaxation, S-procedure, successive convex approximation and rank-one penalty methods, is developed to address this non-convex problem. Simulation results demonstrate that the proposed scheme achieves a superior communication security. And, the results also indicate that, based on user density and radar requirements, the addition or removal of radar sequences can achieve a balance between system complexity and security.

Symbiotic radio (SR), a new green communication technology, is proposed to address staggering amount of energy and spectrum resource requirements from massive internet of things (IoT) devices, where some passive backscatter devices form backscatter-link transmissions. The aim of this letter is to solve the pilot design and channel estimation issues for the SR system. To this end, we propose a novel joint pilot design and channel estimation (JPDCE) scheme based on deep learning, which consists of the pilot designer and channel estimator. Moreover, deep residual network and the successive interference cancellation (SIC) technique are used to eliminate the noise and the interference between different channels, respectively. Numberous simulation results show that the proposed scheme has superior performance of channel estimation.

Power equipment maintenance work orders are vital in power equipment management because they contain detailed information such as equipment specifications, defect reports, and specific maintenance activities. However, due to limited research into automated information extraction, valuable operational and maintenance data remain underutilized. A key challenge is recognizing unstructured Chinese maintenance texts filled with specialized and abbreviated terms unique to the power sector. Existing named entity recognition (NER) solutions often fail to effectively manage these complexities. To tackle this, this paper proposes a NER model tailored to power equipment maintenance work orders. First, a dataset called power equipment maintenance work orders (PE-MWO) is constructed, which covers seven entity categories. Next, a novel position- and similarity-aware attention module is proposed, where an innovative position embedding method and attention score calculation are designed to improve the model’s contextual understanding while keeping computational costs low. Further, with this module as the main body, combined with the BERT-wwm-ext and conditional random field (CRF) modules, an efficient NER model is jointly constructed. Finally, validated on the PE-MWO and five public datasets, our model shows high accuracy in recognizing power sector entities, outperforming comparative models on public datasets.