With integrated sensing and communications (ISAC) and reconfigurable intelligent surface (RIS) emerging as critical enablers for future mobile communications, their combination has attracted increasing attention lately. To effectively utilize RIS for improving ISAC, we propose two novel designs targeting different scenarios. The first design strikes for a spectral-efficient ISAC by seeking to maximize the weighted sum rate (WSR) of communications and minimize sensing radiation pattern approximation error. The second design aims to achieve an energy-efficient ISAC by optimizing the power allocation between communications and sensing subject to communication quality of service (QoS) constraints. Different optimization problems are formulated for the two designs, with practical constraints of RIS considered, including unit modulus and discrete phase shift. Efficient solutions are developed for the non-convex optimization problems by adeptly employing techniques including weighted minimum mean squared error (WMMSE), fractional programming (FP), second-order cone programming (SOCP), semi-definite relaxing (SDR), and feasibility check. Simulation results demonstrate the non-trivial improvements of WSR, communication energy efficiency and sensing radiation patterns achieved by the proposed designs, also highlighting their superiority over the conventional methods.

The mobile edge computing (MEC) heterogeneous network (HetNet) can create a high-performance and low-latency environment for users. However, due to the limitation of small base stations (SBSs) cache and wireless backhaul transmission capacity, how to make the service caching and task offloading decisions and reduce the task offloading time are still key issues to be addressed to achieve decent system computing delay performance. To this end, this paper investigates a reconfigurable intelligent surface (RIS)-aided wireless backhaul MEC HetNet system. The joint service caching, task offloading, and resource allocation problem is designed to minimize the total latency. The joint optimization problem is an NP-hard mixed-integer nonlinear programming problem, which is difficult to tackle directly. We decompose the optimization problem into two sub-problems, i.e., the communication and the computing sub-problems, and propose algorithms to solve them. Numerical results show that the proposed schemes can effectively reduce the total delay by 6.79% and 18.29% in wireless backhaul MEC HetNet in comparison with the scheme without RIS and local computing, respectively.

Target detection technology has been greatly improved for the synthetic aperture radar (SAR) images recently, due to the advancement in deep learning (DL) domain. However, because of the existence of clutter in the SAR images, it's still a challenge to detect small targets with high accuracy and low computational complexity. To solve this problem, a detection algorithm based on feature fusion and cross-layer connection (FFCLC) network is proposed in this paper. Firstly, the attention feature fusion (AFF) is applied to improve the feature fusion ability for the small targets through allocating weights to various feature maps adaptively. Meanwhile, the depthwise separable convolution (DW-Conv) is used to reduce the computational complexity caused by the increasement of network layers. Then, a cross-layer connection (Cross-Connect) submodule is proposed to fuse shallow features with deep features further. Finally, a multi-scale target detection (Multi-Detect) submodule is designed to improve the detection ability for the small targets. We compare the proposed algorithm with the other representative methods on the SAR-Ship-Dataset and SSDD, quantitative evaluations show that our proposed algorithm's can reach the highest computational efficiency. Therefore, because of the superior performance in terms of accuracy and efficiency, the algorithm proposed in this paper is more suitable to detect small targets for the SAR images.

The unmanned aerial vehicle (UAV) with intelligent reflecting surface (IRS) mounted, namely, aerial IRS, has the potential in improving the information freshness (IF) and transmission data rate for wireless networks, where age of information (AoI) is generally utilized to characterize the IF. In this article, we optimize both the AoI and transmission data rate in aerial IRS-assisted Internet of Things (IoT) networks through the joint transmission scheduling, UAV location, and IRS phase shift matrix design. We formulate a multiobjective optimization problem, which simultaneously minimizes the system average AoI and maximizes the overall transmission data rate. The optimal solutions for the two objectives in the formulated problem are not always consistent with each other. Besides, the optimization problem with either objective is nonconvex and difficult to tackle directly. An effective three-step scheme is developed in this article to solve the formulated problem. To be more specific, firstly the UAV locations are optimized through a $Q$ -learning-based scheme to maximize the overall data rate while guaranteeing the signal-to-noise ratio (SNR) constraint of each IoT device; then the IRS phase shift matrices are determined through a low-complexity Tabu-search-based scheme to further improve the overall data rate given the SNR constraint of each device; finally, the transmission scheduling is performed to optimize the system AoI based on a deep $Q$ -network algorithm. Simulation evaluation demonstrates that the proposed scheme outperforms existing ones.

The spatial covariance matrix, as the core of subspace direction of arrival (DOA) estimation, is unavailable in the hybrid analog-digital structure, which is widely used in large-scale transceivers to reduce the number of radio frequency chains. In previous studies, the beam sweeping algorithm (BSA) has been used for effectively reconstructing the spatial covariance matrix. However, the BSA is not practical owning to the intractable computational cost due to the high-dimensional matrix operations. To address this problem, a high-efficiency BSA (HeBSA) is proposed. The main motivation behind the HeBSA is to avoid the high-dimensional matrix multiplication and inversion. By appropriately adjusting weights connected to antennas, the real-valued spatial covariance matrix can be reconstructed according to low-dimensional matrix multiplication. Then the classical U-root multiple signal classification can be adopted to estimate the DOA angle of each signal. Finally, simulation experiments are conducted to verify the performance of the HeBSA. The results indicate that the HeBSA can achieve DOA estimation performance comparable with that of BSA even at considerably lower computational cost.