Sensing-assisted communication is a feasible technology for the rapid development of 5G and 6G communication systems, which can achieve greater channel capacity with limited energy consumption. In this letter, we propose a deep learning-based beam selection and power allocation algorithm (DL-BSPA), which includes a beam selection network and a power allocation network. The beam selection network uses real millimeter wave radar data to learn how to accurately and efficiently select beams. An unsupervised power allocation network learns and optimizes the power allocation based on the results of the beam selection network. In addition, the DL-BSPA algorithm adopts the joint deployment optimization method to maximize the channel capacity between mobile users and base stations (BS). Simulation results show that the proposed scheme is robust and can achieve competitive performance.

In developing modern intelligent transportation systems, integrated sensing and communication (ISAC) technology has become an efficient and promising method for vehicle road services. To enhance traffic safety and efficiency through real-time interaction between vehicles and roads, this paper proposes a searching–deciding scheme for an alternation radar-communication (radar-comm) system. Firstly, its communication performance is derived for a given detection probability. Then, we process the echo data from real-world millimeter-wave (mmWave) radar into four-dimensional (4D) point cloud datasets and thus separate different hybrid modes of single-vehicle and vehicle fleets into three types of scenes. Based on these datasets, an efficient labeling method is proposed to assist accurate vehicle target detection. Finally, a novel vehicle detection scheme is proposed to classify various scenes and accurately detect vehicle targets based on deep learning methods. Extensive experiments on collected real-world datasets demonstrate that compared to benchmarks, the proposed scheme obtains substantial radar performance and achieves competitive communication performance.

Traditional beam prediction approaches are suitable for dealing with point-like targets rather than extended targets, which are commonly detected by the millimeter-wave (mmWave) radar. To solve this problem, a novel adaptive beam prediction approach for complex real-world mmWave vehicle-to-infrastructure (V2I) communications is proposed in this article. By adopting a vehicle kinematic model based on the axle load point and combining it with an adaptive extended Kalman filter (AEKF) algorithm, the proposed approach can accurately predict and track vehicle motion states and beam directions in time division duplexing (TDD) mode. The AEKF algorithm is employed to dynamically update the process and measurement noise covariance matrices to reduce the impact of various noises in real-world scenarios, which can enhance the accuracy of beam prediction, and then improve the received signal-to-noise ratio (SNR) and communication rate. During the downlink phase, the roadside unit (RSU) formulates transmit beam by using predicted beam direction, and then transmits integrated sensing and communication (ISAC) signals to the vehicle. After receiving the echo point clouds reflected by the vehicle body, the RSU can estimate its relative direction and motion states through the AEKF algorithm. In the uplink phase, the RSU uses the above estimated states to further predict the direction and vehicle states, and then estimate and predict them through the AEKF algorithm by using the received uplink signal. Simulation results and DeepSense 6G dataset-based test results show that the proposed approach offers significant improvements in both sensing and communication performance compared with traditional approaches.

Autonomous aerial vehicle (AAVs) are increasingly utilized in smart city applications, such as Artificial Intelligence of Things (AIoT) systems due to their agility, low cost, and independence from ground road conditions. However, they require substantial computing resources and electricity consumption as support, leading to significant carbon emissions and energy consumption. This article presents a novel sustainable scheme for AAV-assisted low-altitude communication enabled by radar point clouds (RPCs) and the moving interaction station (MIS) to address this challenge. Vehicles are represented as rigid shapes with the 4-D RPC, which is close to the real world, and AAVs can ride on buses defined as MIS to conserve energy. Then, a joint optimization of the AAV trajectory, vehicle association, and resource allocation problem is formulated to complete the communication task and minimize energy consumption, which is limited by the Age of Information (AoI) indicators. An advanced hierarchical deep reinforcement learning (HDRL) algorithm containing a central control module and five submodules is proposed to solve the proposed optimization problem. The five submodules optimize the AAV trajectory, vehicle association, and resource allocation to complete the communication task with the constraint of AoI indicators. The central control module manages five submodules to minimize energy consumption by scheduling time. Simulation results demonstrate that the proposed scheme is effective for smart city applications, and the HDRL algorithm achieves competitive performance compared with the benchmarks.

The integrated sensing and communication-enhanced (ISAC-enhanced) intelligent transportation system (ITS) has great development potential and market value in future urban vehicle-to-infrastructure (V2I) scenarios. With limited total power consumption, joint design of power distribution and beam steering of sensing and communication subsystems is required to achieve optimal system performance. This paper introduces an intelligent power allocation and beam-steering control system, where both the system actions are determined based on the radar sensing data cube. The adaptive power allocation module employs the Actor-Critic reinforcement learning agents to optimize power distribution based on vehicle GPS data, communication channel capacity, and historical power allocation, balancing the overall system performance including the sensing and communication functions. The automatic beam control module, utilizing the convolutional neural network (CNN) for dynamic beam control, intelligently adapts communication beams for vehicles navigating complex roadways, enhancing the beam tracking accuracy and quality of service. Validation on real-world datasets showcases the effectiveness of these modules, highlighting the potential to enhance vehicular communication performance.