一种基于现场预训练的无人机信道到达角估计方法
By training the neural network in UAV communication using on-site pre-training, a backpropagation neural network is constructed, which solves the problems of unstable performance and excessive computation time of UAV communication angle estimation methods under different scenarios and hardware systems, and achieves efficient angle estimation.
CN117674917BActive Publication Date: 2026-07-17NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
- Filing Date
- 2023-11-21
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing UAV communication angle estimation methods have unstable performance under different scenarios and hardware systems, and traditional methods take too long to compute, making it difficult to meet the requirements of real-time performance and low latency.
Method used
A UAV channel angle of arrival estimation method based on on-site pre-training is adopted. By training the neural network on-site in each application scenario, channel state information is used to extract channel multipaths in real time, and a backpropagation neural network is constructed for angle estimation.
Benefits of technology
It improves the real-time performance and robustness of angle estimation, reduces computation time, and is suitable for low-latency drone sensing applications.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN117674917B_ABST
Abstract
本发明公开了一种基于现场预训练的无人机信道到达角估计方法，包括：设置信道测量设备频率和带宽，确定信道测量设备地面端位置和无人机飞行轨迹；利用信道测量设备获得实测飞行过程的信道冲激响应数据并提取多径，并计算每根天线阵元对应的多径相位；构建反向传播神经网络，对反向传播神经网络进行现场训练；实时获得实际应用阶段各个天线阵元各条传播路径的相位，输入到训练完成的反向传播神经网络中，实时估计信道多径角度。本发明针对特定的硬件设备和应用场景中对神经网络进行训练，相较于线下一次性训练的网络，提高了角度估计鲁棒性和实时性。
Need to check novelty before this filing date? Find Prior Art