DIAT-µSAT: micro-Doppler Signature Dataset of Small Unmanned Aerial Vehicle (SUAV)

Due to the smaller size, low cost, and easy operational features, small unmanned aerial vehicles (SUAVs) have become more popular for various defense as well as civil applications. They can also give threat to national security if intentionally operated by any hostile actor(s). Since all the SUAV targets have a high degree of resemblances in their micro-Doppler (m-D) space, their accurate detection/classification can be highly guaranteed by the appropriate deep convolutional neural network (DCNN) architecture. In this work, an indigenously developed continuous wave (CW) (X-band: 10 GHz) radar is used to build a diversified “DIAT-µSAT” dataset comprising 4849 micro-Doppler signature images of SUAV targets: RC plane, three-short-blade rotor, three-long-blade rotor, quadcopter, bionic bird, and mini-helicopter + bionic bird. All the SUAV targets are operated at different speeds/orientations/rates as follows—revolution per minute (RPM): 200–1740 RPM, flapping rates: 2–4 flaps/s, azimuth angles: 0◦–360◦ at the angle resolution of 45◦, and elevation angles: 0◦–90◦ at the angle resolution of 30◦, tilted (with respect to radar’s boresight) target positions, so as to ensure the diversification in our dataset.

鉴于小型无人航空器（SUAVs）体积较小、成本较低且易于操作，其在各类国防及民用领域的应用日益广泛。然而，若任何敌对行为者故意操控，此类无人航空器亦可能对国家安全构成威胁。鉴于所有SUAV目标在微多普勒（m-D）空间中均具有高度相似性，通过合适的深度卷积神经网络（DCNN）架构，其精确检测与分类得以高度保障。在本研究中，我们采用自主研发的连续波（CW）雷达（X波段：10 GHz），构建了一个名为“DIAT-µSAT”的多元化数据集，该数据集包含4849幅SUAV目标的微多普勒特征图像，包括遥控飞机、三叶旋翼、三长叶旋翼、多旋翼无人机、仿生鸟和微型直升机+仿生鸟。所有SUAV目标在以下不同速度/方向/速率下进行操作：转速（RPM）：200–1740 RPM，拍动频率：2–4次/秒，方位角：0◦–360◦，角度分辨率为45◦，仰角：0◦–90◦，角度分辨率为30◦，倾斜（相对于雷达的瞄准线）的目标位置，以确保数据集的多样性。