Algorithm 1 Similarity-Based User Clustering Algorithm

1:Initialization: the number of UAV N; the similarity matrix $\mathbf{Sim}$ defined as in (9). 2:Build the adjacency matrix $\mathbf{W}=\mathbf{Sim}$, and calculate the degree matrix $\mathbf{D}$ with diagonal element $d_i={\sum }_{j=1}^U{sim}_{i,j}$. 3:Calculate Laplace matrix $\mathbf{L}=\mathbf{D}-\mathbf{W}$. 4:Calculate normalized Laplace graph matrix ${\mathbf{L}}_{norm}=D^{-1/2}L{D}^{-1/2}$. 5:Pick a number of N eigenvalues of ${\lambda }_1\le {\lambda }_2\le \dots \le {\lambda }_N$ of ${\mathbf{L}}_{norm}$. 6:Calculate the N smallest eigenvectors $z_1\le z_2\le \dots \le z_N$. 7:Let the $\mathbf{Z}$ matrix have the eigenvectors $z_1,z_2,\dots ,z_N$ as columns. 8:Use K-means clustering to cluster the rows of the matrix $\mathbf{Z}$. 9:Output: the division result of cluster $\mathbf{C}\left(c_1,c_2,\dots c_N\right)$.