Algorithm 1 Hierarchical learning approach.

1:Load RF dataset   2:Define the inputs and the outputs   3:Pre-process the data   4:Encoding the outputs with respect to the inputs   5:Split the data into $Sets$: training (70%) and testing (30%) sets $(trainSet,testSet)$   6:Define the input variables and find the optimal parameters $\left(params\right)$ using $GridSearch$   7:Records $startTime$   8:Train the system on $\left(trainSet\right)$ by calling $HierarchicalApproach$ procedure   9:Test the system by using $\left(testSet\right)$ and evaluate the required metrics  10:   11:procedureClassifier($Sets,params$)  12:  Use the optimal parameters $params$ for KNN and XGBoost  13:  Use ensemble learning based on KNN and XGBoost  14:  Train and fit the classifier using $\left(trainSet\right)$  15:  Predict the test samples  16:end procedure  17:   18:procedureHierarchicalApproach($trainSet,testSet,params$)  19:  $Sample$ from $Sets$ will pass the first $Classifier$ to specify the availability of the UAV (2 classes: 0-No UAV, 1-UAV)  20:  if $class==0$ then  21:    $procTime=timeNow-startTime$  22:    go to end procedure and return the predicted class ($predClass==C1$)  23:  else  24:    $Sample$ will pass the second $Classifier$ to specify the type of the UAV (3 classes: 0-Bebop UAV, 1-AR UAV, 2-Phantom3 UAV)  25:    if $class==2$ then  26:      $procTime=timeNow-startTime$  27:      go to end procedure and return the predicted class: Phantom3 UAV ($predClass==C10$)  28:    else  29:      if $class==0$ then  30:       $Sample$ will pass the third $Classifier$ to specify the mode of the Bebop UAV (4 classes: 0-ON ($C2$), 1-Hovering ($C3$), 2-Flying ($C4$), 3-Recording ($C5$))  31:       if $class==0$ then  32:         $procTime=timeNow-startTime$  33:         go to end procedure and return the predicted class: Bebop UAV with ON mode ($predClass==C2$)  34:       else if $class==1$ then  35:         $procTime=timeNow-startTime$  36:         go to end procedure and return the predicted class: Bebop UAV with Hovering mode ($predClass==C3$)  37:       else if $class==2$ then  38:         $procTime=timeNow-startTime$  39:         go to end procedure and return the predicted class: Bebop UAV with Flying mode ($predClass==C4$)  40:       else  41:         $procTime=timeNow-startTime$  42:         go to end procedure and return the predicted class: Bebop UAV with Recording mode ($predClass==C5$)  43:       end if  44:     else  45:       $Sample$ will pass the fourth $Classifier$ to specify the mode of the AR UAV (4 classes: 0-ON ($C6$), 1-Hovering ($C7$), 2-Flying ($C8$), 3-Recording ($C9$))  46:       if $class==0$ then  47:         $procTime=timeNow-startTime$  48:         go to end procedure and return the predicted class: AR UAV with ON mode ($predClass==C6$)  49:       else if $class==1$ then  50:         $procTime=timeNow-startTime$  51:         go to end procedure and return the predicted class: AR UAV with Hovering mode ($predClass==C7$)  52:       else if $class==2$ then  53:         $procTime=timeNow-startTime$  54:         go to end procedure and return the predicted class: AR UAV with Flying mode ($predClass==C8$)  55:       else  56:         $procTime=timeNow-startTime$  57:         go to end procedure and return the predicted class: AR UAV with Recording mode ($predClass==C9$)  58:       end if  59:     end if  60:    end if  61:  end if  62:  Return predicted class ($predClass$) and processing time ($procTime$)  63:end procedure