I am wondering about the evaluation metrics X and Z error, are that RMSE ?

[YieHo-Liu]

Dear authors:
In the paper, the X error and Z error is described as
"Formally, the lane-to-lane cost between Xj and Xk is calculated as the square root of the squared sum of point-wise distances over all y-positions,..."
But in the code , the error is calculated as
x_dist_mat_close[i, j] = np.sum(x_dist[:close_range_idx] * both_visible_indices[:close_range_idx]) / np.sum(both_visible_indices[:close_range_idx])

so the error should be Root Mean Squared , the square root of the mean squared sum , is it?