https://scikit-learn.org/stable/auto_examples/preprocessing/plot_all_scaling.html
https://towardsdatascience.com/scale-standardize-or-normalize-with-scikit-learn-6ccc7d176a02
This tells to remove outliers then there is no need for Robustscalar. This does not include transformers too.
Short answer: it depends!
It depends on the data at hand, the inclusion/exclusion of outliers of the feature, and the subsequent algorithm you want to utilize.
Scaling is preferably done for those algorithms which get impacted by the magnitude of the feature values.
The official documentation also suggests which methods are prone to get affected by outliers. So if outliers cannot be ignored, then, for example, RobustScalar would be appropriate over MinMaxScaler.
However, to really judge between RobustScalar and QuantileTransformer, you may need to subject the data to both the methods, experiment/ test the results of classification/regression, etc., compare the results/ errors/ accuracy and then identify which one helps to achieve an optimal prediction.