A dataset contains observations $X_i$ that are contaminated with outliers. Which of the following estimators would be most robust to these outliers?