This paper reports performance results of qualitative optical flow processing integrated in a fuzzy logic-based behavioral mobile robot control system. The quality and noise robustness of qualitative optical flow processing is improved by a novel optical flow outlier detection filter. This filter performs a qualitative detection of outliers based on comparison between an optical flow vector and the average value of an optical flow pattern related to the considered vector. The performance of qualitative optical flow processing is tested and demonstrated successfully through experiments in a real office environment.