Radar for indoor monitoring is an emerging area of research and development, covering and supporting different health and wellbeing applications of smart homes, assisted living, and medical diagnosis. Different human motion articulations present themselves more vividly in certain joint-variables data domains, most notably, time-frequency (TF) and range vs slow time. In this paper, we present a human motion data-driven classifier that utilizes both domains through a feature fusion approach. With data in each domain considered as an image, the features are extracted from lower dimension projections. These projections recognize the correlations across each image dimension, and are pursued using the generalized principal component analysis (GPCA). It is shown, through the confusion matrices, that feature fusion provides improved classification performance of human daily activities over the case where only the features of either domain are considered.