Ultrasound can noninvasively monitor the mechanical and dynamical properties of the artery. Scattering and overlap of adjacent tissue boundary echoes with those from the artery wall affect the estimation accuracy of the artery properties and impede continuous and automatic monitoring. Decomposition of the ultrasound radiofrequency (RF) signals using matching pursuit with particle swarm optimization is proposed to isolate the echoes arising from the tissue boundaries of the carotid artery wall for subsequent estimation of the wall thickness and diameter changes during the cardiac cycle. The proposed method exhibited less variance in the estimation of artery wall thickness when compared to manual estimation by a clinical method. Artery wall motion tracking by the proposed method was more robust compared to tracking achieved through the conventional cross-correlation technique when applied to the ultrasound RF signals from a wearable ultrasound sensor that contain a high volume of scattering echoes.