Body Posture Determination for Heart Failure Patients From Ankle Orientation Measurements

Abstract

Heart-Failure (HF) is among the leading hospitalization causes in modern healthcare systems. In this paper, a method for performing continuous patient monitoring with a focus on low power consumption is presented. A prototype wearable device to assist clinical practice and improve patient’s outcomes by collecting biological impedance measurements is being developed at the University of Sevilla. Among the sensing components there are two major blocks formed by a commercial biological impedance analog frontend from Analog Devices (AD5940) and an Inertial Motion Unit (IMU) capable of estimating attitude of the device. This information could provide potential new insight into the patient’s status for the physician and help diagnose and remote monitor patients with HF. A major factor that can be analyzed to provide information on patient status is activity level and body states; time spent walking or standing, laying down or seated. In this work, a body tracking / activity estimation method is proposed for low power continuous monitoring. One key factor that can provide information on a patient’s status is their activity level and body position, such as time spent walking, standing, sitting, or lying down. This paper proposes a body tracking/activity estimation method for low-power continuous monitoring. The study reports a 97.6% prediction accuracy for lying down and standing/walking/seated positions and a 67.2% accuracy for discriminating between sitting and walking/standing positions. These results are highly relevant for clinical practice, as body motion and position can serve as important health markers for patients. Moreover, the motion information obtained can be further processed to better understand variations in the analog impedance measurements.

Keywords

• Body tracking
• position estimation
• patient activity
• low power monitoring