Non-Contact Human Gait Identification Through IR-UWB Edge-Based Monitoring Sensor

Non-contact sensors are negating the use of wearables or cameras and providing a rewarding and accepting environment to assist in biomedical applications, such as physiological examinations, physiotherapy, home assistance, rehabilitation success determination, compliance, and health diagnostics. In this paper, physiological parameter identification of human gait has been demonstrated through an edge-based sensor and a heuristic approach. Impulse radio ultra-wide band (IR-UWB) pulsed Doppler radar has been employed with a focus on understanding human walking patterns. This paper extracts an individual's gait trait from the associated biomechanical activity and differentiates lower limb movement patterns from other body areas via a radar transceiver. It is observed that Doppler shifts alone are not reliable to detect human gait because of frequency shifts taking place across the entire body (including breathing, heartbeat, and arm movements) where movement occurs. Thus, a heuristic spherical trigonometrical approach has been proposed to augment radar principles and short-term Fourier transformation (STFT) to identify the gait trait. The experiment presented includes data gathering from a number of male and female participants in both ideal and real environments. Subsequently, the proposed gait identification and parameter characterization has been analyzed, tested, and validated against popularly accepted smartphone applications where the variations are less than 5%.