GaitTracker: 3D Skeletal Tracking for Gait Analysis Based on Inertial Measurement Units

Abstract

Gait rehabilitation is a common method of postoperative recovery after the user sustains an injury or disability. However, traditional gait rehabilitations are usually performed under the supervision of rehabilitation specialists, which implies that the patients cannot receive adequate gait assessment anytime and anywhere. In this article, we propose GaitTracker, a novel system to remotely and continuously perform gait monitoring and analysis by three-dimensional (3D) skeletal tracking in a wearable approach. Specifically, this system consists of four Inertial Measurement Units (IMU), which are attached on the shanks and thighs of the human body. According to the measurements from these IMUs, we can obtain the motion signals of lower limbs during gait rehabilitation. By adaptively synchronizing coordinate systems of different IMUs and building the geometric model of lower limbs, the exact gait movements can be reconstructed, and gait parameters can be extracted without any prior knowledge. GaitTracker offers three key features: (1) a unified 3D skeletal model to depict the precise gait movement and parameters in 3D space, (2) a coordinate system synchronization scheme to perform space synchronization over all the IMU sensors, and (3) an automatic estimation method for the user-specific geometric parameters. In this way, GaitTracker is able to accurately perform 3D skeletal tracking of lower limbs for gait analysis, such as evaluating the gait symmetry and the gait parameters including the swing/stance time. We implemented GaitTracker and evaluated its performance in real applications. The experimental results show that, the average error for skeleton angle estimation, joint displacement estimation, and gait parameter estimation are 3∘, 2.3%, and 3%, respectively, outperforming the state of the art.