Development of a Gait Measurement System Based on a Human-Following Mobile Robot

In this paper, we present a gait measurement system and a human-following mobile robot system through recognition of a 2D human skeleton model using MediaPipe-Pose (MPP), an open source package based on image recognition. The gait measurement system reconstructs the 2D human joint pixel coordinates into a 3d humanoid model using an optimization algorithm and analyzes the joint angle trajectory and foot trajectory to measure step, walking distance, and walking speed. Next, the human-following mobile robot employs a speed control algorithm to maintain a certain angle and distance. This system can be easily operated and measured in conjunction with smartphone applications, such as adjusting the initial position of the mobile robot, estimating the link length, measuring the gait, and visualizing the gait measurement results. In order to verify the performance of the proposed system, the experiment was conducted at a constant step by varying the angle of the camera, and the performance of the gait measurement results for each angle of the camera was verified.