Kalman Filter-Based Human Tracking Using 2D LiDAR

This paper proposes a filter-based approach to track human gait using a 2D LiDAR. A Kalman filter employing a constant acceleration model is utilized to estimate the human walking trajectory based on leg scan data. The center points of the detected legs are used as measurement updates to improve the tracking accuracy. However, during walking, self-occlusion between legs can occurs, leading to data loss. To address this, an algorithm is introduced to estimate the position of the occluded leg. These estimations are subsequently used as pseudomeasurements in the Kalman filter. An experiment is conducted to validate the effectiveness of the proposed method.