Motion Estimation and Control for a Human-Following Robot Under Leg Detection Failure Using a Single-Line Laser Radar

In this article, motion estimation and control are investigated in human-following scenarios for a mobile robot with single-line laser radar (SLR) under conditions of leg detection failure. Leg detection is performed through support vector data description (SVDD), which obtains the raw position of a human using density-based spatial clustering of applications with noise (DBSCAN). The leg detection failure in the raw position from the SLR is compensated by an improved Kalman filter with an event-triggered mechanism to obtain the estimation motion of the human. The estimation motion is then used to generate a reference trajectory through a coordinate mapping method for controlling the mobile robot. Experimental results demonstrate the effectiveness of the proposed motion estimation and control method for a human-following robot under leg detection failure.