A path prediction method for human-accompanying mobile robot based on neural network

Abstract

This paper presents a path prediction method for human-accompanying mobile robot such as robotic wheelchair, domestic robot and tour guide robot. An accompanying human is detected using an in-vehicle laser range sensor (LRS). A new filter gets a smoothed track from raw LRS data of human footprints. Back propagation neural network predicts future positions of the accompanying human from the human track. Based on the future positions, a cubic spline generates a future path of the accompanying human. The experimental result validates the feasibility of the path prediction method. © 2012 Springer-Verlag.