A Novel System for Guiding Unmanned Vehicles Based on Human Gesture Recognition

摘要

Unmanned vehicle driving technology is developing rapidly in recent years. Since the relevant research of unmanned vehicles does not stop at the vehicle control method or multiple vehicle collaboration, vehicle-human interaction problems are gradually being valued by researchers. In this paper, a novel system is proposed for guiding unmanned vehicles based on human gesture recognition. The human gesture instructions are using as the bridge between people and an unmanned vehicle. The Dynamic Time Warping (DTW) method is improved to identify the instructions quickly and accurately. After the vehicle receives the leader’s gesture instruction, it can enter the follow (leader) mode. A continuous path can be formed according to the movement track of the leader. The pure pursuit method is used to control the unmanned vehicle so that the vehicle can follow the leader. Experiments are performed both indoors and outdoors. The results can prove the robustness and feasibility of the proposed system.