A Novel Dynamic Path Re-Planning Algorithm With Heading Constraints for Human Following Robots

Paths will be re-planed timely according the change of obstacles and target position for Human Following Robots (HFRs) so that they can run safely and rapidly. Traditional path planning algorithms are generally focused on how to avoid obstacles or connect with the new target position effectively, however there are rarely studies on the elimination of robot heading fluctuation which is caused by path re-planning. Concerning this issue, we introduce a novel dynamic path re-planning algorithm AT* which is based on A* and walkable area thinning. When robots heading difference value between the current and the last moment which is derived from A* exceeds the threshold, AT* will search for path connection points by walkable area thinning, and then a new path going through optimal connection point and meeting heading constraints will be planned. Thinning algorithm is improved in order to make the passable skeleton can connect path origin and the end. Results from simulation experiments show that AT* algorithm keeps heading stability as a priority at path re-planning and the robot motion stability can be improved effectively.