Navigation of Autonomous Electric Wheelchairs in Crowded Environments Using Pedestrian Activity Patterns and Dominant Moving Directions

This paper proposes a pedestrian-aware navigation method for autonomous electric wheelchairs based on pedestrian flow data. By integrating the degree of activity and the most frequent movement direction-acquired via a 3D LiDAR-based system-into costmaps, the system enables proactive path and motion planning. Real-world experiments demonstrate that the proposed method reduces users' mental workload and enhances both safety and comfort for users and surrounding pedestrians, without compromising navigation performance.