STP4: spatio temporal path planning based on pedestrian trajectory prediction in dense crowds

Area of Science:

• Robotics
• Artificial Intelligence
• Computer Science

Background:

• Autonomous mobile robot navigation in crowded environments presents significant challenges due to unpredictable human movement.
• Existing methods often struggle with real-time adaptation and efficient path planning in dynamic settings.

Purpose of the Study:

• To develop and evaluate a novel autonomous mobile robot navigation system for dense crowds.
• To enhance robot navigation safety and efficiency by predicting pedestrian trajectories.

Main Methods:

• The proposed method integrates pedestrian trajectory prediction with spatiotemporal path planning.
• Predicted pedestrian paths are converted into a time series of cost maps for robot guidance.
• The path planner operates without requiring long-term trajectory predictions.

Main Results:

• Real-world robot testing in a science museum confirmed the effectiveness of the trajectory prediction module.
• Simulations demonstrated that the proposed planning method achieved 26.4% faster arrival times compared to conventional 2D path planning in a 50-person crowd.
• The navigation strategy resulted in smooth robot movement without erratic dodging maneuvers.

Conclusions:

• The developed method offers a robust solution for autonomous mobile robot navigation in dynamic, crowded spaces.
• Predictive trajectory analysis and adaptive path planning significantly improve navigation efficiency and smoothness.
• This approach holds promise for enhancing robot performance in public and high-traffic areas.