An Online Interactive Approach for Crowd Navigation of Quadrupedal Robots

Robot navigation with mobile agents in constrained environments requires the robot to capture crowd interactions for safe and cooperative movements. Challenges arise from the prediction and planning, behavior modeling, and choice of evaluation method.

We present an approach for interactive and human-friendly crowd navigation in complex static environments. The planner models the online interactions among the robot, humans, and the static environment based on game theory, and provides human-friendly navigation commands.

We use various indicators to evaluate the social awareness of the planners and show that our method outperforms existing approaches in success rate to reach the goals and compatibility with humans while maintaining low navigation times.

Our approach stably provides interactive navigation solutions with longer social distances, smoother velocity transitions, and higher human compatibility, while also achieving a high navigation efficiency as it better captures the robot’s geometric shape.

The robot (red) moves through the crowd from top to the goal at the bottom. Both versions of our GTICN approach provide smooth paths in rational travel times without significantly disturbing human behaviors.

Our approaches capture the crowd interaction with the static environment for efficient and smooth navigation. The travel time is further reduced by optimizing the robot’s orientations in our full version.

The planner is successfully deployed on a real-world quadrupedal robot, demonstrating safe and interactive crowd navigation with real-time performance. It can handle different crowd motion patterns in the "random-walk" and the "box-relay" scenarios.

The planner shows cooperative behaviors when a human overtakes or meets the robot in the "corridor" environment.