Robust multiperson detection and tracking for mobile service and social robots

Area of Science:

• Robotics
• Computer Vision
• Artificial Intelligence

Background:

• Mobile robots in public spaces require robust multiperson detection and tracking.
• Existing systems face challenges in crowded and complex environments.

Purpose of the Study:

• To propose an efficient system for robust multiperson detection and tracking for mobile service and social robots.
• To enhance the performance of robots operating in public environments.

Main Methods:

• A novel maximum likelihood (ML)-based algorithm integrating multimodel detections within mean-shift tracking.
• An expectation-maximization (EM)-like algorithm with E-step for detection association and M-step for position localization.
• An improved sequential strategy for mean-shift tracking, prioritizing objects for robustness in crowded scenarios.

Main Results:

• Successful implementation on real-world service and social robots.
• Integration of stereo-based and HOG-based human detections, occlusion reasoning, and sequential mean-shift tracking.
• Significant improvements in multiperson tracking performance demonstrated through quantitative evaluations and real-world examples.

Conclusions:

• The proposed system offers a robust and efficient solution for multiperson detection and tracking from mobile robots.
• The ML-based approach and sequential tracking strategy enhance performance in complex public environments.
• The system demonstrates practical applicability and significant advantages for service and social robots.