A Hierarchical Progressive Perception System for Autonomous Luggage Trolley Collection

Zhirui Sun1,2, Zhe Zhang1, Jieting Zhao3, Hanjing Ye3, and Jiankun Wang1,2

 1Shenzhen Key Laboratory of Robotics Perception and Intelligence, Department of Electronic and Electrical Engineering, Southern University of Science and Technology, Shenzhen, China.

2Jiaxing Research Institute, Southern University of Science and Technology, Jiaxing, China.

3Shenzhen Key Laboratory of Robotics and Computer Vision, Department of Electronic and Electrical Engineering, Southern University of Science and Technology, Shenzhen, China.

Abstract: Advancements in intelligent vehicle technologies and autonomous driving are now used for tasks like collecting luggage trolleys at airports. The robotic autonomous luggage trolley collection system employs robots to gather and transport scattered luggage trolleys. However, existing methods for detecting and locating luggage trolleys often fail when they are not fully visible. To address this, we introduce the hierarchical progressive perception system, which enhances the detection and localization of luggage trolleys under partial occlusion. The proposed system integrates the hierarchical process structure and progressive perception strategy. This innovative structure processes the luggage trolley's position and orientation separately. It can accurately determine the luggage trolley's position with just one well-detected keypoint and estimate its orientation when partially occluded. Once the luggage trolley's initial pose is detected, the progressive perception strategy continuously refines this information until the robot begins grasping. The proposed system only needs RGB images for labeling and training, enhancing the scalability and practical implementation. The experiments on detection and localization demonstrate that the proposed system is more reliable under partial occlusion compared to existing methods. Its effectiveness and robustness have also been confirmed through practical tests in actual luggage trolley collection tasks.

HPPS Framework:

Dataset Examples:

Real-world Experiment Video (Scenarios with Different Visibility or Multiple Luggage Trolleys)

Real-world Experiment Video (Comprehensive Scenario):