Locomotion with Pedestrian Aware from Perception Sensor by Pavement Sweeping Reconfigurable Robot

Area of Science:

• Robotics
• Environmental Engineering
• Computer Vision

Background:

• Public pavement sanitation is crucial for social activities.
• Autonomous cleaning robots face challenges adapting to dynamic environments like varying pavement widths and pedestrian traffic.
• Existing robots lack the adaptability required for complex, real-world urban cleaning scenarios.

Purpose of the Study:

• To develop and demonstrate a self-reconfigurable pavement sweeping robot capable of adapting to environmental dynamics.
• To enable autonomous navigation and cleaning in complex urban settings.
• To address the limitations of current autonomous cleaning systems in handling variable pavement widths and pedestrian avoidance.

Main Methods:

• A novel self-reconfigurable pavement sweeping robot, Panthera, was designed with a width-reconfiguration mechanism using a scissor structure and lead screw motor.
• A control scheme integrating locomotion and reconfiguration was developed, utilizing feedback from a Red Green Blue-D (RGB-D) camera.
• Robot kinematic parameters were published for real-time reconfiguration during autonomous locomotion.

Main Results:

• Panthera successfully demonstrated autonomous width reconfiguration during locomotion in outdoor pavement experiments.
• The robot effectively adapted to varying pavement widths while navigating.
• The system showed capability in avoiding moving pedestrians in real-world scenarios.

Conclusions:

• The proposed self-reconfigurable robot, Panthera, offers a viable solution for autonomous pavement cleaning in dynamic urban environments.
• The integration of width reconfiguration with perception-based control enhances robot adaptability and operational efficiency.
• This technology paves the way for more effective and versatile autonomous cleaning solutions for public spaces.