Perception-Driven Robotic Manipulation and Motion Planning for Packaging of Dynamic Compressible Objects

Venkataramanan, Rohith

Etd

Perception-Driven Robotic Manipulation and Motion Planning for Packaging of Dynamic Compressible Objects

Público Deposited

This thesis presents the development of an innovative robotic system for the efficient and accurate manipulation of envelope stacks moving on a conveyor belt in the mailing and packaging industries. The research aims to address the challenges in envelope stacking and packing by leveraging vision-based techniques for detecting stacks of envelopes and estimating conveyor velocity, as well as trajectory planning and control for robotic manipulation. The proposed solution integrates a custom-designed pneumatic gripper, a camera-based perception system, motion planning techniques, and velocity servoing to precisely pick and place multiple stacks of envelopes from a conveyor belt into a cardboard box. The camera system is designed to detect the envelope stacks and estimate the conveyor velocity by tracking envelope positions over time. The robotic manipulator employs velocity servoing to accurately probe the envelopes and slice into the stacks using the elastic and flexible properties of envelopes. A series of experiments and performance metrics are used to evaluate the system’s effectiveness in terms of grasping success rate, localization accuracy, conveyor velocity estimation error, stack detection performance, cycle times, and box-filling accuracy. Our proposed system demonstrates high success rates in grasping, localization accuracy, and envelope stack packing efficiency while maintaining the integrity of the envelope stacks. The research findings and proposed solution hold the potential to significantly improve automation processes in the mailing and packaging industries, paving the way for further advancements in robotic manipulation, motion planning, and perception-guided control.

Creator