Hybrid robust disturbance rejection position-force control to adjust the motion of mobile manipulator in automatic aluminum foundry recycling industries

This study presents the development of an event-driven hybrid control for position and force tracking applied on a mobile robotic manipulator for metal recycling tasks. The suggested controller operates in a sequenced strategy starting from a fixed spot, moving the mobile device towards a targeted zone ($\Omega _i<^>r$omega ir) from where the i-th piece-to-be-recycled is attainable (considering the arm manipulation). Once the event of entering the zone is completed, the mobile robot is fixed at a position, and the end-effector of the robotic arm is enforced towards the piece-to-be-recycled. When the end-effector touches the piece in a given spot ($\Omega _i<^>e$omega ie), the hybrid control changes to the force tracking intending to carry the piece towards the spot ($\Omega _i<^>p$omega ip) where it ill be processed. Each piece location is identified based on a vision-based system that applies deep learning tools using convolutional neural networks. A multi-physics numerical simulation illustrated the application of the developed controller in a realistic scenario, showing all the elements of the event-driven operation. To validate the suggested controller, the comparison with a robust control that works on a wide range of carrying mass confirms the operational improvement of the event-driven hybrid position and force design.