This thesis will focus on developing an algorithm to enhance vision-guided robotics that is used to separate thin metal products from a stack. These stacks often adhere together due to corrosion-resistant oil and magnetic fields. This poses challenges for automated destacking of these products using robotic manipulators. A promising end-effector prototype for destacking thin products has been developed that uses a bending motion for product separation.
The goal of the thesis is to create an algorithm that utilizes 3D scanner data to optimize the positions and actions of the end-effector for effective separation. This algorithm consists of three main components: effector positioning, degree of bending, and the separation method. Key aspects include:
* Effector Positioning: Identifying critical positioning parameters and using point cloud data to determine optimal effector positions for the separation process.
* Degree of Bending: Calculating the appropriate bending force based on the product's material properties, stiffness, and geometry to ensure effective separation without detachment from the end-effector.
* Separation Method: Choosing between magnetic and vacuum effectors based on product characteristics and implementing a sequence of actions for efficient and reliable destacking.
The algorithm will be developed using Lua scripting language and will require extensive research into point cloud data processing, material mechanics, and the forces involved in the separation process.