KF distance observer and destructive collision avoidance for laser cutting head

Nowadays laser technology is largely employed in industrial application, in particular in metal sheet cutting. The core element is represented by the laser head which is movable relative to the workpiece and has to focus the laser beam at very precise positions. A stiff distance controller keeps the laser head at a stable and very close height from the workpiece (at most few millimeters) to guarantee a good cutting quality. Therefore, it is highly exposed to risk of collisions with production scraps and with cutting table grids. This paper proposes a sensor fusion algorithm based on a discrete-time Extended Kalman Filter (EKF), which provides a noise-free laser head distance measurement and a local estimation of the geometric sheet surface slope. These features allow a quick identification both of collisions with protruded parts and laser head exits from the workpiece's edges. The presented solution aims to use only typical measurements and informations available in a standard CNC system of a laser cutting machine, without any additional sensors.