This work introduces an autonomous, robot-agnostic method for autonomous 3D part measurement. It addresses manual path planning issues, efficiently locating parts, calculating optimal view poses, solving the travelling salesman problem and performing collision-aware path planning. Experimental results confirm its reliability, making it a cost-effective solution when combined with our recently developed in-line robot calibration method which does not need an external tracking system.

This work aimed to validate the secondary kinematic system approach for robotic milling. A one-dimensional prototype was developed with an advanced feedforward controller to linearize the system, for successive modeling of the robot as a number of mutually interconnected masses. The system was identified using a custom Basin-hopping algorithm. A remarkable 200-fold increase in dynamic stiffness outside the 10-20Hz range and precise movements with accelerations up to 2G's and <0.1mm error could be achieved.

This work addresses the challenge of achieving precision and productivity in robotic milling with metals. A 3D high-performance secondary kinematic system, controlled by Beckhoff PLC with 8khz, enables the movement of a 22kW milling spindle with high accelerations of up to 4G's and high overall stiffness achieved by control. Manufacturing and assembly, as well as as first implementation of control have been completed, setting the stage for milling trials and exciting research in the coming years.

This work introduces autonomous robot arm-based workspace exploration, essential for scene reconstruction and path planning. The robot uses a hand-mounted lidar to effectively explore and reconstruct its own workspace, while avoiding collisions. The process involves engineered robot motion, clustering of unknown space, and Next-Best-View problem solving. Experiments on benchmark scenes and a hardware demonstration in an industrial setting confirm its effectiveness.

The cost of industrial foams is directly tied to density control during manufacturing, impacting production expenses. Our solution involves the development of an inline density measurement system coupled with model predictive control loops. This innovation minimizes density while ensuring critical quality characteristics, a collaborative effort with process experts and data scientists from our industry partners.

Traditional robot calibration relies on costly external tracking systems. In this study, an affordable 3D measurement device is attached to the robot's tool-flange to measure arbitrarily placed spheres, enabling precise robot model parameter identification via optimization. Numerical simulations, even with substantial measurement noise, validate the method's viability.