Increasing performance requirements in motion systems necessitate takingthe flexible, dynamic behaviour of these systems explicitly into account. Newapproaches are being investigated to exploit multiple additional actuators andsensors to actively compensate for system deformations. The increasingrequirements justify an increased model dimensionality, complexity andaccuracy, and associated control solutions. To enable this, we developnumerically reliable identification methods for complex systems with high-orderdynamics and a large number of inputs and outputs. By connecting identificationand robust control we can non-conservatively account for model uncertainty, andappropriately account for the system environment, such as by disturbanceidentification. New techniques based on identification and iterative learningcontrol are being developed to accommodate reference-induced errors. Finally,the development of design principles for the mechanical design of high-techsystems, such as ultra-thin stages and adaptive optics systems, focuses on theresearch question of how to design for stiffness with high reproducibility andmanufacturability, and sometimes for low cost and low thermal sensitivity.Applications are in the high-tech systems industry (ASML, Philips, Océ, FEI,various SMEs).