| High-performance motion stage systems are critically important for modern manufacturing processes at all scales, such as high-speed machining, micro machining and nano manufacturing. Although the performance requirements vary for different applications, generally motion systems are expected to have high resolution, fast response and excellent tracking abilities. This thesis research aims to achieve such high-performance motion systems in a systematic manner.; First, parallel kinematics structures are used to build high-performance machines for high-speed machining and nano positioning. Traditional machine-tool kinematic configurations (stacked-axis construction) prevent one from achieving high contouring accuracy over a wide range of speeds, due to the inherent dynamic mismatch (from imbalances in the inertial loads) between the axes as a result of such design. The new motion systems with the parallel kinematics structures overcome these inherent disadvantages and difficulties found in the traditional design, and can provide much better positioning and contouring capabilities.; Second, advanced modeling and control techniques are applied to the new motion stage system, specifically a parallel kinematics nano positioning XYZ stage. Due to the need for monolithic construction, the dynamics and kinematics of all axes get coupled together, making controller design non-trivial for this system. A controller is designed in the mode space instead of the Cartesian space, resulting in a very efficient low-order controller. By applying advanced control approaches to this system, a high-bandwidth and a high-resolution motion system is achieved with a relatively simple controller.; Finally, a provably-optimal feedrate-optimization (trajectory planning) algorithm is developed to exploit the capabilities of a given physical motion system. System capabilities and process requirements are modeled and formulated as constraints to drive the algorithm. The resulting trajectory makes full use of the system capabilities and completes the required task in the minimum time, while satisfying all performance requirements such as dimensional error, maximum cutting force, etc. A bi-directional scan structure is adopted for the algorithm and a sub-optimization structure for each optimized point. These features make the algorithm computationally efficient, extendable to any state-dependent constraints and robust with respect to singularity difficulties often encountered in the optimal control approach. By studying the behavior of the algorithm in the phase plane, global optimality of this feedrate-optimization algorithm is proved. |
