Information Processing Systems Basic Architecture and HW SW Trade-Offs
The governing of mechanical systems is usually performed through actuators for the changing of positions,
speeds, flows, forces, torques, and voltages. The directly measurable output quantities are frequently
positions, speeds, accelerations, forces, and currents.
Multilevel Control Architecture
The information processing of
direct measurable input and output signals
can be organized in several
levels, as compared in Figure 2.5.
level 1: low level control (feedforward, feedback for damping, stabilization, linearization)
level 2: high level control (advanced feedback control strategies)
level 3: supervision, including fault diagnosis
level 4: optimization, coordination (of processes)
level 5: general process management
Recent approaches to mechatronic systems use signal processing in the lower levels, such as damping,
control of motions, or simple supervision. Digital information processing, however, allows for the
solution of many tasks, like adaptive control, learning control, supervision with fault diagnosis, decisions for maintenance or even redundancy actions, economic optimization, and coordination. The tasks of the
higher levels are sometimes summarized as “process management.”
Special Signal Processing
The described methods are partially applicable for
nonmeasurable quantities
that are reconstructed from
mathematical process models. In this way, it is possible to control damping ratios, material and heat
stress, and slip, or to supervise quantities like resistances, capacitances, temperatures within components,
or parameters of wear and contamination. This signal processing may require
special filters
to determine
amplitudes or frequencies of vibrations, to determine derivated or integrated quantities, or
state variable
observers
.
Model-Based and Adaptive Control Systems
The information processing is, at least in the lower levels, performed by simple algorithms or softwaremodules
under real-time conditions. These algorithms contain free adjustable parameters, which have
to be adapted to the static and dynamic behavior of the process. In contrast to manual tuning by trial
and error, the use of mathematical models allows precise and fast automatic adaptation.
The mathematical models can be obtained by identification and parameter estimation, which use the
measured and sampled input and output signals. These methods are not restricted to linear models, but
also allow for several classes of nonlinear systems. If the parameter estimation methods are combined
with appropriate control algorithm design methods, adaptive control systems result. They can be used
for permanent precise controller tuning or only for commissioning

Additional comments :