Professional Documents
Culture Documents
J. Gausemeier et al. (eds.), Design Methodology for Intelligent Technical Systems, 183
Lecture Notes in Mechanical Engineering,
DOI: 10.1007/978-3-642-45435-6_5, c Springer-Verlag Berlin Heidelberg 2014
184 H. Anacker et al.
The principle solution forms the basis of the design and development. Engineers
of the involved domains derive their domain-specific models from the it. This is,
however, an error-prone and tedious task. Therefore, we will introduce a semi-
automatic model transformation techniques (cf. Sect. 5.1) that enables engineers
to, e.g. derive an initial controller hierarchy or an initial software architecture. Af-
terwards, each domain details these models. This may involve changes that have an
impact on the other domains. In order to keep the models of all domains consistent,
we will propose a model synchronization technique (cf. Sect. 5.1.3).
The system must consider several concurrent objectives in different Application
Scenarios.
This requires methods for optimizing the system with respect to these objectives
and appropriate adaption methods. System optimization methods origin from the
research areas of applied mathematics and artificial intelligence. The methods de-
termine the optimal system behavior or a set of optimal compromises for several
concurrent objectives. Practically, this is a formalism to compute optimal controller
parameters or optimal configurations of the system structure (cf. Sect. 5.3). Then,
it is the task of engineers from the domains mechanical, electrical/electronic, con-
trol, and software engineering to specify the corresponding change of the system’s
behavior, i.e. the reconfiguration of the system.
The system can perform reconfigurations on every system level (cf. Sect. 1.4.3).
In particular, this requires new design methods for the application software, the
system software, and the hardware modules to specify reconfiguration. Furthermore,
reconfiguration is often safety-critical and must fulfill hard real-time constraints.
Consider the RailCab’s reconfiguration behavior to build a convoy as an example (cf.
Sect. 2.1.7): The RailCab must reconfigure the controller behavior to consider the
distance to the preceding RailCab if the RailCab joins a convoy as a member. In
fact, if this function is not free from design faults or the system cannot execute the
reconfiguration within a certain time, a crash may happen. Therefore testing and
formal verification methods are crucial to ensure the safety of the system’s complex
behavior and its real-time properties.
On the level of the application software, software engineers specify the com-
munication behavior and the switching between alternative behavior implementa-
tions. We apply a component-based design method called M ECHATRONIC UML
that considers hard-real time constraints for the communication behavior, the re-
configuration of controllers, and the reconfiguration of software components. In
M ECHATRONIC UML, formal verification techniques are applied to ensure safety
constraints and the real-time properties of the system.
As a consequence of reconfigurations of the application software, the software’s
resource and performance demands changes. Usually, the system must reconfig-
ure hardware modules to meet the changed requirements of the application software