Formal, model-based approaches to advance the system sciences

ISR’s long-standing purpose has been to provide a home to interdisciplinary research and education programs in systems engineering and the system sciences, and to develop basic solution methodologies and tools for systems problems in a variety of application domains. These dual missions are tightly coupled in the sense that large-scale science requires systems engineering and, conversely, systems engineering and implementation of modern real-world systems cannot occur without good systems science.

Ephremides and studentsBroadly speaking, most of ISR’s contributions to the system sciences have occurred through a need to model, design, and understand (analyze and simulate) new types of engineering and biological systems that are economically competitive (providing bang for the buck), automated and distributed, readily extensible and adaptable to changing environmental conditions, and resilient to uncertainties.

Today, the frontiers of engineering and biological system development can be pushed because engineers have the tools to support analysis and design. ISR was at the forefront of creating these tools. Basic research has resulted in many new algorithms and sophisticated models for decision making and control (sense-decide-actuate lifecycle), communications, and computing needed to support these activities. New approaches to the planning and multi-objective optimization-based design of engineering systems have been developed.

Since ISR faculty are drawn from 14 departments across five colleges, advances in the system sciences have been driven by a wide range of complex applications, which, of course, have changed over time. A few examples include: integrated product and process design for manufacturing applications, helicopter control systems, air traffic control, healthcare systems, mobile ad-hoc communications networks, sensory systems based on the echolocation systems of bats, and fast and small microrobots.

The Systems Research Center that became ISR was created in part to develop formal model-based approaches to systems analysis and systems engineering. Most of ISR’s contributions to the first part of this mission have occurred through advances to the systems sciences, with support provided through traditional academic funding mechanisms.

Exploratory contributions to systems engineering

ISR’s work and contributions to systems engineering have been far more exploratory. It is important to note that when ISR was created, the systems engineering profession was quite immature—perhaps where control and finite elements were in the mid 1960s. The profession also was stuck in a document-centric mindset. The systems engineering community didn’t even have a professional society.

Advanced Manufacturing LabIn the early days of ISR, a sizeable disconnect existed between ISR’s vision for systems engineering and industry capability. At times, this made the task of forming cooperative working relationships with U.S. industry very difficult. This situation persisted until 1995-2000, when remarkable advances in computing finally created an opportunity for industry to catch up.

During the past decade, there has been an enormous effort within the systems engineering community to develop and implement model-based systems engineering procedures. Visual modeling languages such as UML and SysML are now supported by tools such as MagicDraw and Artisan. The representation and management of very large sets of requirements—millions of requirements—can be handled by tools such as DOORS (Dynamic Object Oriented Requirements System) and IBM Telelogic SLATE. These advances have been facilitated by the International Council of Systems Engineers (INCOSE), which now has a flourishing membership approaching 10,000. For those ISR faculty who have been actively involved in systems engineering research and education, this is nothing less than complete vindication that we were working on the right problems all along.

Today, the discipline of systems engineering can be naturally partitioned into two sub-processes: the systems management process and the technical development process. There are now approximately 20 graduate-level programs in systems engineering offered throughout the U.S. Perhaps for economic reasons, most of these programs emphasize the systems management side of things, where there is considerable overlap with project management.

ISR, in contrast, has focused on issues associated with the technical development of systems. Therefore, over the years, ISR has contributed substantially to system modeling, system requirements modeling and visualization, integration and optimization of product and process development, and recently, formal approaches to validation and verification through present-day and future approaches to model checking.

It is now abundantly clear that to keep the complexities of future system-level designs in check, increased emphasis on system decomposition, systems abstraction and formal approaches to analysis will be needed. This trend will continue into the near future, where we intend to become internationally recognized for contributions to model-based systems engineering methodologies and tools.

Consistent models, changing applications

Some of ISR’s most significant contributions have come in the areas of robust intelligent control, telecommunication wireless and hybrid networks, sensors, optimization, semiconductor manufacturing, signal processing in the auditory cortex, network security, geometric reasoning and planning, databases, tradeoff analysis, human-computer interaction, artificial intelligence, data mining, neuromorphic engineering, and micro and nano electromechanical systems. The abstract models are the same, but the applications change from field to field.

NSF sessionISR’s scientific breakthroughs have included cell-based sensors; micro-ball bearing and three-dimensional grayscale microfabrication technologies for micro-engines; multi-criteria, engineered learning systems; integrating product dynamics and process models; actuation and control based on signal processing; multi-part molding for robotics; mobile adhoc networks (MANETs); queuing for mass vaccination clinics; information visualization; real-time process control; speech processing in noisy environments; AI planning; design of hybrid communication networks (satellite/wireless/wireline); design of biologically inspired microelectronic devices; biological and chemical sensing system design and development for toxin detection; multi-body systems, flocking theory and design; new control approaches for nonlinear systems including stall control of jet engines; and medical applications including control targeting of therapies to tumors.


Top