Systems Design: Overcoming The Cyber-Physical Challenge

Why understanding model-based and platform-based design is key to avoiding the development of automated systems that are too limited to be of long-term use in modern, complex manufacturing operations.

The three key components of a cyber-physical system are computation, communication, and control. Source: National Instruments
The three key components of a cyber-physical system are computation, communication, and control. Source: National Instruments

Even the simplest systems today are fairly complex. In the not-so-distant past, a simple system was largely comprised of physical elements behaving according to the laws of physics (for example, a gear). Today, systems—more often than not—include information technology and computation along with physics. It is this group of fields, combined to engineer systems that can interact with their environment that are referred to as cyber-physical systems.

These types of systems have become so commonplace (just think of smart instruments controlling basic operations such pressure, level and flow) that we don't always think of these systems as being “cyber-physical systems”. Understanding them at this essential level, however, is key to ensuring the systems we design—or even just implement—are well thought out enough to be of real use.

In its recently released Trend Watch 2014, National Instruments considers several key technology trends, including models of computation, RF/wireless, big analog data and, of course, cyber-physical systems (CPS). In the Trend Watch paper, the importance of understanding the cyber-physical nature of today’s systems is deemed critical as “domain experts focus less on new design and more on the integration of cyber and physical components, often with little insight on how a component behaves when integrated with others. [As a result], systems become brittle and intractable.”

To overcome this problem, National Instruments states that the central issue in solving the CPS design challenge is to design beyond basic implementation goals and instead focus on consideration of system level design and functions.

The paper notes two proven methods which engineers can use to address the CPS design challenge: model-based design and platform-based design.

Model-based design “emphasizes modeling to design, analyze, verify, and validate dynamic systems. Engineers derive models from system specifications and the analysis of the environment and use them to design, simulate, synthesize, and test a CPS. These modeling techniques illuminate the interplay of practical design with formal models of systems that incorporate both physical dynamics and computation.”

Platform-based design is often associated with the automotive and aerospace industries where it is used to design scalable platforms encompassing large complex systems with long life spans. The Trend Watch paper states “you can use a platform as an abstraction layer to think about application-level constraints without concerning yourself with implementation refinements. With the right levels of abstraction, you may separate design concerns by defining platform elements with clear interconnections, which results in highly componentized, composable, and modularized designs. Clear interconnections make it possible for you to replace or upgrade platform elements with commercial off-the-shelf hardware to decrease development costs and simplify life-cycle management. You can reuse, repurpose, retool, or leverage platform elements for test frameworks, requirements tracking, verification, and documentation.”

Though platform-based and model-based design are very different approaches to CPS design, they are complementary methodologies that can be and are often used in parallel.

More information on platform-based design as it relates to embedded systems.

More information on model-based design as it applies to control engineering.

Companies in this article
More in Software