Tackling System Design Challenges Through Early Verification

For design teams in a number of industries, the cost of system verification is now their top challenge. With Model-Based Design, system-level verification can be performed earlier in the design process through modeling and simulation, shortening design cycles.

Research and design engineers have successfully used the tools and approaches of Model-Based Design for many years. However, engineering organizations—both those with no experience applying Model-Based Design and those that are long-practiced—have significant opportunities to leverage the work they already do to help address the challenges of system verification.

Three Typical Situations

When looking at the use of Model-Based Design among our customers, we see three types of situations.

One is the set of companies that already use Model-Based Design as part of their standard process. These organizations have built libraries of design models at the component and system levels. They simulate these models to characterize system behavior and analyze their designs, automatically generate code to embed in deployed systems, and reuse models for hardware-in-the-loop and other testing approaches. The remaining challenges these companies face typically rest with managing and analyzing their requirements and designs to confirm—as early as possible and continuously throughout the development process—that they are designing the right thing and meeting requirements, especially as those requirements change based on internal factors and market conditions.

The second situation involves groups that, until now, haven’t perceived sufficient need to leverage modeling in their development process. Perhaps they were designing systems that were simple enough that teams could get by with informal communication, developing systems "on the fly” or through prototypes. But today, the systems they build are more complex, have more subsystems, and are designed to perform multiple, more sophisticated functions. For instance, a robot may integrate motion, vision, navigation, and other functions, often with many operating modes. Quality is increasingly important, either inherent in the system’s use or as a competitive differentiator, as are the traditional goals of faster time-to-market and reduced development costs. As a result, these teams—no matter how large or small they are—need approaches to understand the more complex system behaviors and evaluate different design approaches.

A third situation consists of organizations that have deep experience with modeling, but mostly for implementing components. The models (often expressed as VHDL, Verilog, or SPICE models) are relatively "low level” models of specific digital and analog electronics components, used for chip-level or system-on-chip design. These organizations see the need to work at a higher level given the increasing algorithmic content; the need to explore design tradeoffs more easily across digital, analog, and embedded software domains; and the increasing importance of understanding how components interoperate with other components in the context of the larger system. The high cost of developing electronic components and the risk of expensive market failure make it even more important to get early confirmation that the right things are being developed.

A Shared Approach: Early Verification With Model-Based Design

A common need across these situations is a verification process that works from a system-level approach early in the development process, evaluating design choices and identifying system integration issues as early as possible. Model-Based Design makes this possible, but different engineering teams experience the benefits in different ways.

Scenario 1: Organization is experienced in modeling, simulation, and automatic code generation, with an inventory of design models.

Engineers who are already applying Model-Based Design gain significant additional benefits through the incremental step of systematic, "up front" integration of their models into requirements analysis and management processes.

Scenario 2: Design team is involved in projects that have outgrown prototype-driven development processes.

Organizations are adopting Model-Based Design for the same reasons that many aerospace and automotive companies did years ago: to understand, design, and verify the system while avoiding the time and cost of building physical prototypes. Because these design teams are often relatively small and lack a modeling heritage, the enhancements in modeling tools in recent years are important for building models and verifying system performance easily in one environment that represents the multiple domains of the system.

Scenario 3: Organization is designing electronics with algorithmic content.

Moving to early verification means verifying that algorithms work with real-world constraints and confirming that important subsystems work together even before implementations are available. This requires modeling at a higher system level, and incorporating existing implementation models when appropriate. This workflow enables engineers to evaluate floating-point versus fixed-point algorithm performance, explore the impact of different HDL implementations, and confirm the design choice in the context of the larger system.

Modern Tools Tackle The Task

The ability to model multiple domains in one environment, reuse research algorithms in the design phase, simulate designs on the desktop and in real time, and automatically generate implementations for testing and production make it possible to virtually test system-level performance much earlier than in the past.

By enabling early verification, Model-Based Design tools help teams find defects, validate requirements, and confirm that design strategies are on track when there is still time to address any problems that are discovered. Design defects and other issues discovered late in the development process are costly to fix and often result in missed deadlines. When found early, however, these same problems can often be resolved with minimal impact on the schedule and the budget.

Engineering groups across a wide range of industries, disciplines, and design approaches are gaining a competitive advantage through early verification, because it leads to improved quality, lower costs, and shorter time-to-market.