Inside Track with Ken Karnofsky, Senior Strategist, Signal Processing, MathWorks

KK: Today’s RF engineers have very different needs, because today’s RF devices and radios integrate RF and digital technologies to a degree never seen before. Emerging approaches for 5G design include hybrid RF and digital signal processing (DSP) techniques that require even more highly integrated technology. As a result, RF engineers—as well as system and digital engineers, for that matter—need to understand how the RF front end affects system performance and how to partition designs between RF/analog and digital components. 

The enhancements to our MATLAB and Simulink platforms integrate highly accurate RF and antenna modeling with advanced DSP algorithm design and implementation. This enables more effective collaboration among RF, digital, and system engineers to allow for faster development cycles and more thorough design verification.

CD: How have modeling and simulation software requirements changed in comparison to five or 10 years ago?

KK: The changes in RF technology are driving a need for several improvements in modeling and simulation software:

1. Improved integration of RF, antenna, and digital modeling and simulation.
2. Faster simulation of complex RF architectures to facilitate rapid design exploration.
3. Connectivity to a range of software-defined radio (SDR) and RF test hardware to accelerate and lower the cost of prototyping and design verification.

CD: How can design teams that consist of engineers who specialize in various disciplines work together to create more efficient solutions?

KK: Design teams can use software that enables them to model and simulate digital and RF components in the same environment, at multiple levels of fidelity. This enables system engineers to quickly build a reference design, and for each design team to elaborate the design with high-fidelity behavioral models that incorporate DSP, RF, antenna, mixed-signal, and control models. System-level simulation using these models provides insight into component interactions, exposes integration issues before building hardware, and enables more rigorous system verification much earlier in the development process.

CD: What are some of the challenges associated with modeling the latest wireless systems?

KK: Developing wireless systems today is a task that requires multiple design skills, including system architecture, DSP, RF, antenna, mixed-signal, digital hardware, and embedded software. Most teams don’t have expertise in all those areas. Even when they do, each specialist typically uses their favored tool. This makes system integration increasingly difficult, and pushes discovery of critical problems to the end of the development process when they’re most expensive to fix.

This challenge has different impacts at different stages of development. For example, researchers can’t effectively explore 5G hybrid beamforming techniques when they use different tools for digital and RF design. Advanced technology teams can’t prove their concepts in hardware prototypes when they have to rely on other teams for RTL implementation. And design teams are spending far too much time debugging highly integrated radio designs in the hardware lab or in the field.

CD: What are the benefits of a complete (i.e., algorithm-to-antenna) simulation?

KK: System-level (algorithm-to-antenna) modeling and simulation pay large benefits in several ways. First, simulation can eliminate many system-level and integration errors before building hardware. This is the first step in model-based design where system models automatically generate code for hardware and software implementation of algorithms, enabling algorithm designers to prototype on hardware without having to find programmers or HDL engineers from other teams. The models also provide a reusable test bench throughout the development process, saving time and ensuring consistency of testing. These combined capabilities enable faster design iterations and streamline verification. An upfront investment in modeling has been proven to reduce overall development time by 30% or more.

CD: How have aerospace and defense applications changed in recent years, and how has that affected simulation requirements?

KK: Current applications are driving the adoption of highly integrated RF front ends and adaptive radio technologies. These include development of robust tactical radios and networks, interference and spectrum management, electronic countermeasures, satellite and space systems, and signal intelligence.  

Defense electronics and military system designers use commercial wireless standards in a variety of ways. Researchers want to anticipate the impact of emerging technologies such as mm-wave and massive multiple-input, multiple-output (MIMO) systems. System designers are looking to adapt those standards to lower cost and improve reliability of military communications systems. Other engineers are concerned with minimizing interference from other operational systems in a shared spectrum environment. And the signal intelligence community needs to understand how to extract information from systems using these standards. 

These tasks are complicated by the scope and rapid change of commercial wireless standards. Defense system designers are users of these standards; they can’t afford to maintain comprehensive knowledge or in-house tools to keep up with them. 

Low-cost, highly capable SDR technology is driving innovation and broader adoption. COTS SDR hardware can be connected to a PC to create highly capable testing and prototyping systems. The challenge is that those first-generation SDR tools limited even broader adoption of the technology by forcing engineers to maintain low-level programming environments, or use software tools that work only with a single vendor’s hardware. 

CD: What additional challenges do you think 5G technology will create when it comes to modeling and simulation?

KK: The technologies being developed for 5G such as massive MIMO, mm-wave, and new modulation schemes require innovative combinations of new baseband technologies and RF architectures. These technologies only deepen and accelerate the need for highly integrated design environments and flexible connectivity to prototyping and test hardware. 

CD: In the future, how will the Internet of Things (IoT) impact design software needs?

KK: Today, most IoT designers purchase RF modules to add wireless connectivity to their products. If something goes wrong, they face a lot of time in the lab debugging a design they didn’t create. Designers can instead use model-based design to simulate the integration of RF front ends into their designs, which helps them identify and fix issues earlier and at lower cost.