Defining a New Methodology for Radar System Design
By Daren McClearnon, Product Marketing Manager, System-level Design, Agilent Technologies, EEsof EDA

For radar system architects, the challenges are many. One such challenge is that modern radar systems may work in environments with strong clutter, noise and jamming and therefore must be designed using advanced digital signal processing techniques. To design these complex systems, most designers use analysis suites that they have personally collected and maintained over many years, using general purpose tools. Often times though, these tools do not mesh well with enterprise design environments, modeling and language-based intellectual property (IP) lifecycle tools, and test and measurement platforms. Moreover, they lack sufficient fidelity in the modeling of the RF, clutter and propagation effects to perform aggressive, predictive RF-baseband co-design.

Using this older methodology results in reduced collaboration between teams, greater dependence on expensive verification assets (e.g., ranges, chambers and “live” testing), and overdesign with wider design margins. It also reduces the real-world feedback to the system architect, which is critical for influencing subsequent generations of the radar design. What’s required is a system design methodology that is based on an open modeling environment which connects to a variety of design flows and can easily bridge the gap between design and verification. Bridging this gap brings “range test” fidelity into the architecture design phase, enabling the design of superior system architectures and facilitating early R&D hardware verification.

A Radar Architecture “Cockpit”
An example of just such a radar system design methodology is the SystemVue 2010.07 electronic system-level (ESL) software with W1905 radar model library. The W1905 radar library provides 2 dozen highly-parameterized primitive blocks, as well as higher-level reference designs that can be used to create working radar systems, including both the radar signal processing and its environment. The library is especially useful for pulsed-doppler (PD) radar architectures, and can be used by radar architects in the aerospace/defense community to study performance under various conditions, or to generate precise wideband signals for algorithm and hardware verification. In addition to military applications, the library can also be used by researchers in the academic, consulting, automotive, medical, and commercial wireless fields.

A key benefit of the radar simulation block set is that it blurs the boundaries between design and verification. SystemVue’s value is that as an open modeling environment that connects to a variety of design flows, it pulls together RF modeling, measurement science, reference IP, and simulation technology. This helps bridge the gaps in system design and enables superior physical layer systems design and verification. A targeted personality for radar designers completes this front-end radar architecture “cockpit.”

The SystemVue environment sits above both the RF and baseband DSP hardware design flows, bringing functional verification directly into the design environment itself. As a result, hardware design is performed with higher confidence and with reduced need for exhaustive R&D verification farther downstream.

The Methodology in Action
In this radar system design methodology, the W1905 radar model library block set and example workspaces serve as algorithmic and architectural reference designs to verify radar performance under a variety of signal conditions, target and radar cross section (RCS) scenarios, clutter conditions, jammers and environmental interferers, and different receiver algorithms. An example of this is shown in Figure 1. Here, echoes from a wideband measured signal are incorporated as a stimulus in an algorithmic simulation study and demodulated using the environment previously described. Conversely, the same environment can generate degraded reference waveforms for either baseband or modulated RF hardware testing.

This radar system design methodology makes it possible for the radar architect to fill in gaps in the system design where proprietary models and data are not available. This stems from using an open environment that essentially puts many tools at the designer’s disposal. In this case, the openness of the SystemVue 2010.07 platform for multilingual, polymorphic modeling in a variety of IP formats, including MATLAB®, C++, graphical user interface blocks and VHDL/Verilog, allows the radar designer to easily incorporate existing radar IP and regression scripting, while taking advantage of SystemVue to connect to RF simulators (including nonlinear X-parameter component performance), measured waveforms and existing designs.

A Platform for Integrated R&D Verification
Figure 2 hints at additional creative configurations that are possible using the proposed radar system design methodology. The parameterized radar library can be used to generate baseband and RF test vectors at any location in the block diagram, facilitating model-based verification at several levels of abstraction as a design matures. Math and C++ algorithms can be tested using simulation only, and compared using regression scripts; VHDL hardware blocks can be co-simulated using Mentor ModelSim and compared to the original algorithmic reference; or the exact same simulated waveforms can be downloaded to baseband pattern generators (or wideband arbs in modern RF signal generators) to test live hardware (Figure 2).

The ability to create precisely controlled amounts of impairment and signal degradation is critical for stressing receiver algorithms and RF transceiver hardware, and works both ways. Early in a design, the measurement environment helps add realism to the scenarios; later in the design, the simulation environment helps automate the coverage testing of the hardware.

Benefits of Design/Test Integration
The benefit of integrated design and test is twofold. First, there is direct continuity between the original architecture intent and verified hardware. This leverages IP, reduces engineering “glue” and test overhead, and also feeds back important performance metrics to subsequent design iterations of the architecture. In other words, the design process “learns” and produces better architectures.

Second, the ability to leverage the simulation environment to make creative, customized measurements and proprietary personalities using standard commercial off-the-shelf (COTS) test equipment can provide large savings for reconfigurable, early R&D testing. It saves time, increases asset reuse, and allows a geographically dispersed team to collaborate using standard, sharable tools. More expensive range testing, indoor and outdoor chambers, and hardware simulators can then be saved for a smaller suite of final design validation, while early R&D validation can be more timely and insightful.

Conclusion
While designing modern radar systems can be challenging, using a system design methodology specifically geared toward such systems can alleviate many issues. Not only does this help bridge the gap between design and verification, but it also provides today’s radar system architects with the tools they need to design and verify superior system architectures early in the design process.

Agilent Technologies, EEsof EDA
http://www.agilent.com/find/eesof-systemvue-radar-library
Email this article to a friend!