Signal-processing challenges ongoing and evolving
Advancements in processor technology mean that engineers and designers are able to meet the demand of signal-processing requirements in applications ranging from radar and software-defined radios to high-performance embedded computing (HPEC), and space systems. Designing for exacting high- bandwidth as well as size, weight, and power (SWaP) requirements are no longer a major challenge, but now are simply a checkbox in the design process. Even so, challenges remain, but designers are finding ways to face those issues by using fiber-optic technology, multicore processors, commercial off-the-shelf (COTS) components, and codesign methodologies.
Today’s reality: Servers are smaller, while their processing power and capabilities have increased. Users have the option to either build a custom-made system or use COTS components and reap those benefits. Intel processors have, arguably, taken over the market and solved processing capabilities that systems may require.
Intel is the dominant player in this world for a simple reason, “Intel has been instrumental in defining not just the core microprocessor and instruction-set architecture, but also the architecture of peripherals,” says Dirk Finstell, EVP for the Module Computer Product Segment at Adlink Technology in Frankfurt, Germany. “Companies that provide embedded-computing products based on the x86 architecture have been able to leverage that (chip-level) expertise by providing either proprietary or open-standard products that use a common I/O interface. Through the use of common connector pinouts, it is possible for customers to select from a wide range of hardware- and software-compatible peripherals with which they can customize their end products.”
Customization is one avenue that designers go down, depending on the need of the user. The technology that has become the game changer in 2015? “The advent of multicore processors from Intel and Freescale as well as the use of multicore ARM processors in nearly every smartphone, everywhere,” states Doug Patterson, VP – Military and Aerospace Business at Aitech Defense Systems in Chatsworth, California. “Integrating the memory crossbar switch into the processor silicon took the bane of memory bandwidth limitations out of the ‘lower-performance’ and ‘memory-thrashing’ equations altogether. Faster and larger memory has fueled more software developer creativity and more functional and capable systems. Software and operating systems – supporting virtual memory, multiuser and multiprocessing parallel executions of applications simultaneously – have advanced to finally start meeting the promises of true portability and application auto-level loading.”
Another avenue is fiber-optic technology, which has enabled applications that require large amounts of data to be transferred at high speeds. It provides a solution that, says Thierry Wastiaux, senior vice president of sales at Interface Concept, is “suitable for use in connecting the thousands of transmit/receive modules of the active antennas used in an AESA radar platform to the signal-processing system and more generally for connecting sensors generating important flows of data.”
Radar and electronic-warfare applications have been able to implement fiber-optic technology in ways never considered in the past. Gérald Persaud, vice-president of product management of Reflex Photonics in Pointe-Claire, Canada states, “Radar is a key platform for fiber optics, especially phased-array radar, which requires a tremendous amount of bandwidth between the antenna array and the beamforming computer. When we get to the signal-processing computer, we’re also seeing fiber optics being used among processors and FPGA [field-programmable gate array] boards to scale processing.”
Design challenges are not extinct, however, as engineers grapple with puzzles ranging from cost to ruggedization to meeting tomorrow’s requirements. The good news is that companies have new avenues open to solve these issues, whether that means specifying COTS components or covalidating during the design process to keep costs down.
“Sometimes it’s cheaper to solve it with a better signal-processing algorithm,” says Dr. Murthy Upmaka, an application engineer at Keysight Technologies in New York City. “Knowing the best place to solve an issue and validating the solution before the final product is assembled is a benefit of an integrated codesign approach, as is the ability to minimize over design and improve system performance for a given budget.” says.
Sometimes using COTS components is not always the best option as in space applications for example. While COTS components are attractive because they offer so much in terms of bandwidth and performance, critical space applications call for proven, radiation-hardened components that will last the life cycle of the end use and ensure reliability.
To save money, the industry is heading toward “cheaper multicore processors [which] will be the norm for nearly all applications, with operating systems and apps taking advantage of the huge increases in parallelism,” Patterson says. Prices of the processors and memory will drop quickly as demand increases almost exponentially across all markets: commercial/consumer, industrial and defense. Interprocessor communications will continue to rise quickly and efficiently, passing multiple gigabytes-per-second of data between local [onboard] nodes as well as nodes across board boundaries via high-speed copper, then optical pathways, as the fears of implementing and using optical fiber are swiftly moved to the past.”