Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > RF/Microwave

Next-gen radios bend toward flexibility

Posted: 01 Feb 2006 ?? ?Print Version ?Bookmark and Share

Keywords:peter varhol? ee times? software defined radio? sdr? radio technology?

One major stumbling block in the transition of advanced wireless standards into commercially viable communications devices is the inability to easily modify those products to adapt to multiple link-level standards and bandwidths. This limits the useful commercial life of hardware, which in turn affects design parameters. In short, wireless handsets that can't adapt when technology or market requirements change won't receive the investment in engineering resources they need to compete.

Software-defined radio (SDR) provides a set of radio technologies that can be dynamically programmed to support different waveforms, meet new and emerging standards for link-level connectivity, provide new communications protocols and features, improve performance, and deliver new services. The military has used SDR to provide soldiers with radios that can download software modules and establish contact with support aircraft, surveillance flights and other weapons systems that use different waveforms and frequencies.

Can the technology offer similar advantages to commercial wireless-handset vendors faced with short product-viability windows, differing equipment standards and a high level of cost sensitivity? In principle, SDR seems to offer a ready and compelling solution. But there are significant challenges in designing and implementing an optimal solution. Generally speaking, a hardware platform that is generic enough to support a variety of features in software must incorporate expensive components, use a lot of power, or both. Is there a way to break this trade-off cycle?

There is certainly ample motivation to do so. The payoff may be especially large on the far side of the connection. The realities of the handset life cycle may not make an explicit software solution as compelling as it is for the wireless base station, which is considerably more expensive and has a longer expected useful life. Replacing dedicated waveform-processing hardware that lacks a multiband capability with software processing can save hundreds of thousands of dollars over the life of the base station.

But the rationale is different for handsets. Here, the goal is to build in support that makes it possible for the user to add new services, travel across national boundaries and still obtain a valid transmission signal. This requires the ability to obtain and decode a wide variety of waveforms, potentially across multiple bandwidths.

Back to the specs
But design methods tied to particular hardware architectures make it necessary to start again from the specifications to port to a new hardware platform. That is one reason selection of the hardware components can make a big difference.

The sooner the signal is digitized, the sooner software modules can be applied to it to adapt to differing characteristics. A/D conversion represents the first opportunity for considering design trade-offs. Doing it in a way that optimizes power and cost depends on the number of communications channels that must be processed to satisfy a specific need. Web surfing can potentially be allocated more channels than voice communication, for example.

Pentek Inc.'s solution is to consider replacing standard hardware components with FPGAs that implement special-purpose functions. "If an FPGA can be coded to process a digitized input at an equivalent performance to that of several standard DSPs, the cost and power consumption could well be lower," said Rodger Hosking, VP at Pentek.

Texas Instruments Inc., in contrast, focuses on building standard hardware with the flexibility to manage increasing numbers of waveforms, bandwidths and protocols. "It started a number of years ago," said Bill Krenik, chief technology officer of TI's wireless division. "Handsets have needed expanded capabilities to work with new bandwidth requirements; it's more cost-effective to build this capability into the hardware platform and use software for control."

TI can refine its manufacturing process to meet requirements. The company's current 65nm process can combine hardware for multiple standards. As needs grow over the next several years, TI anticipates being able to migrate to 32nm. That should provide room to incorporate more standards while lowering overall power consumption, according to the company.

Modeling, simulation
The risk of creating an inadequate hardware platform for future and possibly unknown communications features is one that won't be determined until those features are implemented in software. Designers may determine that a hardware platform lacks the performance or battery power to support needed features after it is deployed, requiring that the platform be phased out to move forward.

A second risk is that of making assumptions concerning the platform that cannot easily be proven correct without completing significant parts of the software implementation. While the hardware platform may not yet be deployed, it could result in a great deal of reengineering and delays in releasing the product.

Both types of risk beg the question of whether it is possible to get a good indication of the likelihood of a hardware-software (HW/SW) mismatch much earlier in the game. One solution is to model the components and simulate their interactions. Hardware engineers have used modeling and simulation for years. Is it possible to add software into that mix?

Models of both hardware and software are becoming much more realistic, sometimes to the point where the model can be used as part of the implementation. Vendors of analytical software, such as The MathWorks Inc. and Wolfram Research Inc., enable designers to create comprehensive models of HW/SW interaction and to simulate that interaction. By doing so, they can address technical issues and change parameters of the model to analyze "what if" scenarios.

Best solution
Because the features that will be needed in the future are impossible to predict accurately, design of the hardware platform boils down to intelligent guesswork. Nonetheless, based on past requirements, there are some generalities that should provide a working guide:

? Use as much processing power as you can afford. That refers to both the GPP and the DSP. Adding software modules with advanced digital features will use all that horsepower and then some.
? Specifically with handsets, given higher performance, more memory and software, battery life is at a premium. Use power-management software or firmware to make the most of it.
? Model HW/SW interaction as early in the development stage as possible. This can save costly redesign of both hardware and software if one of them changes before the platform is deployed or if standards evolve after the fact.
? Use standard hardware platforms and software modules where possible. Trade-offs can certainly be made in the areas of power, performance and cost, but a standard platform provides known physical characteristics from which to offer a variety of software modules.

There are no hard-and-fast rules that will provide the best SDR wireless transceiver under all conditions, since those conditions are too diverse. Base stations have somewhat more common characteristics, so the need to adapt over longer periods of time makes the technical trade-offs different, but just as complex.

Vendors are increasingly making it possible to extend the useful life of both device and infrastructure through the use of multiple hardware standards plus add-on software. Designers can intelligently use these components to ensure a higher level of compatibility in wireless products.

- Peter Varhol
EE Times

Article Comments - Next-gen radios bend toward flexibil...
*? You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.

Back to Top