To read and download the complete Boeing Technical Journal paper titled “On-Demand Waveform Design for Software-Defined Radio Applications,” click here.
The exponential growth in demand for wireless communications, both for civilian and military use, has created an urgent need to use as much of the increasingly scarce available radio frequency (RF) bandwidth as possible. This remaining available bandwidth is usually non-contiguous, time-varying and non-uniform in propagation characteristics.
The ideal would be to have a single physical radio that implements all existing communication signals and protocols (one radio that does it all), as well as one that adapts to new conditions to effectively use as many communications resources as possible, while still meeting applicable Federal Communications Commission and World Radio Conference requirements.
Software-defined radio (SDR) technology brings these benefits of flexibility and cost efficiency to end users. An SDR typically includes a collection of hardware and software technologies where some or all of the radio’s operating functions are implemented through modifiable software or firmware running on devices including field programmable gate arrays (FPGA), digital signal processors (DSP), general purpose processors (GPP), programmable System on Chip (SoC) and other application specific programmable processors. This allows new wireless features and capabilities to be added to existing SDR systems without requiring new hardware.
Software-defined radios typically use a set of fixed modulation waveforms for a fixed set of channel band-widths and choose among them to meet their operational goals. For example, when a set of non-contiguous channels become available, the SDR chooses one signal from this fixed set for each empty channel that matches that particular channel’s conditions (maximum energy, bandwidth, noise, channel characteristics, etc.). There are drawbacks to this method.
First, the fixed modulation waveforms do not exactly match each channel and so must be chosen conservatively so that all channel conditions are met (especially conditions that define the maximum signal leakage into adjacent channels). This mismatch leads to lower performance.
Second, coding is typically done only within each channel, rather than across channels. This leads to worse performance for certain channels when they are affected by worsening channel impairment over time.
Third, each channel must do its own synchronization (symbol, bit and frame, for example) because each is handled independently. This increases the complexity of the receiver because of the requirement for separate independent real-time parallel signal processing.
This paper describes a way to operate typical SDR systems through the on-demand custom design of modulation waveforms for communication on nearly any available spectrum under an unlimited set of operational conditions.
In summary, the technical paper describes the concept of on-demand designed waveform generation and an approach to implement this concept in both standard and specialized software-defined radio architectures.
- An optimization method is described to design modulation waveforms for SDR applications which enhances, among other things, both synchronization and bandwidth usage compared to traditional modulation methods. In particular, our approach allows:
- Longer range and lower power communications under comparable conditions.
- Communications allowing more efficient spectrum usage.
- Spreading of synchronization and coding functions across multiple channels and disjoint spectral regions.
- A numerically controlled waveform generator is described which allows hardware/firmware to create a sampled version of the designed modulation waveform at any particular frequency for modulation by the symbol sequence of a transmitter or conversely by correlating in the receiver. This waveform can be further tailored to reduce storage requirements through judicious interpolation and various symmetries can be imposed within the optimization to reduce storage as well.
- An on-demand architecture is described that allows creation and utilization of a designed modulation function within a typical software-defined radio architecture. A method of operation is also described that would allow on-demand waveform design to function within an existing SDR system.
Figure 1 shows a standard SDR system while Figure 2 shows the corresponding on-demand designed waveform architecture introduced by this paper. The three main differences are the replacement of a fixed set of modulation waveforms with a processing unit that solves an optimization problem to design a new modulation waveform particular to the measured channel conditions and required data rates, together with modified baseband processing using these new modulation waveforms. Implementation of on-demand waveform design with SDR requires four items:
- A method of mapping communications channel conditions into a mathematical description that can be used via standard optimization processes to generate the desired waveforms.
- A transmitter and receiver architecture that implements on-demand designed waveform communication using the numerically controlled waveform generator.
- On-demand waveform SDR operation implemented within the larger software-defined network operation.
These four items are described in sections III-VI within the full Boeing Technical Journal paper available online. Specifically, section III covers the mathematical description of the design process for modulation waveform generation; section IV covers the numerically controlled waveform generator; and section V describes the on-demand designed waveform transmitter and receiver architecture, amplifying what was shown in Figures 1 and 2.
The approach described was tested on several design problems and the results are further described in the full paper. Finally, this architecture was implemented using the Universal Software Radio Peripheral (USRP) platform (GNU radio hardware) from National Instruments/Ettus Research at Boeing’s 7-107 laboratory complex. This implementation and testing are described in the full paper.
This waveform method creates a single modulation function across the likely non-contiguous available spectrum with near-optimal synchronization and correlation properties and with near-optimal bandwidth efficiency. Because of its designed structure, this modulation waveform can increase the bit-error rate performance and/or range of almost any existing wideband SDR, while increasing robust performance under varying channel conditions. In addition, an SDR designed using the new hardware architecture described in this paper has a simpler structure than a traditional set of parallel digital modems, thus simplifying its design.
Many applications of our approach are possible for both the military and commercial communications. In the commercial arena, possible application areas include:
- Better spectrum usage for public channels such as ISM.
- Operation within White Space (IEEE 802.22), such as within newly available UHF channels (formerly allocated to television).
- Ad hoc wireless system and emergency service providers.
- Mobile satellite and wireless systems.
- Increasing manufacturing network capacity while operating heritage communications networks.
Future work is planned to address real-world performance metrics to compare on-demand waveform design with traditional modulation approaches in given SDR scenarios. This work could also compare competing approaches, such as orthogonal frequency-division multiplexing, to design waveforms for specific tasks such as spanning non-contiguous transmission bands. We also plan to address the computational requirements and latency in more detail, especially for embedded applications such as mobile radios, as well as coordination among radios using the proposed on-demand protocol.
By Gary Ray
Gary Ray is a Technical Fellow in Boeing Research & Technology, with more than 30 years in the field of communications and signal processing, including experience at Hughes Aircraft and Westinghouse Hanford.