Program Summary
SDR and software development kits (SDK) such as GNU Radio exist as free and open source technologies that are widely used in research, industry, academia, government, and hobbyist environments to support both wireless communications research and real-world radio systems. However, even with high end multi-core x86 central processing units (CPU) there are adaptive radar, electronic warfare (EW), and communications applications that cannot be implemented onto SDR with a purely homogeneous CPU due to high latency and power consumption. Furthermore coprocessors like field programmable gate arrays (FPGA) and graphics processing units (GPU) are difficult to program, and prior attempts abstract them to support direct use of open source SDKs have failed to achieve sufficient processing efficiency to support more complex waveforms and protocols.
Therefore, SDR 4.0 will develop technology that significantly improves the ability for SDRs to process radio frequency signals and efficiently execute flexible, easy-to-program open source software. The research effort will optimize how the GNU radio scheduler controls allocation of memory buffers in order to improve the efficiency of data transfer to and from coprocessor devices. Improving data transfer efficiency will significantly improve signal processing capabilities and enhance mission performance associated with radio frequency applications. The research effort will also improve signal processing capabilities available in GNU Radio for use in heterogeneous processing stacks that include a mix of CPUs, FPGAs, and GPUs. Offloading key signal processing tasks to an FPGA or GPU while managing memory input/output will enable faster and more power-efficient computation of the specific signal-related mathematics and therefore lower the overall power consumption of the system.
SDR 4.0 will develop an improved foundation of SDR processing for use in both government and commercial research. It is not creating new hardware or developing the means of expanding SDR multi-functionality. In the commercial world, accelerated processing will enable GNU Radio open–source software users to innovate on a whole host of new ideas with broad customer appeal. Commercial research coupled with DoD research positively impacts numerous defense applications of SDRs like those mentioned previously. For DARPA, diverse radio functionality enabled by improved signals processing contributes to advanced Mosaic Warfare objectives such as system recomposability.