Software Defined Radios are computer based systems that emulate the behavior of traditional radio systems by processing digitized radio signals. A SDR replaces the traditional fixed hardware radio with a system that may be reconfigured, both during operation to provide greater flexibility and by providing software upgrades to add new capabilities without requiring new hardware. These are powerful reasons for using SDR technology; however this flexibility comes at the cost of increased hardware cost and greater power consumption compared with traditional hardware radios.
This report presents measurements of memory and processor usage for a Software Communication Architecture (SCA) waveform running on an OMAP starter kit and a desktop PC. The process used to build software, originally targeted for a desktop computer, on an embedded machine with a different processor architecture is described. OSSIE, an open source SCA implementation developed at Virginia Tech, was ported to the ARM processor by adding support for building OSSIE into the OpenEmbedded build system. Once the port for the OMAP starter kit was complete, it became possible to easily re-target OSSIE for a variety of other hardware platforms.
For testing purposes a simple waveform capable of transmitting several common digital modulation formats was developed. A SCA device for the Universal Software Radio Peripheral was developed to interface the waveform to the antenna.
One method to reduce the cost and power consumption is to limit the amount of memory used in the radio. This reduces both cost and power consumption. This report describes the memory manager portion of the Linux kernel and how it helps reduce the memory used by the system. The exmap tool for accurately measuring memory usage is described and used to measure the memory usage of the OSSIE based test waveform. These techniques help radio developers measure and reduce the amount of memory required for the SDR, reducing system cost and power consumption.
Finally, the oprofile was used to measure relative processor usage of the waveform components. Oprofile can also provide details about specific sections of waveform code that use the most processor cycles. This information helps the radio designer reduce the number of processing cycles required. This allows the hardware to use a lower speed part, or add more capability to the radio design. / Master of Science
| Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/36272 |
| Date | 21 January 2008 |
| Creators | Balister, Philip John |
| Contributors | Electrical and Computer Engineering, Reed, Jeffrey H., Pratt, Timothy J., Dietrich, Carl B. |
| Publisher | Virginia Tech |
| Source Sets | Virginia Tech Theses and Dissertation |
| Detected Language | English |
| Type | Thesis |
| Format | application/pdf |
| Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
| Relation | thesis.pdf |
Page generated in 0.0022 seconds