MSc Dissertation: Joseph Wamicha



Adobe-PDF-downloadWamicha, Joseph Nguya. Software Defined Radio OFDM Implementation, Mixed Signal Circuit Design and EBG Antenna Design. MSc Dissertation. Department of Electrical Engineering, University of Cape Town, 2011.



So how did my interest in software defined radio and antennas come about? It all started in the month of February, 2006 when two of us deranged and hungry friends were trying our hands at entrepreneurship. We managed to convince one of the 60 kbps – 100kpbs CDMA2000 mobile networks at the time, Popote, to provide a video streaming premium rated service to their clients. The market was there and the  mobile network was interested; all we had to do was deliver. So for months we hacked at the video streaming codecs and protocols. We finally managed to get video streaming working to the mobile device working (which was quite a feat at the time as video streaming codecs weren’t nearly as versatile as today), but no matter how hard we tried to optimise the codecs, we still couldn’t get a high definition video stream to the mobile cdma2000 device. It goes without saying that we lost a potentially lucrative product, Popote could not survive since they were unable to provide the premium rated services larger GSM networks such as Safaricom (of MPESA fame) were providing, and yes, we burnt our hands at entrepreneurship like so many others before us and continued to remain hungry and broke.

There was a positive outcome from the experience though. The question still remained: why were we unable to efficiently stream a video, despite our having used highly optimized video streaming algorithms? It was then that it struck me that this was a problem I’d never be able to solve using software alone. The bottle neck was in the hardware and I’d need to get into it’s murky guts to solve this problem once and for
all. It aroused inside me a curiosity inside me about wanting to find out in detail how mobile networks work so that I could finally solve my problem. This curiosity finally led me into UCT, where I pursued my research. During my research, I discovered that the two bottle necks were with the antenna and base station hardware. I sought to understand how these two modules of the mobile network work and it is the results that form the bulk of my thesis. Since I believe that the only way Africa can develop is by developing it’s own intellectual property using open source software tools, I sought to use only open source software and hardware during my research. Besides, open source software is free of licensing costs and is a lot more affordable for the price conscious African market. It is also open to anyone who wants to understand how things work. Hopefully with the end of my Masters research project, I can finally set this curious case to rest.

Due to its potential to revolutionize modern digital communications, Software Defined Radio (SDR) has become a hot research topic in recent years. This is because using Software Defined Radio, we can re-create radio communications blocks such as modulators, demodulators, filters and amplifiers from their traditional hardware implementations into new next generation, re-configurable sofware based radio communications components. The software radio components will run on a base hardware platform such as the traditional x-86 computing platform, GPUs (General Processing Units) and more recently FPGAs (Field Programmable Gate Arrays).

In theory, since the software running on the base hardware is inter-changeable, this would make software defined radio communications a lot more versatile than its hardware counterpart. The software defined radio may for example act as aWiMAX base station today, and through altering the software running on the base hardware, act as an LTE base station tomorrow. Without having to do any modifications on the base hardware, the Software Defined Radio is also upgradeable. As a result, software defined radios would possibly have much longer lifetimes than pure hardware-based implementations, thereby making them a lot cheaper to invest in, in the long run. For example, through a software upgrade, the base hardware may change from being an LTE base station to an Advanced LTE base station.

In the first part of my thesis I seek to verify the plausibility of SDR causing a revolution in next generation radio communications by doing an actual OFDM implementation for the IEEE, 802.11g WLAN (Wireless Local Area Network) specification using the Gnuradio SDR platform. In the second part of my thesis I examine open source electronic design automation using gEDA, by designing a front-end for a new low cost FPGA board called Rhino (Reconfigurable Hardware Interface for Computing and Radio). Rhino is capable of far more powerful DSP (Digital Signal Processing) than the more popular Gnuradio, and would provide communications engineers with a much better platform for wireless communications prototyping. Since the Rhino FPGA board does not have ADC or DAC chips for analog to digital or digital to analog conversion processes respectively, my work was to design the Rhino Expansion Board which is an ADC/DAC mezzanine board designed for Gnuradio RF Acquisition daughterboards to plug into the Rhino FPGA board. This would enable the SDR community to utilize the processing power of the Rhino FPGA Board. Also, the Rhino expansion board would allow re-usability of already proven hardware, instead of designing new RF Acquisition hardware from scratch.

In the second part of my thesis I also seek to establish whether open source EDA has matured enough for it to be used by Industry. If it is established that this is indeed the case, then this could potentially lower the cost of designing new hardware, which would be a huge benefit to lowering the barriers of entry to technology manufacturing for the nascent African electronics industry.

Finally in the last part of my thesis I use the open source Gnuradio SDR platform to test an EBG (Electronic Band Gap) edge-fed microstrip patch antenna. The open source Openmoko mobile phone can be used to collect receive sensitivity (power) and antenna range measurements. EBG structures are used in patch antenna designs since they allow us to make meta-materials with new EM properties that are not ordinarily found in nature. As a result, further research that may be carried out on EBG antennas could potentially be used to design far more sensitive nano antennas in future.

We recommend that the pdf version of the thesis be read, to aid in better navigation of the content using the hyperlinks.



Leave a Reply