The Global Positioning System represents the pinnacle of navigation technology for the 21st century. As new technologies integrate GPS services, the limited availability of GPS in environments where the signal is severely attenuated, subject to strong multipath or high dynamics becomes an obstacle to a rapidly growing industry. A novel scheme for processing the GPS signal, namely a software radio employing block-processing techniques similar to those used for image processing has proven to enhance the usability of GPS in such environments. However, these techniques have huge computational requirements that are impossible to meet with a microprocessor. Custom designed hardware, such as an application specific integrated circuit (ASIC) would handle the processing requirement, but defeats the philosophy of a software radio since the algorithms cannot be changed. Field programmable gate arrays (FPGAs) are beginning to replace ASICs in certain applications since they feature software-like re-programmability while approaching ASIC-like performance. FPGAs are excellent candidates for research since they lack the NRE costs associated with ASICs. Hence, FPGAs are the most attractive implementation platform for developing a real-time block-processing GPS receiver.
This work lays the groundwork for the implementation of a real-time block-processing GPS receiver in FPGA hardware. It is a feasibility study since the problem is approached at a high-level of abstraction. The original block-processing approach is re-analyzed for implementation in FPGA hardware. Implementing the 5000-point FFTs in finite-precision hardware represents one of the biggest challenges in this work. This requires analysis of the FFT error bound to determine the minimum precision required that would yield acceptable results while minimizing hardware cost. Even though the analytical error bound for finite-precision FFTs is well documented in past literature, its direct application to the block-processing problem becomes too complex. This work employs statistical results of simulations to deduce the optimum hardware architecture and concludes that real-time capability can be achieved with currently available technology.