Filtering IF samples to reduce computational load of frequency domain acquisition in GNSS receivers.

Abstract

Frequency domain search is a preferred method for rapid acquisition of GNSS signals. In a frequency domain acquisition, the spectra of both local and incoming signals are processed and the result is then returned to the time domain. In this process, most of the computational resources are consumed in inter-domain translation of signals. An FFT algorithm is generally used for these translations. Hence most of the computational resources, in a frequency domain search, are consumed by the FFT algorithm. The computational load of the FFT algorithm depends on the number of samples to be processed. Reducing the number of samples required by the FFT algorithm can thus reduce the computational load of frequency domain acquisition. The minimum number of samples required by the FFT algorithm in a signal search depends on the length of the PRN code and its chipping rate. For a half-chip spacing correlator, at least 2fcT samples are required by the FFT algorithm, where fc and T are the code chipping rate and code period (in seconds), respectively. The authors propose an approach to reduce the required number of samples below 2fcT. In this approach, the IF signal is passed through an anti-aliasing filter and then down sampled to a frequency that is twice the filter cutoff. Selection of the filter cutoff is a trade off between desired improvement in the computational load and the correlation loss (caused by the filter). A lower cutoff is desired to minimize the computational load while keeping the correlation loss within acceptable limits. Acquisition of the new civilian GPS L2C signal and the L1 C/A signal is performed with the proposed method on both software and hardware platforms. A comparison of the performance of the proposed method is made with the conventional approach for frequency domain acquisition. It is shown that the proposed method achieves a significant reduction in the computational load at the cost of minor correlation loss.