ION GNSS 2012
Session D3: GNSS Algorithms & Methods 1: Signal Processing

Title: Acquisition of Weak GNSS Signals Using Fast Orthogonal Search
Author(s): M. Tamazin, A. Noureldin and M. Korenberg, Queen´s University, Canada
Date/Time: Thursday, September 20, 2012, 8:35 a.m.
Room: 204 (NCC)

The demand for personal navigation is driving research and development towards enhanced civilian GNSS receiver technology for use in increasingly difficult operational environments. Existing GNSS receivers suffer significant degradations in many environments such as in indoor environments and urban canyons due to restricted visibility of available satellites. Even if a sufficient number of satellites are available, the noise and multipath can often be large, leading to large errors in position. The signal attenuation and heavy multipath which are found in these environments distort the code delay estimate and lead to inaccurate positioning. In addition, these two elements make acquisition and tracking processing of standard GNSS receivers very difficult. Traditional GNSS signal processing is ineffective, particularly, for initial detection, or acquisition, of the GNSS signals. Fast Orthogonal Search (FOS), introduced by Michael Korenberg at 1987, is a general-purpose non-linear modelling technique that finds functional expansions using an arbitrary set of non-orthogonal candidate functions. The algorithm can be applied to a variety of practical problems such as signal processing, and applications of FOS in other research fields have shown successful detection of signals buried in both short-term and long-term errors. Fourier transformation of a signal, which used in traditional GNSS acquisition methods, is a numerically ill-posed problem as small amounts of noise corruption of the time-series data can result in large errors in the frequency domain, likewise, missing samples or unequally spaced data may result in an inaccurate spectral model. The advantage of FOS is its ability to cope with missing or unequally spaced data and recover signals heavily contaminated with noise. There are two significant differences between FOS and conventional Fourier transforms techniques (i.e. discrete Fourier transforms (DFT) or FFT): [1] FOS yields a parsimonious sinusoidal series representation by selecting the most significant sinusoidal components; and [2] the frequencies of the sinusoids selected do not have to be commensurate nor integral multiples of the fundamental frequency corresponding to the record length. The latter difference is significant as it allows FOS to model frequency components that are separated by less than the FFT bin. This translates to better frequency resolution in the spectral model. Therefore, this research utilizes FOS to increase the availability of the navigation solution for acquisition of weak GNSS signals. FOS was used in this paper to significantly reduce the effect of noise and enhance the accuracy of the GNSS signal contaminated by noise and multipath. The advantages of this approach in reducing some of the undesired signals are also explored in this research. The outcome of post processing acquisition attempts using FOS algorithm on actual GNSS signals of power levels significantly below the nominally received power levels are discussed and demonstrated in this paper. The target here is to focus on the acquisition rather than tracking since it is possible to derive coarse (1-100m) position estimates from the acquisition result alone. The parallel frequency space search acquisition method is used as a reference to evaluate the performance of the proposed method. The code phase and Doppler frequency are then estimated by cross-correlation in the frequency domain which uses a Fast Fourier transform (FFT) to perform a transformation from the time domain into the frequency domain. Data has already been collected in two indoor test scenarios, namely in a suburban home and in an engineering laboratory. Processing is conducted on raw Intermediate Frequency (IF) samples collected synchronously from antennas at two locations, one with a clear view of the sky (the reference), and one in a degraded environment (the rover). The front-ends used to collect this data are driven by the same local oscillator, thereby ensuring that both reference and rover data are subject to the same clock bias and drift effects. Processing is conducted on raw Intermediate Frequency (IF) samples collected from the antenna. The data is processed by a NavINST GPS Software Receiver (NGSRx), developed by the Navigation and Instrumentation (NavINST) research group. For the scope of this paper, we focus our analysis on the GPS system with L1 C/A code signals. We decided to also exclude multipath signals when considering signal acquisition analysis and processing. The use of FOS for Multipath mitigation will be explored in future research. The analysis of the GPS signals in indoor environments using the FOS algorithm enhanced the GPS signal power relative to the noise and improved the solution availability. We observed that the noise variance was reduced by using the FOS method. In general, if compared to FFT, FOS has a greater computational load and relatively more complex to implement; however, our research team was able to optimize the FOS algorithm for real - time operation. Therefore, this paper offers a reliable method based on FOS algorithm that can: (1) efficiently replace present techniques (mostly relying on FFT); (2) acquire weak GNSS signals in harsh environments; and (3) capable of operating in real-time.

Return to Session D3 Next Abstract