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

Title: A Cross-Correlation Mitigation Method Based on Subspace Projection for GPS Receiver
Author(s): L. Chen, W. Meng, S. Han, E. Liu, Harbin Institute of Technology, China
Date/Time: Thursday, September 20, 2012, 9:43 a.m.
Room: 204 (NCC)

GPS is a CDMA system using DSSS, and the length of the C/A code used by GPS L1 band civil signal is 1023 chips. As the cross-correlation peak between the C/A codes is not zero, it will degrade the performance of the detection probability and the bit error rate when the strong one coexists. This is of particular concern in applications of GPS Pseudolite and some special signal propagation environments, such as indoors and urban crayons. A number of different cross-correlation mitigation techniques have been developed for CDMA communication system. The direct interference cancellation technique is relatively applicable for GPS receiver. It´s a subtraction type method in which strong signal is subtracted from the input signal prior to the correlation of the weaker one. For reconstructing the strong signals, the parameters of strong signals such as code frequency, code phase, carrier frequency, carrier phase, signal amplitude and data-bit modulation should be accurately estimated. But in fact, the estimated parameters contain inevitable errors, especially signal amplitude and carrier phase. Further discussion about this method will be found in [1]. Literature [2] provides a sub-optimal approach to dealing with the near-far problem in GPS, but it is suitable only for the application in which the number of strong signals is less than four. In this paper, we analyze the problem caused by the cross-correlation and propose a cross-correlation mitigation method based on subspace projection. In comparison with direct interference cancellation technique, only the code phases, carrier frequencies and date-bit modulation of strong signals are required and the method is independent of the carrier phases of the strong signals. The major components of this paper are given as follows. Firstly, we analyze the basic principle in theory and present the architecture of the cross-correlation mitigation method based on subspace projection. If we have acquired the code phases, carrier frequencies and date-bit modulation of all strong signals, we can construct the strong signal subspace matrix. The projection of the input signals onto the strong signal subspace is the sum of the strong signals and the projection of the noise. Then we can use the projection as an estimate of the strong signal. After subtracting the estimate from the input signals, the result consists of weak signals, noises and the residue error resulted from the projection operation. Then the acquisition and tracking of the weak signals can be done. We also prove that the performance of this method is independent of the carrier phases of the strong signals. And two different scenarios have been discussed. One is that the strengths of the strong signals are almost the same. Another scenario is that the strengths of the strong signals are largely different. Then, we consider the SINR (weak signal is desired signal and strong signal is interference) after mitigation, the estimated C/N0 (Weak Signal to Noise Ratio), the bit error rate and detection probability as the evaluation of the performance of this cross-correlation mitigation method. The simulation results indicate that the method can suppress the cross-correlation interference effectively and provide at least 33 dB improvement of SINR when the SINR of input signals is -30 dB to -10 dB and C/N0 is 44 dB-Hz. The C/N0 estimation algorithms selected in this paper are Beaulieu´s Method and Moment Method. As the C/N0 of input signals is 30 dB-Hz to 60 dB-Hz and the SINR is -20 dB, the estimated C/N0 after cross-correlation mitigation perfectly matches with the value without the strong signals for the Beaulieu´s Method and Moment Method respectively.

The BER curve and detection probability curve obtained after mitigation also match with the curves without strong signal and are much better than the curves obtained before cross-correlation mitigation. For the simulation of the detection probability, the C/N0 of input signals is 36 dB-Hz to 52 dB-Hz in 1 dB step and the SINR is -13 dB, 1ms coherent integration is used. During the simulations, only one weak signal and one strong signal exist. And we also test that when the residual Doppler frequency of strong signals is less than 10 Hz, the algorithm can work as well. Thirdly, as the commercial GPS receivers are always equip 2-bit ADC, the influence of the quantization loss on the performance of the mitigation method is analyzed. When the C/N0 of input signals is 30 dB-Hz to 45 dB-Hz and SINR is -20 dB, the experimental results used 2-bit ADC is that the estimated C/N0 after mitigation has 1 dB loss compared to that is without strong signals. The detection probability also has almost 1 dB loss. Additionally as the C/N0 of input signals increases from 45 dB-Hz to 60 dB-Hz, the loss of estimated C/N0 is greatly increasing. When the C/N0 is 55 dB-Hz, the loss increases to 20 dB and the detection probability is almost zero. As one knows when the C/N0 increases and SINR is -20 dB, the power of the strong signals also increases and will exceed the noise power. Due to the strong signals will augment the quantization noise of the weak signals, the loss of estimated C/N0 will increase. When using 4-bit ADC and the C/N0 increasing from 30 dB-Hz to 60 dB-Hz, the loss of estimated C/N0 after mitigation will be very small. And the loss used 7-bit ADC is almost zero. Therefore the mitigation method requires 4-bit ADC at least. Finally, we summarize the analysis and solutions. The cross-correlation mitigation method based on subspace projection proposed in the paper can effectively mitigate the cross-correlation interference. The method requires less parameters and has better stability than the direct cancellation method. Besides, it can be easily implemented on software receiver.

[1]P. H. Madhani, P. Axelrad, K. Krumvieda, J. Thomas, "Application of Successive Interference Cancellation to GPS Pseudolite Near-Far Problem," IEEE Trans. on Aerospace and Electronic Systems, vol. 39, no. 2, pp. 481-488, April 2003. [2]E. P. Glennon and A. G. Dempster, "A novel cross correlation mitigation technique," ION GNSS, Long Beach, CA, Sep. 2005, pp. 190 - 199.

Previous Abstract Return to Session D3 Next Abstract