Blind interference mitigation in multi-antenna GNSS receivers

The vulnerability of Global Navigation Satellite Systems (GNSS) to jamming is a well-known issue that threatens the use of GNSS for safety-critical applications. In recent years, the development of interference mitigation techniques on the receiver side has become a major research area especially due to the increasing availability of mobile jamming devices and their growing popularity among civilians. In case of a single-antenna receiver, adaptive filter methods in the frequency/time domain can be used to mitigate frequency-limited or time-limited interference signals. However, these methods are ineffective against wide-band interference. This kind of interference can efficiently be detected and mitigated in the spatial domain, i.e. using multi-antennas and adaptive spatial signal processing. Spatial processing techniques are, however, associated with higher implementation costs and higher power consumption due to their multidimensional character and their higher computational complexity. The main aim of this thesis is to explore effective, area and energy-efficient as well as easy-to-implement signal processing techniques for interference mitigation in GNSS receivers. The thesis focuses primarily on interference mitigation but it strives also for an appropriate multipath rejection performance at the same time. The first step is the evaluation of existing beamforming techniques and a careful consideration of their advantages and disadvantages as well as their implementation costs. Since the choice of the word lengths are directly linked to the hardware costs, special attention is given to quantization effects. A low-complexity 2-step adaptive beamformer, that minimizes the Minimum Mean Square Error (MMSE), is a major outcome of the thesis. The two steps of this method are (a) a pre-correlation interference mitigation and (b) a post-correlation multipath mitigation. Like the conventional MMSE beamformer, the presented 2-step beamformer is a so-called “blind approach”: It does not require any knowledge about the antenna characteristics (geometry and coupling), signal directions, or any kind of calibration. Consequently, its implementation becomes relatively simple compared to many other spatial processing techniques (such as the deterministic beamformer, the Minimum Variance Distortionless Response (MVDR) beamformer, or parametric spatial processing methods). The pre-correlation step is an adaptive pre-whitening filter, whose coefficients are computed based on the Eigenvalue decomposition of the input covariance matrix. A recursive form of this filter is presented, which reduces the computational complexity even further. The post-correlation step is a refined version of the MMSE beamformer. The conventional MMSE beamformer is susceptible to multipath-induced errors. The proposed modifications give the MMSE beamformer a multipath-rejective character. In the final chapter, the proposed 2-step beamformer is combined with temporal filtering (with a bank of notch filters in particular) to form a spatiotemporal filter. The advantage of the combined filter is that it can mitigate a significantly increased number of interference signals by making use of the spatial as well as the temporal characteristics of the received signal. The thesis presents the performance improvement of the combined filter relative to the spatial-only filter. Furthermore, an estimate of the additional resource requirement of the combined approach is provided