Adaptive Spatial Filter for Autocorrelated Interference Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing adaptive spatial filtering methods fail to effectively suppress interference signals with autocorrelation properties, leading to inadequate detection of target signals, especially when noise and interference power exceeds that of the target signal, and often result in self-extinction of the target signal due to imprecise determination of its bearing value.
Innovation Solution
A method for spatial filtering that optimizes the sum of squares of autocorrelations of the output signal for multiple time delays under a constant filter response constraint, specifically targeting interference signals with autocorrelation properties while minimizing suppression of temporally uncorrelated noise, using time-lagged cross-covariance matrices and PARAFAC decomposition to enhance signal extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If traditional MVDR spatial filtering is used to suppress interference, then noise and interference are minimized, but interference signals with autocorrelation properties cannot be effectively suppressed and target signals may self-extinguish
Solution Approach 1:
The patent changes the optimization parameter from minimizing total power (MVDR) to minimizing the sum of squares of autocorrelations at multiple time delays. This parameter change allows the filter to specifically target and suppress interference signals with autocorrelation properties while preserving target signals through their distinct autocorrelation patterns.
Solution Approach 2:
The patent extends the filtering criterion from a single-time-point power minimization to a multi-dimensional optimization across multiple time delays. By considering autocorrelations at different time lags, the filter gains additional dimensional information to distinguish target signals from interference, resolving the contradiction between interference suppression and target signal preservation.
2Loss of energy
If adaptive spatial filtering minimizes total power at filter output, then noise is suppressed, but autocorrelated interference signals with higher power cannot be adequately suppressed
Solution Approach 1:
The optimization criterion is changed from minimizing instantaneous power to minimizing the sum of squared autocorrelations across multiple time delays. This parameter transformation enables the filter to identify and suppress autocorrelated interference signals based on their temporal correlation structure rather than just their power level.
Solution Approach 2:
The patent introduces autocorrelation functions at multiple time delays as an intermediary measure. This intermediary allows the filter to indirectly detect and suppress autocorrelated interference signals by targeting their temporal correlation properties, which serve as a characteristic signature distinguishing them from uncorrelated noise.
3Measurement precision
If precise bearing value determination is required for optimal filter performance, then signal-to-interference ratio is maximized, but self-extinction of target signal occurs due to imprecise bearing determination
Solution Approach 1:
The patent changes the optimization approach from bearing-value-dependent filtering to autocorrelation-based filtering. By optimizing the sum of squared autocorrelations at multiple time delays, the filter becomes less sensitive to imprecise bearing determination while maintaining effective interference suppression through the characteristic autocorrelation patterns.
Solution Approach 2:
The filter utilizes the inherent autocorrelation properties of the signals themselves as the basis for suppression, rather than relying on externally determined bearing values. This self-service approach allows the filter to automatically distinguish target signals from interference based on their intrinsic temporal correlation characteristics, reducing dependency on precise bearing measurement.
Data Source
Figure 1
Figure 2~4
Figure 5
AI summary
To selectively suppress interference signals exhibiting temporal correlations more effectively than time-uncorrelated noise signals, an adaptive spatial filter is proposed. This filter can be calculated by optimizing the sum of the squared magnitudes of the autocorrelations of the filtered signal, subject to the constraint of a constant filter response with respect to the control vector of the target signal. Furthermore, the optimization problem, as illustrated in the exemplary implementations, is initialized using loading vectors of the PARAFAC decomposition of the time-delayed cross-covariance matrices for freely selectable time delays and the inverse matrix of the weighted sum of the outer product of any two loading vectors of the PARAFAC decomposition.