Abstract
For the coexistence of SUT (System Under Test) radiative emission signal and ambient interference signal, the amplitude of SUT signal will be submerged by the amplitude of interference signal, so it is difficult to accurately measure the amplitude of SUT signal. In this paper, a two-level nested array is used as the receiving array antenna, and the mixed matrix estimation method based on Blind Source Separation (BSS) is used to separate the coherent groups of the signal. Then the Sparse Reconstruction method is used for the DOA (Degree Of Arrival) estimation of each coherent group of the signal. After the DOA information of each signal is obtained, beamforming method is used to form beams of the main channel and auxiliary channel. The beam of the main channel outputs without distortion in the direction of the SUT signal and forms zero traps in the direction of the coherent signals, while the beam of the auxiliary channel forms zero traps in both the direction of the SUT signal and the direction of the coherent signal. The data received by the array are respectively multiplied by the weights of the main channel and auxiliary channel to obtain the output signals of the two channels. The output signals of the two channels are respectively fed into the Adaptive Noise Cancellation (ANC) system, and the ANC method is used to suppress the ambient interference signals and restore the SUT signal. Simulation and experiment results show that this method can accurately estimate DOA of radiation emission signals, effectively suppress ambient signals and restore the signal of SUT in field measurement of radiation emission.
Introduction
In Electromagnetic Compatibility (EMC) testing, measuring radiation emissions in-situ (on-site) is often necessary for large equipment that cannot fit into anechoic chambers. However, field environments are typically contaminated with high levels of ambient interference from sources like radio towers and Wi-Fi. Standards such as CISPR 16-2-5 require ambient noise to be at least 6 dB below the signal of interest, a condition rarely met in complex industrial settings. This paper proposes a novel measurement method that combines non-uniform antenna arrays with advanced signal processing to isolate the System Under Test (SUT) signal from overwhelming background noise.
Proposed Methodology
The proposed system follows a multi-stage pipeline designed to enhance spatial resolution and suppress interference.
1. Two-Level Nested Array
The system utilizes a non-uniform "nested" array configuration. By strategically spacing a small number of physical antennas, the system creates a "virtual" array with a much larger aperture. This allows for higher resolution Degree of Arrival (DOA) estimation without the hardware complexity of a massive uniform linear array.
2. Signal Separation and DOA Estimation
The received signals undergo Blind Source Separation (BSS) to be categorized into coherent groups. Following this, Sparse Reconstruction is applied to estimate the DOA for each signal source. This approach is robust even when the number of sources exceeds the number of physical sensors, leveraging the virtual aperture of the nested array.
3. Dual-Channel Beamforming and ANC
Two distinct beams are formed:
- Main Channel: Designed to have unity gain in the SUT direction while placing "zero traps" (nulls) in the directions of identified interference.
- Auxiliary Channel: Designed to null the SUT signal entirely, thereby capturing only the ambient interference.
These two channels are then processed by an Adaptive Noise Cancellation (ANC) system. The auxiliary channel serves as a pure noise reference, which is mathematically subtracted from the main channel to recover the clean SUT signal.
Results and Discussion
The method was validated through both numerical simulations and physical experiments. In simulations involving a 6-element nested array and 11 signal sources, the system demonstrated high precision in identifying signal directions.
Key performance metrics include:
- Interference Suppression: Environmental noise was suppressed by over 40 dB in simulations.
- DOA Accuracy: Estimation errors were maintained within a range of 2° to 5°.
- Signal Restoration: In physical experiments at 2.9–3.0 GHz, the SUT signal amplitude was restored with an error of 0.55 dB for incoherent noise and 3.43 dB for more complex coherent noise scenarios.
Conclusion
The combination of nested arrays, BSS-based sparse reconstruction, and dual-channel ANC provides a powerful framework for in-situ EMC measurements. While currently optimized for a 100 MHz bandwidth, the method successfully extracts SUT signals even when they are initially submerged in ambient noise. Future work will focus on reducing computational complexity and extending the frequency agility of the system for broader industrial applications.
Summary updated