Eigenvalue-based passive intermodulation detection
A method and network node for detection of passive intermodulation (PIM) are disclosed. According to one aspect, a method includes determining values of a covariance matrix based at least in part on uplink signals received by the network node radio. The method also includes determining a plurality of uplink (UL) channel 5 eigenvalues based at least in part on an eigen-component acquisition technique. The method further includes determining a presence or absence of PIM based at least in part on a plurality of the UL channel eigenvalues.
1. A method for detection of passive intermodulation, PIM, in a network node radio based at least in part on uplink, UL, channel eigenvalues, the method comprising:
determining values of a covariance matrix based at least in part on uplink signals received by the network node radio;
determining a plurality of UL channel eigenvalues based at least in part on an eigen-component acquisition technique; and
determining a presence or absence of PIM based at least in part on the plurality of the UL channel eigenvalues.
2. The method of claim 1 , wherein the eigen-component acquisition technique is one of determining eigenvalues from the covariance matrix values or subspace tracking.
3. The method of claim 1 , wherein the plurality of UL channel eigenvalues are determined one or more of in a time-domain, a frequency-domain, using an entire channel bandwidth and a frequency resolution less than the entire channel bandwidth.
4. The method of claim 1 , wherein different UL channel eigenvalues are determined for each of a plurality of polarizations of antennas of the network node radio and the presence or absence of PIM is determined for each of the plurality of polarizations.
5. The method of claim 1 , wherein the plurality of UL channel eigenvalues are mapped against a corresponding downlink transmitted power to determine the presence or absence of PIM.
6. The method of claim 1 , wherein UL channel eigenvalues failing to meet an objective criteria are excluded from the process of determining the presence or absence of PIM.
7. The method of claim 1 , wherein a machine learning algorithm processes at least some of the UL channel eigenvalues to determine the presence or absence of PIM.
8. The method of claim 7 , wherein the machine learning algorithm is one of a regression method, a classification method or a combination of the two.
9. The method of claim 1 , wherein the plurality of UL channel eigenvalues are evaluated without aid of a machine learning algorithm to determine the presence or absence of PIM.
10. The method of claim 1 , wherein the process of determining the presence or absence of PIM is performed in both a time domain and a frequency domain to produce a time domain report of PIM presence and a frequency domain report of PIM presence.
11. The method of claim 10 , wherein the time domain report and the frequency domain report are separately weighted and combined to produce a composite report of PIM presence.
12. A network node configured to detect passive intermodulation, PIM, based at least in part on uplink, UL, channel eigenvalues, the network node comprising processing circuitry ( 48 ) configured to:
determine values of a covariance matrix based at least in part on uplink signals received by the network node radio;
determine a plurality of UL channel eigenvalues based at least in part on an eigen-component acquisition technique; and
determine a presence or absence of PIM based at least in part on a plurality of the UL channel eigenvalues.
13. The network node of claim 12 , wherein the eigen-component acquisition technique is one of determining eigenvalues from the covariance matrix values or subspace tracking.
14. The network node of claim 12 , wherein the plurality of UL channel eigenvalues are determined one of more of in a time domain, in a frequency domain, using an entire channel bandwidth and a frequency resolution less than the entire channel bandwidth.
15. The network node of claim 12 , wherein different UL channel eigenvalues are determined for each of a plurality of polarizations of antennas of the network node radio and the presence or absence of PIM is determined for each of the plurality of polarizations.
16. The network node of claim 12 , wherein the plurality of UL channel eigenvalues are mapped against a corresponding downlink transmitted power to determine the presence or absence of PIM.
17. The network node of claim 12 , wherein UL channel eigenvalues failing to meet an object criteria are excluded from the process of determining the presence or absence of PIM.
18. The network node of claim 12 , wherein a machine learning algorithm processes at least some of the UL channel eigenvalues to determine the presence or absence of PIM.
19. The network node of claim 18 , wherein the machine learning algorithm is one of a regression method, a classification method or a combination of the two.
20. The network node of claim 12 , wherein the plurality of UL channel eigenvalues are evaluated without aid of a machine learning algorithm to determine the presence or absence of PIM.
21. The network node of claim 12 , wherein the process of determining the presence or absence of PIM is performed in both the time domain and the frequency domain to produce time domain report of PIM presence and a frequency domain report of PIM presence.
22. The network node of claim 21 , wherein the time domain report and the frequency domain report are separately weighted and then combined to produce a composite report of PIM presence.