Method and apparatus for detection of radio signal
Embodiments of the present disclosure provide methods and apparatus for detection of radio signal. A method performed by a reception apparatus may comprise: obtaining (S 101 ) time domain samples of a radio signal; processing (S 102 ) the time domain samples by a plurality of match filters respectively, to generate a plurality of filtered results of the time domain samples; classifying (S 103 ) a filtered result of the plurality of filtered results by a classifier; discarding (S 104 ) a filtered result classified as being not associated to the code sequence; and determining (S 105 ) whether the code sequence is used in the radio signal, when a filtered result of the radio signal is classified as being associated to the code sequence. Some processing procedures may be avoided based on the classification result of the detected radio signal, and computational complexity may be reduced.
1. A method performed by a reception apparatus, comprising:
obtaining time domain samples of a radio signal;
processing the time domain samples by a plurality of match filters respectively, to generate a plurality of filtered results of the time domain samples, wherein the plurality of match filters corresponds to a plurality of respective code sequences;
classifying a filtered result of the plurality of filtered results by a classifier, wherein a classifying result indicates the filtered result as being associated to a code sequence in the plurality of code sequences or as not being associated to the code sequence in the plurality of code sequences;
discarding a filtered result classified as not being associated to the code sequence; and
determining whether the code sequence is used in the radio signal when a filtered result of the radio signal is classified as being associated to the code sequence.
2. The method of claim 1 , further comprising:
processing a filtered result classified as being associated to the code sequence, by peak search for the code sequence;
wherein whether the code sequence is used in the radio signal is determined, based at least on a result of the peak search.
3. The method of claim 1 , wherein the classifier comprises a machine learning (ML) classifier.
4. The method of claim 3 , wherein the ML classifier comprises a neural network (NN) classifier.
5. The method of claim 1 , further comprising:
processing each filtered result of the plurality of filtered results by non-coherent combining, before classifying.
6. The method of claim 5 ,
wherein each filtered result comprises data from a plurality of reception antennas; and
wherein the data from the plurality of reception antennas is non-coherently combined during the non-coherent combining.
7. The method of claim 6 ,
wherein data from a reception antenna of the plurality of reception antennas comprises a complex vector or matrix; and
wherein a plurality of complex vectors or matrices corresponding to the plurality of reception antennas of a filtered result is non-coherently combined to a real vector or matrix, during the non-coherent combining.
8. The method of claim 7 , wherein the real matrix is classified, during classifying.
9. The method of claim 8 , wherein a dimensionality reduction for the real vector or matrix is performed during classifying.
10. The method of claim 9 , wherein the dimensionality reduction comprises a principal component analysis (PCA).
11. The method of claim 1 , further comprising:
processing the time domain samples by fast Fourier transform (FFT) before the plurality of match filters; and
processing the plurality of filtered results by inverse fast Fourier transform (IFFT), after the plurality of match filters.
12. The method of claim 1 , further comprising:
processing the time domain samples by pre-whitening, before the plurality of match filters.
13. The method of claim 1 , further comprising:
calculating a round trip time (RTT) relating to the radio signal, when the code sequence is determined to be used in the radio signal.
14. The method of claim 1 , wherein the reception apparatus comprises a base station.
15. The method of claim 1 , wherein the plurality of code sequences is configured root sequences for a physical random access channel (PRACH).
16. A reception apparatus, comprising:
a processor; and
a memory, the memory containing instructions executable by the processor, whereby the reception apparatus is configured to:
obtain time domain samples of a radio signal;
process the time domain samples by a plurality of match filters respectively, to generate a plurality of filtered results of the time domain samples, wherein the plurality of the match filter corresponds to a plurality of respective code sequences;
classify a filtered result of the plurality of filtered results by a classifier, wherein a classifying result indicates the filtered result as being associated to a code sequence in the plurality of code sequences or as not being associated to the code sequence in the plurality of code sequences;
discard a filtered result classified as not being associated to the code sequence; and
determine whether the code sequence is used in the radio signal when a filtered result of the radio signal is classified as being associated to the code sequence.
17. The reception apparatus of claim 16 , wherein the reception apparatus is further configured to process a filtered result classified as being associated to the code sequence by peak search for the code sequence, wherein whether the code sequence is used in the radio signal is determined based at least on a result of the peak search.
18. The reception apparatus of claim 16 , wherein the classifier comprises a machine learning (ML) classifier.
19. The reception apparatus of claim 16 , wherein the reception apparatus is further configured to process each filtered result of the plurality of filtered results by non-coherent combining, before classifying.