System and method for acoustic detection of emergency sirens
A system and method for detecting multi-tone sirens despite environmental noises that may be present obtains a microphone input signal, applies, in real time, a time-frequency analysis to the microphone input signal to determine a time-frequency representation, provides at least one multi-tone model that has a plurality of tone duration patterns, performs multi-tone siren detection on the time-frequency representation, the detection based on the at least one multi-tone model and factoring of doppler shifts, and generates a detection result that can be used in systems for automated vehicles.
1. A computer-implemented method comprising:
obtaining, by a computing device, a microphone input signal;
applying in real time, by the computing device, a time-frequency analysis in order to obtain a time-frequency representation of the microphone input signal;
providing, by the computing device, at least one multi-tone model that has a plurality of tone duration patterns; performing, by the computing device, multi-tone siren detection on the time-frequency representation of the signal, wherein the multi-tone detection is based on the at least one multi-tone model and factoring of doppler shifts; and
generating, by the computing device, a detection result;
wherein performing the multi-siren detection further comprises grouping tone duration patterns which have common onsets and the same fundamental frequency.
2. The method of claim 1 , wherein the performing of the multi-siren detection further comprises:
siren pattern matching by a temporal alignment of tone duration patterns for at least two different multi-tone models.
3. The method of claim 1 , wherein the multi-siren detection is performed for a set of different time stretching factors for each tone duration pattern.
4. The method of claim 1 , wherein the multi-tone siren detection is performed by an integral signal representation.
5. The method of claim 4 , wherein the integral signal representation is two-sided.
6. A computer program product residing on a non-transitory computer readable storage medium having a plurality of instructions stored thereon which, when executed across one or more processors, causes at least a portion of the one or more processors to perform operations comprising:
obtaining a microphone input signal;
applying in real time a time-frequency analysis in order to obtain a time-frequency representation of the microphone input signal;
providing at least one multi-tone model that has a plurality of tone duration patterns;
performing multi-tone siren detection on the time-frequency representation of the signal, wherein the multi-tone detection is based on the at least one multitone model and factoring of doppler shifts; and
generating a detection result;
wherein performing the multi-siren detection further comprises grouping tone duration patterns which have common onsets and the same fundamental frequency.
7. The computer program product of claim 6 , wherein the performing multi-siren detection further comprises:
siren pattern matching by a temporal alignment of tone duration patterns for at least two different multi-tone models.
8. The computer program product of claim 6 , wherein the multi-siren detection is performed for a set of different time stretching factors for each tone duration pattern.
9. The computer program product of claim 6 , wherein the multi-tone siren detection is performed by an integral signal representation.
10. The computer program product of claim 9 , wherein the integral signal representation is two-sided.
11. A computing system including one or more processors and one or more non-transitory memories storing program instructions that, when executed, perform operations comprising:
obtaining a microphone input signal;
applying in real time a time-frequency analysis in order to obtain a time-frequency representation of the microphone input signal;
providing at least one multi-tone model that has a plurality of tone duration patterns;
performing multi-tone siren detection on the time-frequency representation of the signal, wherein the multi-tone detection is based on the at least one multitone model and factoring of doppler shifts; and
generating a detection result;
wherein performing the multi-siren detection further comprises grouping tone duration patterns which have common onsets and the same fundamental frequency.
12. The computing system of claim 11 , wherein the performing of the multi-siren detection further comprises:
siren pattern matching by a temporal alignment of tone duration patterns for at least two different multi-tone models.
13. The computing system of claim 11 , wherein the multi-siren detection is performed for a set of different time stretching factors for each tone duration pattern.
14. The computing system of claim 11 , wherein the multi-tone siren detection is performed by an integral signal representation.
15. The computing system of claim 14 , wherein the integral signal representation is two-sided.