Peak time detecting apparatus and method

• US 6,681,199 B2
• Filed: 03/02/2001
• Issued: 01/20/2004
• Est. Priority Date: 10/13/1999
• Status: Expired due to Term

• Current Assignee: Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Original Assignee: Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Inventors: Iyoda, Motomi, Imai, Katsuji, Oshima, Masuji
• Primary Examiner(s): Barlow, John
• Assistant Examiner(s): Bhat, Aditya
• Application Number:
  US09/796,802
• Publication Number:
  US 20020152054A1
• Time in Patent Office: 1,054 Days
• Field of Search: 702/189, 702/190, 702/1, 702/14, 702/16, 702/17, 702/191, 702/6, 702/35, 702/66, 702/85, 600/410, 600/431, 600/437, 600/500, 600/509, 382/236, 382/232, 382/100, 382/103, 382/290, 382/296, 367/72, 367/53, 367/63, 367/39, 367/31
• US Class Current: 702/189
• CPC Class Codes: B60R 21/013 including means for detecting collisions, impending collisions or roll-over
• CPC Class Codes: B60R 21/01332 by frequency or waveform analysis

Abstract

A product-sum operation portion for performing a product-sum operation (wavelet transformation) with respect to an input time-series signal by using as a base of integral a complex function in which the imaginary number portion is π/2 shifted in phase from the real number portion, a phase calculation portion for calculating a phase θ from the ratio between the real number portion and the imaginary number portion of a result of the product-sum operation, a peak time detection portion for detecting a time point at which the calculated phase θ changes from 2π to zero, as a peak time, are provided. Since the wavelet transformation is performed by using a basic wavelet function that is localized in terms of time and frequency, a peak time can be promptly detected. Furthermore, since a differential operation is not employed but the product-sum operation is performed, false detection caused by noises or the like can be prevented.

Claims

1. A peak time detecting apparatus for detecting a peak time of a time-series signal by using a wavelet transformation, comprising:

signal input means for inputting the time-series signal;

product-sum operation means for performing a product-sum operation with respect to the time-series signal inputted, by using a predetermined complex function as an integral base;

phase calculation means for calculating a phase based on a real number portion and an imaginary number portion of a result of the product-sum operation; and

peak time determination means for determining a peak time of the time-series signal based on the phase calculated,

wherein the product-sum operation means is means that uses a Gabor function as the predetermined complex function.

2. A peak time detecting apparatus according to claim 1, wherein the peak time determination means is means for determining, as the peak time, a time point at which the phase calculated by the phase calculation means changes from 2π

to zero.

3. A peak time detecting apparatus for detecting a peak time of a time-series signal by using a wavelet transformation, comprising:

signal input means for inputting the time-series signal;

product-sum operation means for performing a product-sum operation with respect to the time-series signal inputted, by using a predetermined complex function as an integral base;

phase calculation means for calculating a phase based on a real number portion and an imaginary number portion of a result of the product-sum operation; and

peak time determination means for determining a peak time of the time-series signal based on the phase calculated,

wherein the product-sum operation means is means that uses, as the predetermined complex function, a function that includes a real number portion having a localized waveform and an imaginary number portion having a localized waveform that is delayed by π

/2 in phase from the real number portion.

4. A peak time detecting apparatus according to claim 3, wherein the peak time determination means is means for determining, as the peak time, a time point at which the phase calculated by the phase calculation means changes from 2π

to zero.

5. A peak time detecting apparatus for detecting a peak time of a time-series signal by using a wavelet transformation, comprising:

signal input means for inputting the time-series signal;

product-sum operation means for performing a product-sum operation with respect to the time-series signal inputted, by using a predetermined complex function as an integral base;

phase calculation means for calculating a phase based on a real number portion and an imaginary number portion of a result of the product-sum operation; and

peak time determination means for determining a peak time of the time-series signal based on the phase calculated; and

validity determination means for determining a validity of a result of determination made by the peak time determination means.

6. A peak time detecting apparatus according to claim 5,

wherein the phase calculation means is means for calculating a phase regarding a result of the product-sum operation with respect to a peak time detection-purposed transformation frequency and a phase regarding a result of the product-sum operation with respect to a validity determination-purposed transformation frequency that is higher than the peak time detection-purposed transformation frequency, and

wherein the validity determination means is means for determining the validity based on the phase calculated regarding the result with respect to the validity determination-purposed transformation frequency.

7. A peak time detecting apparatus according to claim 6, wherein the validity determination means is means for determining that a valid determination is made if a peak time is determined within a predetermined time before and after a time point at which the phase calculated regarding the result with respect to the validity determination-purposed transformation frequency changes from 2π

to zero.

8. A peak time detecting apparatus according to claim 6, wherein the validity determination-purposed transformation frequency is 1.0 to 2.0 times the peak time detection-purposed transformation frequency.

9. A peak time detecting apparatus for detecting a peak time of a time-series signal by using a wavelet transformation, comprising:

signal input means for inputting the time-series signal;

product-sum operation means for performing a product-sum operation with respect to the time-series signal inputted, by using a predetermined complex function as an integral base;

phase calculation means for calculating a phase based on a real number portion and an imaginary number portion of a result of the product-sum operation; and

peak time determination means for determining a peak time of the time-series signal based on the phase calculated,

wherein the time-series signal is a signal formed by removing a high-frequency component from a signal detected by deceleration detection means provided in a vehicle, and

wherein the phase calculation means is means for calculating a phase regarding a result of the product-sum operation with respect to a predetermined transformation frequency within a range of 100 to 150 Hz.

10. A peak time detecting method for detecting a peak time of an input time-series signal by using a wavelet transformation, comprising:

performing a product-sum operation with respect to the input time-series signal by using a predetermined complex function as an integral base;

calculating a phase based on a real number portion and an imaginary number portion of a result of the product-sum operation; and

detecting a peak time based on the phase calculated,

wherein a Gabor function is used as the predetermined complex function.

11. A peak time detecting method according to claim 10, wherein the peak time detecting step is a step of detecting, as the peak time, a time point at which the phase calculated changes from 2π

to zero.

12. A peak time detecting method for detecting a peak time of an input time-series signal by using a wavelet transformation, comprising:

performing a product-sum operation with respect to the input time-series signal by using a predetermined complex function as an integral base;

calculating a phase based on a real number portion and an imaginary number portion of a result of the product-sum operation; and

detecting a peak time based on the chase calculated,

wherein a function that includes a real number portion having a localized waveform and an imaginary number portion having a localized waveform that is delayed by π

/2 in phase from the real number portion is used as the predetermined complex function.

13. A peak time detecting method according to claim 12, wherein the peak time detecting step is a step of detecting, as the peak time, a time point at which the phase calculated changes from 2π

to zero.

14. A peak time detecting method for detecting a peak time of an input time-series signal by using a wavelet transformation, comprising:

performing a product-sum operation with respect to the input time-series signal by using a predetermined complex function as an integral base;

calculating a phase based on a real number portion and an imaginary number portion of a result of the product-sum operation;

detecting a peak time based on the phase calculated; and

determining a validity of a result of determination made by the peak time detecting step.

15. A peak time detecting method according to claim 14,

wherein the phase calculating step is a step of calculating a phase regarding a result of the product-sum operation with respect to a peak time detection-purposed transformation frequency and a phase regarding a result of the product-sum operation with respect to a validity determination-purposed transformation frequency that is higher than the peak time detection-purposed transformation frequency, and

wherein the validity determining step determines the validity based on the phase calculated regarding the result with respect to the validity determination-purposed transformation frequency.

16. A peak time detecting method according to claim 15, wherein the validity determining step determines that a valid determination is made if a peak time is determined within a predetermined time before and after a time point at which the phase calculated regarding the result with respect to the validity determination-purposed transformation frequency changes from 2π

to zero.

17. A peak time detecting method according to claim 15, wherein the validity determination-purposed transformation frequency is 1.0 to 2.0 times the peak time detection-purposed transformation frequency.

18. A peak time detecting method for detecting a peak time of an input time-series signal by using a wavelet transformation, comprising:

performing a product-sum operation with respect to the input time-series signal by using a predetermined complex function as an integral base;

calculating a phase based on a real number portion and an imaginary number portion of a result of the product-sum operation; and

detecting a peak time based on the phase calculated,

wherein the time-series signal is a signal formed by removing a high-frequency component from a signal detected by deceleration detection means provided in a vehicle, and

wherein the phase calculating step calculates a phase regarding a result of the product-sum operation with respect to a predetermined transformation frequency within a range of 100 to 150 Hz.