Period estimation method, device and server of signal

Abstract

The present disclosure provides a signal cycle estimation method, device and server, relating to signal cycle estimation technology, comprising: obtaining target object data of a sensor, and determining a target signal sequence corresponding to the target object according to the target object data; determining an average amplitude difference function and an amplitude difference sum function of the target signal sequence according to the target signal sequence; determining extreme value points and extreme points of the amplitude difference sum function within a preset cycle range, and determining a decision threshold according to a function value at the extreme points and a preset decision threshold proportion constant; determining a target extreme value point satisfying a preset condition from the extreme value points according to the decision threshold and the extreme points, and determining the target extreme value point as a cycle estimation value of the target signal sequence. The present scheme can utilize information between all signal cycles by using the amplitude difference sum function, thereby reducing the influence of noise to a greater extent, and further improving the accuracy of cycle estimation.

Description

Technical Field

[0001] This disclosure relates to signal period estimation techniques, and more particularly to a signal period estimation method, apparatus, and server. Background Technology

[0002] Periodic signals are widely present in various observational data. Examples include speech signals, physiological signals such as respiration, heart rate, and blood oxygenation in medical monitoring, micro-Doppler and radar cross-section (RCS) of oscillating objects in radar echoes, and signals generated by the periodic rotation and vibration of machinery. Although these phenomena have different mechanisms, the signals they produce are all periodically repetitive and have unknown waveforms (a "waveform" refers to the waveform of a signal within one repetition period; the observed signal contains multiple waveforms that repeat at a certain period). Accurately and quickly estimating the signal period from the digital sequence acquired by sensors is of great significance.

[0003] In existing technologies, methods such as the autocorrelation function (ACF) or the average magnitude difference function (AMDF) are used to estimate the period of an unknown signal.

[0004] However, the accuracy of the signal period estimation using the above method needs to be improved. Summary of the Invention

[0005] This disclosure provides a method, apparatus, and server for estimating the period of a signal, thereby improving the accuracy of signal period estimation.

[0006] According to a first aspect of this disclosure, a method for estimating the period of a signal is provided, comprising:

[0007] Acquire target object data from the sensor, and determine the target signal sequence corresponding to the target object based on the target object data;

[0008] Based on the target signal sequence, determine the average amplitude difference function of the target signal sequence; and based on the average amplitude difference function of the target signal sequence, determine the amplitude difference sum function of the target signal sequence.

[0009] Determine the extreme points and maximum / minimum points of the amplitude difference and function within a preset period range, and determine the decision threshold based on the function value at the maximum / minimum point and a preset decision threshold proportional constant.

[0010] Based on the decision threshold and the extreme point, a target extreme point that meets the preset conditions is determined among the extreme points, and the target extreme point is determined as the period estimate of the target signal sequence; the period estimate is used to assist in the identification of the target object.

[0011] According to a second aspect of this disclosure, a signal period estimation apparatus is provided, comprising:

[0012] The acquisition unit is used to acquire target object data from the sensor and determine the target signal sequence corresponding to the target object based on the target object data.

[0013] The function determination unit is used to determine the average amplitude difference function of the target signal sequence based on the target signal sequence; and to determine the amplitude difference sum function of the target signal sequence based on the average amplitude difference function of the target signal sequence.

[0014] The threshold determination unit is used to determine the extreme points and maximum / minimum points of the amplitude difference and function within a preset period range, and to determine the decision threshold based on the function value at the maximum / minimum point and a preset decision threshold proportional constant.

[0015] The period estimation unit is used to determine a target extreme point that meets a preset condition among the extreme points based on the decision threshold and the extreme point, and to determine the target extreme point as the period estimate of the target signal sequence; the period estimate is used to assist in the identification of the target object.

[0016] According to a third aspect of this disclosure, a server is provided, including a memory and a processor; wherein,

[0017] The memory is used to store computer programs;

[0018] The processor is configured to read a computer program stored in the memory and execute the signal period estimation method as described in the first aspect according to the computer program in the memory.

[0019] According to a fourth aspect of this disclosure, a computer-readable storage medium is provided, wherein computer-executable instructions are stored therein, which, when executed by a processor, implement the period estimation method for a signal as described in the first aspect.

[0020] According to a fifth aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the period estimation method for a signal as described in the first aspect.

[0021] The signal period estimation method, apparatus, and server disclosed herein include: acquiring target object data from a sensor and determining the target signal sequence corresponding to the target object based on the target object data; determining the average amplitude difference function of the target signal sequence based on the target signal sequence; determining the amplitude difference sum function of the target signal sequence based on the average amplitude difference function of the target signal sequence; determining the extreme points and maximum / minimum points of the amplitude difference sum function within a preset period range, and determining a decision threshold based on the function value at the extreme / minimum points and a preset decision threshold proportional constant; determining a target extreme point that meets a preset condition among the extreme points based on the decision threshold and the extreme / minimum points, and determining the target extreme point as the period estimate of the target signal sequence; the period estimate is used to assist in identifying the target object. The signal period estimation method, apparatus, and server provided in this solution utilize information between all signal periods by employing the amplitude difference sum function method, which greatly reduces the impact of noise and thus improves the accuracy of period estimation. Furthermore, using a preset period range and a decision threshold method to estimate the period further improves the accuracy of period estimation. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A schematic flowchart illustrating a signal period estimation method as an exemplary embodiment of this disclosure;

[0024] Figure 2 A flowchart illustrating a method for estimating the period of a signal, as shown in another exemplary embodiment of this disclosure;

[0025] Figure 3 A schematic diagram illustrating the magnitude difference and function as an exemplary embodiment of this disclosure;

[0026] Figure 4 A structural diagram of a signal period estimation apparatus shown in an exemplary embodiment of this disclosure;

[0027] Figure 5 This is a structural diagram of a server shown as an exemplary embodiment of the present disclosure. Detailed Implementation

[0028] Periodic signals are widely present in various observational data. Examples include speech signals, physiological signals such as respiration, heart rate, and blood oxygenation in medical monitoring, micro-Doppler and radar cross-section (RCS) of oscillating objects in radar echoes, and signals generated by the periodic rotation and vibration of machinery. Although these phenomena have different mechanisms, the signals they produce are all periodically repetitive and have unknown waveforms (a "waveform" refers to the waveform of a signal within a repetition period; an observed signal contains multiple waveforms that repeat at a certain period). Accurately and quickly estimating the signal period from the digital sequence acquired by sensors is of great significance. Specifically, the estimated signal period can be used to assist in identifying the target object corresponding to that signal.

[0029] In existing technologies, methods such as the autocorrelation function (ACF) or the average magnitude difference function (AMDF) are used to estimate the period of an unknown signal. Specifically, the autocorrelation function or the average magnitude difference function can be calculated using the signal sequence corresponding to the target object, and then the period can be estimated based on the extreme points and maximum / minimum points of the autocorrelation function or the average magnitude difference function.

[0030] However, the above method only utilizes information between adjacent signal periods, resulting in significant noise interference and consequently low accuracy in signal period estimation. To address this technical problem, the solution provided in this disclosure employs the Sum Average Magnitude Difference Function (Sum AMDF) method, which utilizes information between all signal periods, significantly reducing the impact of noise and thus improving the accuracy of period estimation. Furthermore, using a preset period range and a decision threshold to estimate the period further enhances the accuracy of period estimation.

[0031] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0032] The technical solutions of this disclosure and how they solve the aforementioned technical problems will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of this disclosure will now be described with reference to the accompanying drawings.

[0033] Figure 1 This is a schematic flowchart illustrating a signal period estimation method as an exemplary embodiment of the present disclosure.

[0034] like Figure 1 As shown, the signal period estimation method provided in this embodiment includes:

[0035] Step 101: Obtain the target object data from the sensor, and determine the target signal sequence corresponding to the target object based on the target object data.

[0036] The method provided in this disclosure can be executed by a server with computing power.

[0037] The server can acquire target data from the sensor and determine the target signal sequence corresponding to the target based on the target data.

[0038] The sensor can be a radar sensor. Specifically, it can acquire target data from the radar sensor and extract a target signal sequence of length K from the target data.

[0039] The target signal sequence can be a periodic discrete sequence.

[0040] Specifically, the target signal sequence may include a periodic signal corresponding to the target object and noise. The noise may be white noise.

[0041] Step 102: Determine the average amplitude difference function of the target signal sequence based on the target signal sequence; and determine the amplitude difference sum function of the target signal sequence based on the average amplitude difference function of the target signal sequence.

[0042] Specifically, the formula for the average amplitude difference function can be pre-set based on the actual situation and the method of the average amplitude difference function. To reduce the influence of noise, the difference term of the sampling points themselves can be removed from the summation term of the average amplitude difference function, thus obtaining the amplitude difference sum function of the target signal sequence. Therefore, the formula for the amplitude difference sum function can be pre-set based on the actual situation and the method of the amplitude difference sum function.

[0043] Then, based on the target signal sequence and the pre-set formula for the average amplitude difference function, the average amplitude difference function of the target signal sequence can be calculated. Then, using the average amplitude difference function of the target signal sequence and the pre-set formula for the amplitude difference sum function, the amplitude difference sum function of the target signal sequence can be calculated.

[0044] Step 103: Determine the extreme points and maximum / minimum points of the amplitude difference and function within a preset period range, and determine the decision threshold based on the function value at the maximum / minimum point and the preset decision threshold proportional constant.

[0045] The preset decision threshold ratio constant is a value that is set in advance based on the actual situation.

[0046] The preset period range is the period range of the target signal sequence that is pre-set according to the actual situation.

[0047] Specifically, based on the amplitude difference and function of the target signal sequence, all extreme points and maximum / minimum points of the amplitude difference and function within a preset period range can be determined. There may be multiple extreme points.

[0048] Then, based on the amplitude difference and function of the target signal sequence, as well as the extreme points, the function value at the extreme points can be determined. This function value is then multiplied by a preset decision threshold proportional constant to obtain the decision threshold.

[0049] Step 104: Based on the decision threshold and the extreme point, determine the target extreme point that meets the preset conditions among the extreme points, and determine the target extreme point as the period estimate of the target signal sequence; the period estimate is used to assist in the identification of the target object.

[0050] Specifically, the extreme point may be the period of the target signal sequence or a multiple of the period.

[0051] Specifically, the absolute value of the amplitude difference corresponding to the period estimate of the target signal sequence and the function value in the function can be greater than the decision threshold.

[0052] The preset conditions are the criteria for determining the target extreme value point, set in advance based on the actual situation. Specifically, these preset conditions can be set based on the extreme value point and the decision threshold.

[0053] Specifically, based on the extreme points, decision thresholds, and preset conditions, a target extreme point that meets the preset conditions can be determined among the extreme points, and the target extreme point can be determined as the period estimate of the target signal sequence.

[0054] Furthermore, the period estimate of the target signal sequence can be used to assist in identifying the target object corresponding to the target signal sequence.

[0055] The signal period estimation method disclosed herein includes: acquiring target object data from a sensor, and determining the target signal sequence corresponding to the target object based on the target object data; determining the average amplitude difference function of the target signal sequence based on the target signal sequence; determining the amplitude difference sum function of the target signal sequence based on the average amplitude difference function of the target signal sequence; determining the extreme points and maximum / minimum points of the amplitude difference sum function within a preset period range, and determining a decision threshold based on the function value at the maximum / minimum point and a preset decision threshold proportional constant; determining the target extreme point that satisfies a preset condition among the extreme points based on the decision threshold and the extreme / minimum point, and determining the target extreme point as the period estimate of the target signal sequence; the period estimate is used to assist in identifying the target object. The signal period estimation method provided in this solution utilizes information between all signal periods by employing the amplitude difference sum function method, thus mitigating the influence of noise to a greater extent and improving the accuracy of period estimation. Furthermore, using a preset period range and a decision threshold method to estimate the period further improves the accuracy of period estimation.

[0056] Figure 2 This is a schematic flowchart illustrating a method for estimating the period of a signal, which is another exemplary embodiment of this disclosure.

[0057] like Figure 2 As shown, the signal period estimation method provided in this embodiment includes:

[0058] Step 201: Obtain the target object data from the sensor, and determine the target signal sequence corresponding to the target object based on the target object data.

[0059] Specifically, assume that the target signal sequence consists of P s The waveform s at each sampling point is repeated for M periods, and there are a total of K sampling points. The target signal sequence contains white noise w. Then the target signal sequence can be represented as follows:

[0060]

[0061] Where x[k] represents the target signal sequence; k represents the sampling time; l is a natural number; s[k] represents the periodic signal corresponding to the target object included in the target signal sequence; M is a positive integer representing the number of repetition periods of s[k]; P s represents the actual number of sampling points (i.e., the actual period length) contained in s[k] in one repetition period; w[k] represents the white noise contained in the target signal sequence.

[0062] Where s[k] can represent an unknown waveform. Then the unknown parameters in the target signal sequence are θ = [s[1], ... s[P] s ], P s ], total Ps +1 parameter to be estimated.

[0063] Step 202: Interpolate the target signal sequence according to the preset interpolation factor, and determine the average amplitude difference function of the target signal sequence according to the interpolated target signal sequence; determine the amplitude difference sum function of the target signal sequence according to the average amplitude difference function of the target signal sequence.

[0064] Specifically, let the variance of white noise be σ. 2 The log-likelihood function of the target signal sequence is as follows:

[0065]

[0066] Where, ln(p(θ)) represents the log-likelihood function of the target signal sequence; p(θ) represents the maximum likelihood function of the target signal sequence; K represents the length of the target signal sequence; k represents the sampling time; x[k] represents the target signal sequence; σ 2 The variance of the white noise included in the target signal sequence is represented by ; l is a positive integer; s[k] represents the periodic signal included in the target signal sequence; P s Let represent the actual number of sampling points (i.e., the period length) contained in s[k] within one repetition period; P represents the number of sampling points contained in s[k] within one repetition period; since P s Since it is unknown, it can be represented by P. s ; This indicates that the number inside the parentheses is rounded down to the nearest integer.

[0067] Then, according to the likelihood function, the periodic signal s[k] and the period length P are... s The formula for maximum likelihood estimation is as follows:

[0068]

[0069]

[0070] in, The estimated signal represents the periodic signal included in the target signal sequence; l is a natural number; P represents the number of sampling points contained in s[k] in one repetition period; K represents the length of the target signal sequence; k represents the sampling time. This represents an estimate of the number of sampling points contained in s[k] within one repetition period; This indicates that the number inside the parentheses is rounded down to the nearest integer.

[0071] Specifically, the above formula essentially measures the interval P. s The similarity between the signals can be expressed as the sum of the squared amplitude differences of the signals, as shown below:

[0072]

[0073] in, The value represents the estimated number of sampling points in one repetition period of the periodic signal s[k] included in the target signal sequence; l, m, and n are all natural numbers; k represents the sampling time; K represents the length of the target signal sequence; P represents the number of sampling points in one repetition period of s[k]; x[k] represents the target signal sequence. This indicates that the number inside the parentheses is rounded down to the nearest integer.

[0074] However, directly searching based on the above formula involves a large amount of computation. On the other hand, methods based on summing power spectrum lines or averaging autocorrelation functions can be used to simplify the calculation, but these methods are still relatively complex.

[0075] Therefore, the computational load can be reduced by minimizing signal multiplication. The above formula is based on the 2-norm between any two sampling points at any interval P. By changing the metric to the 1-norm, i.e., the absolute value of the difference, signal multiplication can be avoided. This leads to the following formula:

[0076]

[0077]

[0078] in, This represents the estimated number of sampling points in one repetition period of the periodic signal s[k] included in the target signal sequence; m and n are both natural numbers; k represents the sampling time; K represents the length of the target signal sequence; P represents the number of sampling points in one repetition period of s[k]; x[k] represents the target signal sequence. This indicates that the number inside the parentheses is rounded down to the nearest integer.

[0079] Specifically, the above estimation process is based on the classic maximum likelihood trajectory. This estimation method cannot distinguish between the signal's own period T and multiples of the period kT, because the likelihood function of the target signal sequence has extrema at these locations, as does the amplitude difference function.

[0080] Meanwhile, when noise is present, the estimation results tend to overestimate the period. This is because noise has no period, or rather, the period of noise is infinite, causing the period of the noise-contaminated signal to also tend to be infinite. In the above formula, this manifests as, as P increases, the term in the amplitude difference term that subtracts the sampled value from itself, i.e., |x[k+mP]-x[k+nP]|, where m=n, increases more and more. The summation result gradually tends to 0, hindering the search for the minimum value.

[0081] The impact of noise can be mitigated by removing the difference terms from the sampling points themselves in the summation terms. This reduces the number of summation terms from N. 2 The number of elements has been reduced to N(N-1), and a weighted value is required accordingly. Replace with This provides an estimate of the period length. And define the amplitude difference and function a1(P) as shown in the following equation:

[0082]

[0083] in, The value of represents the estimated number of sampling points in one repetition period of the periodic signal s[k] included in the target signal sequence; m and n are natural numbers; k represents the sampling time; K represents the length of the target signal sequence; P represents the number of sampling points in one repetition period of s[k]; x[k] represents the target signal sequence; a1(P) represents the amplitude difference function of the target signal sequence. This indicates that the number inside the parentheses is rounded down to the nearest integer.

[0084] In practical estimation, the period of the signal is not necessarily an integer multiple of the sampling interval, so improving the accuracy of period estimation may be important. If the method given above is used directly, the estimation accuracy is limited by the sampling rate of the target signal sequence x[k], which is often insufficient.

[0085] An intuitive approach is to upsample the signal to reduce the sampling interval, but this method consumes a lot of computational power, defeating the purpose of simplifying the algorithm using absolute value amplitude. Similarly, any other complex interpolation algorithm suffers from the same problem.

[0086] Therefore, simple 0th-order spline interpolation, i.e., rounding, can be used. This interpolation aligns with the idea of ​​the amplitude difference sum function, improving accuracy without adding new complex operations involving multiplication and division. For direct interpolation of the target signal sequence x[k] in the time domain, the amplitude difference sum function a1(P) for non-integer periods can be expressed as follows:

[0087]

[0088] Where [] represents taking the nearest integer within the brackets; a1(P) represents the amplitude difference function; m and n are both natural numbers; k represents the sampling time; K represents the length of the target signal sequence; P represents the number of sampling points contained in one repetition period of the periodic signal s[k] included in the target signal sequence; x[k] represents the target signal sequence; This indicates that the number inside the parentheses is rounded down to the nearest integer.

[0089] Specifically, after interpolation, the original non-integer position x[k+mP] is replaced by the nearest sampling point x[k+[mP]]. When the interpolation is very dense or the possible period range is very large, there are many possible P values, making direct calculation of the above formula difficult and computationally intensive, involving many repetitive operations. Therefore, the summation order of the above formula can be further changed, as shown in the following equation:

[0090]

[0091] Where [] indicates taking the nearest integer within the brackets; The parentheses indicate that the number inside is rounded down; a1(P) represents the amplitude difference function; k represents the sampling time; K represents the length of the target signal sequence; P represents the number of sampling points contained in one repetition period of waveform s; x[k] represents the target signal sequence; l is a positive integer. In the above formula, This is a typical AMDF where the period includes the values ​​corresponding to [lP] sampling points. a1(P) can be approximated as the summation of the positions corresponding to [lP] selected in the AMDF. Since it is not necessary to sum each interpolated period separately, the computational load can be reduced when there are many values ​​of P.

[0092] In one possible implementation, the amplitude difference and function can be pre-set as follows:

[0093]

[0094] Where a1(P) represents the amplitude difference function of the interpolated target signal sequence; P represents the number of sampling points contained in one repetition period of the target signal sequence; K represents the length of the target signal sequence; a amdf ([lP]) represents the average amplitude difference function of the interpolated target signal sequence; l is a positive integer; [] indicates taking the nearest integer within the brackets; This indicates that the number inside the parentheses is rounded down to the nearest integer.

[0095] In one possible implementation, the average amplitude difference function can be pre-defined as follows:

[0096]

[0097] Among them, a amdf (p) represents the average amplitude difference function of the target signal sequence; p is a positive integer; K represents the length of the target signal sequence; x[k] represents the target signal sequence.

[0098] Step 203: Determine the extreme points and maximum / minimum points of the amplitude difference and function within a preset period range, and determine the decision threshold based on the function value at the maximum / minimum point and the preset decision threshold proportional constant.

[0099] The preset decision threshold ratio constant is a value that is set in advance based on the actual situation.

[0100] In one possible implementation, the preset decision threshold ratio constant has a range of (0.5, 1).

[0101] Specifically, the range of the decision threshold proportional constant can be preset to (0.5, 1) according to the actual situation. For example, the preset decision threshold proportional constant can be set to 0.75.

[0102] The preset period range is the period range of the target signal sequence that is pre-set according to the actual situation.

[0103] Specifically, based on the amplitude difference and function of the target signal sequence, all extreme points and maximum / minimum points of the amplitude difference and function within a preset period range can be determined. There may be multiple extreme points.

[0104] Then, based on the amplitude difference and function of the target signal sequence, as well as the extreme points, the function value at the extreme points can be determined. This function value is then multiplied by a preset decision threshold proportional constant to obtain the decision threshold.

[0105] Step 204: Divide the extreme point by n to obtain the first intermediate value; where n is a natural number greater than 0; determine the first intermediate value within the preset period range as the second intermediate value; determine the first extreme point closest to the second intermediate value among the extreme points.

[0106] Specifically, the amplitude difference and function a1(P) at the actual period length P s The period of a waveform has almost identical minimum values ​​at its integer multiples, so other methods are still needed to ensure the accuracy of the estimation. This is unavoidable for any period estimation method for an unknown waveform. One approach is to utilize prior knowledge, giving a range of possible period lengths for the signal, and ensuring that this range is small enough that there are no multiples of the possible period lengths. Alternatively, a period determination method can be added.

[0107] If the possible period range is unknown, or if the possible period range varies greatly, it is necessary to accurately estimate the period. Therefore, this embodiment provides a simple method for determining the period based on the amplitude difference and the amplitude of the function a1(P).

[0108] Assuming that the samples of the signal are uncorrelated within the period, the amplitude difference and the function a1(P) will have similar extreme values ​​at the actual period and integer multiples of the period. At positions such as 0.5 times and 1.5 times the actual period, there are secondary peaks with amplitudes approximately half that of the first peak.

[0109] Based on the above conditions, a judgment method can be designed. Remove the mean of a1(P), and using a preset period range, the range of the number of sampling points included in one repetition period corresponding to this preset period range can be expressed as [P]. min P max ], in [P min P max Search within the range for extreme points of the magnitude difference and function a1(P), and find the minimum value a. min and the position P corresponding to its minimum value multi P multi It could be the actual period or a multiple of the period. P multi That is, the extreme point, a min This refers to the magnitude difference corresponding to the extreme points and the function value within the function. In other words, the true period is all the points falling within [P]. min P max P within ] multi / n, n = one of 1, 2, 3...

[0110] Specifically, P multi / n, where n = 1, 2, 3... is the first intermediate value. The first intermediate value within a preset period range can be determined as the second intermediate value. Then, the difference between the extreme points and the second intermediate value can be calculated to determine the first extreme point that is closest to the second intermediate value.

[0111] Specifically, among all extreme points, find the distance from P. multi / n, n = 1, 2, 3... The closest extreme point. If a position is the true period or an integer multiple of the period, then its extreme value should be close to a. min If a position is not a true period or an integer multiple of the period, then it should be greater than 0.5a. min .

[0112] Step 205: If the absolute value of the first extreme point is greater than the absolute value of the decision threshold, then the first extreme point is determined as the third intermediate value; the smallest value among the third intermediate values ​​is selected as the target extreme point. The target extreme point is then determined as the period estimate of the target signal sequence; the period estimate is used to assist in the identification of the target object.

[0113] Specifically, the absolute value of each first extreme point is first compared with the absolute value of the decision threshold. If the absolute value of the first extreme point is greater than the absolute value of the decision threshold, then the first extreme point is determined as the third intermediate value.

[0114] Then, the smallest value among the third intermediate values ​​is selected as the target extreme point. This target extreme point is then used as the period estimate of the target signal sequence. Furthermore, this period estimate of the target signal can be used to assist in the identification of the target object.

[0115] Specifically, it is possible to design based on a min The judgment threshold. The threshold value can be set according to the actual situation, between a... min and 0.5a min The specific value should be selected based on the actual situation. multi Among the peak values ​​corresponding to / n, n = 1, 2, 3..., the smallest P that makes the peak value less than the threshold. multi / n is determined to be the true period.

[0116] Actual effect as Figure 3 As shown, a1(P) in [P min P max There are multiple peaks within the range, with the minimum value around 6.6s. The decision threshold can be set to 0.75a. min Based on the decision threshold, only the peak value around 3.3s represents the true period. Peak values ​​at positions such as 1.7s are secondary peaks that are multiples of the period. Figure 3 The horizontal axis represents time, the vertical axis represents amplitude, the solid line represents the amplitude difference and sum of functions a1(P), i.e., Sum AMDF, and the dashed line represents the value of the decision threshold.

[0117] Figure 4 This is a structural diagram of a signal period estimation apparatus shown as an exemplary embodiment of the present disclosure.

[0118] like Figure 4 As shown, the signal period estimation device 400 provided in this disclosure includes:

[0119] The acquisition unit 410 is used to acquire target object data from the sensor and determine the target signal sequence corresponding to the target object based on the target object data.

[0120] The function determination unit 420 is used to determine the average amplitude difference function of the target signal sequence based on the target signal sequence; and to determine the amplitude difference sum function of the target signal sequence based on the average amplitude difference function of the target signal sequence.

[0121] The threshold determination unit 430 is used to determine the extreme points and maximum points of the amplitude difference and function within a preset period range, and to determine the decision threshold based on the function value at the maximum point and the preset decision threshold proportional constant.

[0122] The period estimation unit 440 is used to determine the target extreme point that meets the preset conditions among the extreme points according to the decision threshold and the extreme point, and to determine the target extreme point as the period estimate of the target signal sequence; the period estimate is used to assist in the identification of the target object.

[0123] The function determination unit 420 is specifically used to interpolate the target signal sequence according to a preset interpolation multiple, and determine the average amplitude difference function of the target signal sequence according to the interpolated target signal sequence; and determine the amplitude difference sum function of the target signal sequence according to the average amplitude difference function of the target signal sequence.

[0124] In one feasible implementation, the magnitude difference sum function is:

[0125]

[0126] Where a1(P) represents the amplitude difference function of the interpolated target signal sequence; P represents the number of sampling points contained in one repetition period of the target signal sequence; K represents the length of the target signal sequence; a amdf ([lP]) represents the average amplitude difference function of the interpolated target signal sequence; l is a positive integer; [] indicates taking the nearest integer within the brackets; This indicates that the number inside the parentheses is rounded down to the nearest integer.

[0127] In one feasible implementation, the average amplitude difference function is:

[0128]

[0129] Among them, a amdf (p) represents the average amplitude difference function of the target signal sequence; p is a positive integer; K represents the length of the target signal sequence; x[k] represents the target signal sequence.

[0130] The period estimation unit 440 is specifically used to divide the extreme point by n to obtain the first intermediate value; where n is a natural number greater than 0.

[0131] The first intermediate value within the preset period range will be determined as the second intermediate value;

[0132] Among the extreme points, identify the first extreme point that is closest to the second intermediate value;

[0133] If the absolute value of the first extreme point is determined to be greater than the absolute value of the decision threshold, then the first extreme point is determined as the third intermediate value;

[0134] Select the smallest value among the third intermediate values ​​and determine it as the target extreme point.

[0135] In one possible implementation, the preset decision threshold ratio constant has a range of (0.5, 1).

[0136] Figure 5 This is a structural diagram of a server shown as an exemplary embodiment of the present disclosure.

[0137] like Figure 5As shown, the server provided in this embodiment includes:

[0138] Memory 501;

[0139] Processor 502; and

[0140] Computer programs;

[0141] The computer program is stored in memory 501 and configured to be executed by processor 502 to implement the period estimation method for any of the signals described above.

[0142] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement any of the signal period estimation methods described above.

[0143] This embodiment also provides a computer program product, including a computer program that, when executed by a processor, implements the period estimation method for any of the above-described signals.

[0144] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0145] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for estimating the period of a signal, characterized in that, include: Acquire target object data from the sensor, and determine the target signal sequence corresponding to the target object based on the target object data; Based on the target signal sequence, determine the average amplitude difference function of the target signal sequence; and based on the average amplitude difference function of the target signal sequence, determine the amplitude difference sum function of the target signal sequence. Determine the extreme points and maximum / minimum points of the amplitude difference and function within a preset period range, and determine the decision threshold based on the function value at the maximum / minimum point and a preset decision threshold proportional constant. Based on the decision threshold and the extreme point, a target extreme point that meets the preset conditions is determined among the extreme points, and the target extreme point is determined as the period estimate of the target signal sequence; the period estimate is used to assist in the identification of the target object; The amplitude difference function is: in, The sum of amplitude differences in the interpolated target signal sequence is represented by the function; P represents the number of sampling points in one repetition period of the target signal sequence; K represents the length of the target signal sequence. The function representing the average amplitude difference of the interpolated target signal sequence; It is a positive integer; [] indicates taking the nearest integer within the brackets; This indicates that the number inside the parentheses is rounded down to the nearest integer.

2. The method according to claim 1, characterized in that, The step of determining the average amplitude difference function of the target signal sequence based on the target signal sequence, and determining the amplitude difference sum function of the target signal sequence based on the average amplitude difference function of the target signal sequence, includes: The target signal sequence is interpolated according to a preset interpolation factor, and the average amplitude difference function of the target signal sequence is determined based on the interpolated target signal sequence; the amplitude difference sum function of the target signal sequence is determined based on the average amplitude difference function of the target signal sequence.

3. The method according to claim 1, characterized in that, The average amplitude difference function is: in, This represents the average amplitude difference function of the target signal sequence; p is a positive integer; K represents the length of the target signal sequence; This represents the target signal sequence.

4. The method according to any one of claims 1-3, characterized in that, The step of determining a target extreme point that satisfies a preset condition among the extreme points based on the decision threshold and the extreme point includes: Divide the extreme point by n to obtain the first intermediate value; where n is a natural number greater than 0. The first intermediate value within the preset period range is determined as the second intermediate value; Among the extreme points, determine the first extreme point that is closest to the second intermediate value; If the absolute value of the first extreme point is determined to be greater than the absolute value of the decision threshold, then the first extreme point is determined as the third intermediate value; The smallest value among the third intermediate values ​​is selected as the target extreme point.

5. The method according to any one of claims 1-3, characterized in that, The preset decision threshold ratio constant has a range of values ​​of (0.5, 1).

6. A signal period estimation device, characterized in that, include: The acquisition unit is used to acquire target object data from the sensor and determine the target signal sequence corresponding to the target object based on the target object data. The function determination unit is used to determine the average amplitude difference function of the target signal sequence based on the target signal sequence; and to determine the amplitude difference sum function of the target signal sequence based on the average amplitude difference function of the target signal sequence. The threshold determination unit is used to determine the extreme points and maximum / minimum points of the amplitude difference and function within a preset period range, and to determine the decision threshold based on the function value at the maximum / minimum point and a preset decision threshold proportional constant. The period estimation unit is used to determine a target extreme point that meets a preset condition among the extreme points based on the decision threshold and the extreme point, and to determine the target extreme point as the period estimate of the target signal sequence; the period estimate is used to assist in the identification of the target object; The amplitude difference function is: in, The sum of amplitude differences in the interpolated target signal sequence is represented by the function; P represents the number of sampling points in one repetition period of the target signal sequence; K represents the length of the target signal sequence. The function representing the average amplitude difference of the interpolated target signal sequence; It is a positive integer; [] indicates taking the nearest integer within the brackets; This indicates that the number inside the parentheses is rounded down to the nearest integer.

7. A server, characterized in that, Includes memory and processor; among which, The memory is used to store computer programs; The processor is configured to read the computer program stored in the memory and execute the method described in any one of claims 1-5 according to the computer program in the memory.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the method described in any one of claims 1-5.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1-5.

Images