METHOD FOR TEMPORAL SYNCHRONIZATION OF A FIRST SIGNAL AND A SECOND SIGNAL
By resampling and calculating optimal time shifts for synchronization, the method addresses the limitations of existing synchronization methods, enhancing accuracy and efficiency in signal alignment.
Patent Information
- Authority / Receiving Office
- FR · FR
- Patent Type
- Applications
- Current Assignee / Owner
- STELLANTIS AUTO SAS
- Filing Date
- 2024-12-06
- Publication Date
- 2026-06-12
AI Technical Summary
Existing methods for time synchronization of signals recorded by separate digital recorders during vehicle tests are limited by initial hypotheses without testing for relevance, affecting accuracy and computation time.
A method involving resampling the second set of signals to match the sampling period of the first set, calculating point-to-point differences for various time offsets, determining the optimal time shift for minimal deviation, and applying this shift to achieve synchronization.
This approach ensures optimal synchronization between signals from unsynchronized recorders, improving accuracy and reducing computational complexity.
Smart Images
Figure 00000000_0000_ABST
Abstract
Description
Title of the invention: METHOD FOR TIME SYNCHRONIZING A FIRST SIGNAL AND A SECOND SIGNAL
[0001] The invention relates to the time synchronization of signals recorded during tests. These tests may be tests carried out in the field of vehicle design and development, for example vehicle tests on roller test benches.
[0002] Patent application EP3296767 describes a method comprising a step of acquiring a first signal by a first digital recorder and a step of acquiring a second signal by a second digital recorder, the first recorder and the second recorder measuring, in separate time bases, a parameter of the vehicle's motion at a distance from the first and second digital recorders, for example, the measured motion parameter is the vehicle's speed. The method includes a step of applying a plurality of assumptions, including time offsets, to the first signal. The method also includes a step of calculating an estimate of the time synchronization between the first and second signals for each of the applied assumptions. One estimate is selected from among the calculated estimates.
[0003] However, the estimates derived from the hypotheses thus formulated are limited to comparing the first signal and the second signal. Therefore, the accuracy of the estimate is contingent upon the hypotheses as initially formulated, without testing, during the calculation of the estimates, whether one hypothesis is more relevant than the others, which saves computation time.
[0004] The objective of the present invention is to remedy this drawback and to improve the time synchronization between signals recorded during tests by separate recorders.
[0005] To achieve this objective, the invention proposes a method for time synchronizing a first set of signals with a second set of signals, the method comprising the following steps: - a step of acquiring the first set of signals by a first digital recorder; - a step of acquiring a second set of signals by a second digital recorder; each of these signal sets includes a so-called common signal corresponding to the acquisition of the same quantity, - a step of resampling the common signal of the second set of signals to the sampling period of the common signal of the first set of signals; - a step in which, for time offsets between a minimum time offset and a maximum time offset, the second common signal is offset relative to the first common signal and a point-to-point difference value is calculated between the first common signal and the second offset common signal, this for each of the time offsets assigned to the second common signal; - a step to determine the optimal time shift for which the deviation value is the smallest among those calculated; - a step of applying the optimal time shift to the second set of signals.
[0006] This makes it possible to obtain optimal synchronization between the first set of signals and the second set of signals, from two unsynchronized recorders.
[0007] Various additional features may be provided, alone or in combination:
[0008] In one embodiment, the time offsets assigned to the second common signal are multiples of the sampling period of the common signal.
[0009] In one embodiment, the point-to-point difference value between the first common signal and the second common signal is calculated using the following formula:
[0010] (S2(p) -S^(p))2 + nnc x maxÇs'jj 2
[0011] where VC(x) is the point-to-point difference value between the first common signal and the second common signal calculated for the time offset denoted x; ne is the number of common points between the first signal and the second signal for the time shift x; is the value of the first signal for a point, denoted p; S2(p) is the value of the second signal for the point, denoted p; nnc is the number of non-common points between the first signal and the second signal for the time shift x; max(S1) is the maximum value of the first signal.
[0012] In one embodiment, the minimum time offset is chosen so that the last point of the second common signal is aligned with the first point of the first common signal.
[0013] In one embodiment, the maximum time offset is chosen so that the last point of the first common signal is aligned with the first point of the second common signal.
[0014] In one embodiment, the method includes a step of calculating a coherence indicator between the first common signal and the second common signal.
[0015] In an embodiment, the consistency indicator is defined by the following formula:
[0016] VC^p^-WK^max^S' J2 the~ not
[0017] where is the value of the consistency indicator; VC(xopti) is the smallest deviation value among those calculated; ncc is the number of absent values of the second signal relative to the first signal, max(S|) is the maximum value of the first signal.
[0018] In one embodiment, the method includes a step of issuing an alert when the value of the consistency indicator is greater than a predetermined consistency indicator value.
[0019] The invention also relates to a computer program comprising program code instructions for executing the steps of the process defined as above, when said program is running on a computer.
[0020] The invention further relates to an assembly comprising means for processing, by means of software instructions stored in a memory, for the implementation of the computer program defined as above.
[0021] The invention will be further detailed by describing a non-limiting embodiment and based on the accompanying figures in which:
[0022] [Fig. 1] illustrates a flowchart representing the steps of a method for synchronizing a first set of signals with a second set of signals, according to an embodiment of the present invention.
[0023] [Fig.2] illustrates a graph showing, on the y-axis, the value of the first signal common, in solid line, and of the second common signal, in dashes, as a function of time according to the same temporal reference frame.
[0024] [Fig.3] illustrates a graph analogous to that of [Fig.2], the second signal common being temporally offset by -30 seconds relative to the first common signal.
[0025] [Fig.4] illustrates a graph analogous to that of [Fig.2], the second signal common being temporally offset by +20 seconds relative to the first common signal.
[0026] [Fig.5] illustrates a graph analogous to that of [Fig.2], comprising a value the calculated difference between the first common signal and the second common signal, the the second common signal being temporally offset by -30 seconds relative to the first common signal.
[0027] [Fig.6] illustrates a graph analogous to that of [Fig.2], the second signal common being temporally offset by -15 seconds relative to the first common signal, the graph includes, in points connected by solid lines, the calculated deviation values.
[0028] [Fig.7] illustrates a graph analogous to that of [Fig.6], the second signal common being temporally offset by -5 seconds relative to the first common signal.
[0029] [Fig.8] illustrates a graph analogous to that of [Fig.6], the second signal common being time-shifted by 0 seconds relative to the first common signal.
[0030] [Fig.9] illustrates a graph analogous to that of [Fig.6], the second signal common being temporally offset by 5 seconds relative to the first common signal.
[0031] [Fig.10] illustrates a graph analogous to that of [Fig.6], the second common signal being time-shifted by 10 seconds relative to the first common signal.
[0032] [Fig. 11] illustrates a graph analogous to that of [Fig.6], the second common signal being time-shifted by 20 seconds relative to the first common signal.
[0033] Figure 1 illustrates a flowchart of a method for synchronizing a digital recording of a first set of signals by a first digital recorder with a digital recording of a second set of signals by a second digital recorder. The first and second recorders are separate. The two recorders are not synchronized; in other words, the recordings are never triggered simultaneously. Furthermore, the sampling frequency is not necessarily the same for the two recorders. Each of these two recordings contains a specific signal corresponding to the acquisition of the same quantity, which will also be referred to as the common signal.
[0034] Thus, since the two recorders are not synchronized, the first set of signals is measured according to a first time reference frame, while the second set of signals is measured according to a second time reference frame. In general, the first time reference frame is different from the second time reference frame. Moreover, the duration of the second recording is greater than the duration of the first recording, which implies that the latter temporally covers the duration of the first recording.
[0035] The steps of the process described below illustrate one embodiment of the present invention.
[0036] In a non-limiting embodiment, tests are performed with a vehicle on a roller test bench. The roller test bench is configured to move the vehicle by actuating the rollers in contact with the vehicle's wheels. In this configuration, the first recorder is connected to the roller test bench and the second recorder is connected to the vehicle.
[0037] In an acquisition step El, the first set of signals is acquired by the first digital recorder.
[0038] In an acquisition step E2, the second set of signals is acquired by the second digital recorder.
[0039] In this example, the common signal is the vehicle's speed as a function of time. Thus, the first recorder and the second recorder both measure this same physical quantity. The signal from the first recorder will be referred to as the first common signal, and the signal from the second recorder as the second common signal.
[0040] By way of illustration, [Fig.2] represents on the same time line, the common signal of each recording, the acquisition of the second recording having been launched 5 seconds earlier (in absolute time) than the acquisition of the first recording.
[0041] For the remainder of this text, the time base of the common signal of the first recording serves as the time reference. All time shifts made for the common signal of the second recording are expressed in this reference time base.
[0042] In a resampling step E3, the common signal of the second recording is resampled at the same sampling period as the common signal of the first recording, which serves as the reference signal.
[0043] Figure 2 shows an example of common signals, in this case the vehicle speed, from two recordings of different durations that were not started at the same time during the test (in absolute time). The common signal from the first recording lasts 20 seconds and is sampled at a sampling frequency of 1 Hz, thus comprising 21 points. The common signal from the second recording lasts 30 seconds and was resampled at a frequency of 1 Hz, thus comprising 31 points. The acquisition of the common signal from the second recording was started 5 seconds (in absolute time) before the start of the recording of the common signal from the first recording. In this example, the values (on the y-axis of Figure 2) of the common signal from the second recording to be synchronized are slightly higher than those of the common signal from the first recording. Reference recording. The maximum signal value of the reference recording is 25 km / h.
[0044] In a step E4, a time offset is assigned to the second common signal with respect to the first common signal.
[0045] In a calculation step E5, a point-by-point difference value is calculated between the first common signal and the second common signal, for the time offset assigned to the second signal, this for the points of the two signals which are common in time.
[0046] For example, when a time offset, denoted x, is applied to the second signal with respect to the first signal, the first point-to-point difference value between the first signal and the second signal, denoted VC(x), is calculated by the following formula:
[0047] v(^x^^p=i(S2(p)-S^pY^ + nnc xmax^J 2
[0048] where VC(x) is the point-to-point difference value between the first signal and the second signal calculated for the time shift denoted x, ne is the number of common points between the first signal and the second signal for the time shift x, S^p) is the value of the first signal for a point, denoted p, S2(p) is the value of the second signal for the point, denoted p, nnc is the number of non-common points between the first signal and the second signal for the time shift x, max(S1) is the maximum value of the first signal.
[0049] Steps E4 and E5 are repeated for time offset values between a minimum and a maximum time offset. In other words, the point-to-point offset values between the first common signal and the second common signal are calculated for time offsets ranging from the minimum to the maximum time offset, in predetermined time steps. The time step between two time offsets is equal to the sampling period of the first common signal and therefore of the second resampled signal. The time offsets assigned to the second common signal are thus multiples of this sampling period.
[0050] As illustrated in [Fig.3], the minimum time offset is chosen so that the last point of the second common signal is aligned with the first point of the first common signal.
[0051] For example, the minimum time offset value is equal to -30 seconds as illustrated in [Fig.3].
[0052] As illustrated in [Fig.4], the maximum time offset is chosen so that the last point of the first common signal is aligned with the first point of the second common signal.
[0053] For example, the maximum time offset value is equal to 20 seconds, as illustrated in [Fig.4].
[0054] In a step E6, the optimal time shift, xopti, is determined, for which the deviation value VC(x) is the smallest among those calculated in steps E4 and E5.
[0055] In step E7, this optimal time offset is applied to the second common signal. This achieves optimal synchronization of the first common signal with the second common signal. This offset is also applied to all other signals in the second recording. Thus, all signals in both recordings are realigned. The other signals in the second recording are not necessarily resampled. The realigned signals can then be used and analyzed.
[0056] Figures 5 to 11 are described to illustrate one embodiment of the present invention.
[0057] In [Fig. 5], the point-to-point deviation value is calculated for a time offset -30 seconds. A single point is common between the first and second signals. The point-to-point difference value is calculated using the following formula:
[0058] yc(-30) =0 + 20 x 252 = 12500km2 / h
[0059] In [Fig. 6], the point-to-point deviation value is calculated for a time offset -15 seconds. 8 points are common between the first and second signals. The point-to-point difference value is calculated using the following formula:
[0060] VC(-15) = 2371.4+13 x 252 = 9871.4 km2 / h
[0061] In [Fig. 7], the point-to-point difference value is calculated for a time offset of -5 seconds. 21 points are common between the first signal and the second signal. The point-to-point difference value is calculated using the following formula:
[0062] VC(-5) =5.4 + 0 x 252 = 5.4km2 / h
[0063] In [Fig. 8], the point-to-point difference value is calculated for a time offset of 0 seconds. 21 points are common between the first and second signals. The point-to-point difference value is calculated using the following formula:
[0064] VC(O) =3889.8 + 0 x 252 = 3889.8km2 / h
[0065] In [Fig.9], the point-to-point deviation value is calculated for a time offset of 5 seconds. 16 points are common between the first signal and the second signal. The point-to-point difference value is calculated using the following formula:
[0066] VC (5) = 6707.1 + 5 x 252 = 9832.1km2 / h
[0067] In [Fig. 10], the point-to-point difference value is calculated for a time offset of 10 seconds. Seven points are common between the first and second signals. The point-to-point difference value is calculated using the following formula:
[0068] yc(10) = 1492.2 + 14 x 252= 10242.2km2 / h
[0069] In [Fig. 11], the point-to-point difference value is calculated for a time offset of 20 seconds. Seven points are common between the first and second signals. The point-to-point difference value is calculated using the following formula:
[0070] VC (20) = 0 + 20 x 252 = 12500 km 2 / h
[0071] Therefore, in this example, the calculated deviation value for the time offset of -5 seconds is the smallest of the calculated deviation values, so the time offset of -5 is applied to the second signal in the application step E7.
[0072] In a step E8, a coherence indicator is calculated between the first common signal and the second common signal. This coherence indicator is calculated for optimal synchronization.
[0073] For example, when the time offset is obtained in step E6, the consistency indicator, denoted by the, is defined by the following formula:
[0074] VcÇrop(1)rHMCxmax(S J2
[0075] where is the value of the coherence indicator, VC(xopti) is the smallest VC(x) deviation value among those calculated in steps E4 and E5, ncc is the number of missing values of the second signal relative to the first signal, max^'J is the maximum value of the first common signal ne is the number of common points between the first common signal and the second common signal, with the time offset being applied to the second common signal.
[0076] The formulation of this consistency indicator has the advantage of being able to define a threshold which does not depend on the number of points acquired and therefore on the duration of the recording.
[0077] In the present example, the coherence indicator value between the first signal and the second signal is equal to 5.4 / 21, or 0.257 km2 / h.
[0078] In step E9, an alert is issued when the consistency indicator has a value greater than a consistency indicator threshold value. This allows for alerting to a lack of consistency between the first common signal and the second common signal, resulting, for example, from the outcome of an attempt to synchronize recordings not originating from the same trial.
Claims
Demands
1. A method for time synchronizing a first set of signals with a second set of signals, the method comprising the following steps: - an acquisition step (E1) of the first set of signals by a first digital recorder, - an acquisition step (E2) of a second set of signals by a second digital recorder, each of these sets of signals comprising a so-called common signal corresponding to the acquisition of the same quantity, - a resampling step (E3) of the common signal of the second set of signals to the sampling period of the common signal of the first set of signals, - a step (E4, E5) in which, for time offsets between a minimum time offset and a maximum time offset,The second common signal is offset (E4) relative to the first common signal, and a point-to-point offset value is calculated (E5) between the first common signal and the second offset common signal. This is done for each of the time offsets assigned to the second common signal. The process then follows a step (E6) of determining the optimal time offset (xopti) for which the offset value is the smallest among those calculated, and a step (E7) of applying the optimal time offset (xopti) to the second set of signals, so as to obtain optimal synchronization between the first set of signals and the second set of signals.
2. A method according to claim 1, characterized in that the time offsets assigned to the second common signal are multiples of the sampling period of the common signal.
3. A method according to claim 1 or 2, characterized in that the point-to-point difference value between the first common signal and the second common signal is calculated by the following formula: ( $2 ( p ) - ( p ) )2+ nnc x max(s J 2 where VC(x) is the point-to-point difference value between the first common signal and the second common signal calculated for the time offset denoted x, ne is the number of common points between the first signal and the second signal for the time shift x, $](?) eSt the value of the first signal for a point, denoted p, S^p) is the value of the second signal for the point, denoted p, nnc is the number of non-common points between the first signal and the second signal for the time shift x, max(S|) is the maximum value of the first signal.
4. A method according to any one of claims 1 to 3, characterized in that the minimum time offset is chosen so that the last point of the second common signal is aligned with the first point of the first common signal.
5. A method according to any one of claims 1 to 4, characterized in that the maximum time offset is chosen so that the last point of the first common signal is aligned with the first point of the second common signal.
6. A method according to any one of the preceding claims, characterized in that it comprises a calculation step (E8) of a coherence indicator between the first common signal and the second common signal.
7. Method according to the preceding claim, characterized in that the coherence indicator is defined by the following formula: *C “ fie where le is the value of the coherence indicator, VC(xopti) is the smallest deviation value among those calculated, nnc is the number of non-common points between the first signal and the second signal for the time offset x, max(Sj) is the maximum value of the first signal.
8. A method according to the preceding claim, characterized in that it includes a step of issuing an alert (E9) when the value of the consistency indicator is greater than a predetermined consistency indicator value.
9. Computer program comprising program code instructions for carrying out the steps of the process according to any one of claims 1 to 8, when said program is running on a computer.
10. Assembly comprising means for processing, by means of software instructions stored in memory, for the implementation of the computer program according to claim 9.