JPH09315260A - Self-diagnostic method of collision judging device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To diagnose the presence of something abnormal with a turnover strain preventing low-pass wave filter circuit connected to an acceleration sensor in use of the collision judgment algorithm of an collision judging device to a virtual collision signal. SOLUTION: Output of an acceleration sensor 2 is made into analog-to-digital conversion via a turnover strain preventing low-pass waver filter circuit 3a, and an operated value generated on the basis of accelerating data subjected to this AD conversion is applied to an collision judging device giving an collision judgment after judging its threshold, issuing a diagnostic command, and thereby the output of the acceleration sensor 2 is imaginarily changed suddenly, and the said operated value threshold-judged at the fixed point of time, thereby diagnosing the low-pass wave filter circuit 3a. A self-disnostic circuit 22 concerned to this diagnosis judges whether four values of absolute value DE, short interval integral value BV, short interval integral value/differential value DB and long-interval integral value CV is in the specified threshold range held between both upper and lower threshold values or not, and in the case where even one value comes off the threshold range, it is so judged that there is something abnormal.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention makes it possible to diagnose whether or not there is an abnormality in a low pass filter circuit for preventing aliasing distortion connected to an acceleration sensor by using a collision determination algorithm of a collision determination device for a virtual collision signal. The present invention relates to a self-diagnosis method for a collision determination device.

[0002]

2. Description of the Related Art A collision determination device 1 shown in FIG. 4 is a device for deploying an in-vehicle airbag when a collision occurs. This collision determination device is a bandpass filter in which an acceleration signal G obtained by the pulse signal output of the acceleration sensor 2 is cascade-connected with a low-pass filter circuit 3a for removing aliasing distortion and a high-pass filter 3b for digital signal processing. Band-pass filtering by 3 and AD-converting it into the digital signal processing unit 1a, and in the digital signal processing unit 1a, there are four types of absolute value determination, short interval integration determination, short interval integration value differential determination, and long interval integration determination. Make a collision judgment by combining the judgments. A deployment signal is output to the airbag simultaneously with the collision determination, and the airbag deploys as the detonator element is energized.

As the acceleration sensor 2, there is used one in which a stress-strain gauge utilizing a piezo resistance change is incorporated on a semiconductor substrate with its pressure receiving surface facing the traveling direction of the vehicle.
Not only the semiconductor acceleration sensor that detects a change in piezoresistance but also a piezoelectric acceleration sensor or a pure mechanical elastic spring can be used. The pulse output signal G output from the acceleration sensor 2 has its high-frequency component removed in order to eliminate the influence of aliasing distortion in the low-pass filter circuit 3a having an analog circuit configuration before being converted into discrete-value data, and the subsequent digital signal processing unit. It is sent to the high-pass filter circuit 3b, the short-range integrator 4, and the long-range integrator 5 of the digital circuit configuration in 1a via the AD conversion input port. In the case of this example, the AD conversion input port is provided with an AD conversion function, and therefore the acceleration signal detected by the acceleration sensor 2 and band-limited is converted into digital acceleration data when passing through the AD conversion input port. To be converted. Further, the low-pass filtering circuit 3a for preventing aliasing distortion has a filtering characteristic of removing high-frequency components exceeding 70 to 260 Hz contained in the acceleration signal, for example, high-frequency components exceeding 70 Hz, and the high-pass filtering circuit of the next stage. 3b is a low-pass filtering circuit 3a
Has a filtering characteristic for removing low-frequency components of 20 to 170 Hz or less included in the output of, for example, low-frequency components of 20 Hz or less.

The acceleration data G (k) extracted by the bandpass filter 3 is sent to an absolute value circuit 6 where it is converted into an absolute value │G (k) │ which is independent of the positive / negative of the acceleration signal and continues. After the threshold value is discriminated by the comparator 7 with reference to a constant reference value Er, it is sent to the waveform shaper 8. In the waveform shaper 8, the impact force that exceeds the threshold Er is detected by the comparator 7.
When it is obtained as the output of, the output of the comparator 7 is expanded in the time axis direction for a fixed time and shaped into a waveform that lasts at least a fixed time. The waveform shaper 8 shown in this example is a one-shot circuit 8a that is triggered by the rising edge of the output of the comparator 7 to generate a one-shot pulse that lasts 30 ms.
And an OR gate circuit 8b which takes the logical sum of the output one-shot pulse of the one-shot circuit 8a and the original output of the comparator 7. Therefore, when the output of the absolute value circuit 6 shows a sudden change in impact force exceeding the dangerous value,
The fact that the output of the waveform shaper 8 has a high possibility of collision detection will continue to be clearly shown over the period of 30 ms.

On the other hand, the short-term and long-term interval integrators 4,
In the interval integration performed by 5, the acceleration data G (k), which has been digitized when passing through the AD conversion input port,
The short interval integrator 4 performs interval integration for 10 to 14 ms, for example, 14 ms, and the long interval integrator 5 performs integration for 90 ms, for example. The long-interval integrated output is supplied to the subsequent comparator 9 with a constant reference value E.
When a threshold value is discriminated based on rl and a long-interval integrated value exceeding a constant reference value Erl is obtained, a high-level signal is supplied to the 3-input 1-output type OR gate circuit 11 in the decision circuit 10. The determination circuit 10 includes two AND gate circuits 12 and 13 in addition to the OR gate circuit 11, and the outputs of the AND gate circuits 12 and 13 and the output of the comparator 9 are supplied to the OR gate circuit 11. It

The output of the short interval integrator 4 is the output of the pair of comparators 1.
4, 15 and the differentiator 16. Comparator 14, 1
5 is a reference value VrS which is a constant reference value for each of the short-term integrated outputs.
The threshold value is discriminated based on 1 and VrS2. Differentiator 16
Uses the algorithm according to Simpson's formula to time-differentiate the short-term integrated output, and supplies the obtained time differential value to the comparator 17. Incidentally, according to Simpson's formula, the time differential value is the current value V (k) of the short-term integral output and the data V (k-1), V (k-
2) and V (k-3), {(V (k) + 3V (k-1) -3V (k-2) -V
(K-3)} / 6. The comparator 17 performs a threshold value determination on the time differential value of the short-term integrated output with reference to a constant reference value αr and supplies the threshold value determination output to the AND gate circuit 13 to perform a logical product operation of the outputs of the comparators 15. To serve.

By the way, the conventional collision determination device described above is
The low-pass filter circuit 3a for removing the aliasing distortion due to the D conversion has an analog circuit configuration, and when a short circuit or an open of a resistor or a capacitor which is a component of the low pass filter circuit 3a occurs, the airbag should not be deployed and a light collision. Or, it is a cause of causing troubles such as causing a burst when riding on a curb or traveling in a pothole, or vice versa, although the deployment trigger signal may be delayed despite the accident that the airbag should be deployed quickly. It was For this reason, conventionally, a redundant circuit configuration in which the resistors and capacitors forming the low-pass filtering circuit 3a are divided or doubled is adopted, which causes an accident such as a burst or a miss and a delay in the deployment time. Has been configured to prevent the above. For example, in the low-pass filtering circuit 3a illustrated in FIG. 5, in forming an active low-pass filter using an operational amplifier, a pair of cascade-connected input resistors R1 and R2 are connected to a non-inverting input terminal, and an output is provided. Feedback resistors R3 and R4 are connected in series in a feedback path connecting the terminal and the inverting input terminal, and the connection point of the input resistors R1 and R2 is connected to the output terminal via the feedback capacitor C1 and the non-inverting input terminal is connected to the capacitor C2. It was grounded via. However, the total resistance value of the feedback resistors R3 and R4 is equal to the total resistance value of the input resistors R1 and R2.

[0008]

The low-pass filtering circuit 3a for preventing the aliasing distortion of the above-described conventional collision determination device has a redundant circuit configuration because it is advantageous in view of the probability of occurrence of an accident in which a circuit element fails. Therefore, if both of the divided or double-provided circuit elements fail, the return distortion cannot be removed, or even the collision cannot be determined. For this reason, it is necessary to thoroughly check that there is no abnormality in the circuit elements forming the low-pass filtering circuit 3a in shipping inspection or the like. However, even if a thorough quality control system is put in place, anomalies that occur after shipping can only be found in regular inspections, and abnormalities that cannot be found in regular inspections are a serious problem.
There were also problems with the method of regular inspection itself.

For example, in the low-pass filtering circuit 3a shown in FIG. 5, when the input resistor R1 is short-circuited, the step response waveform of the low-pass filtering circuit 3a has a rising time constant as shown by the chain line in FIG. As a result of the shortened period, it will start up fairly early. Therefore, it is possible to detect the abnormality of the circuit element of the low pass filter circuit 3a by analyzing the step response waveform. However, when an open accident of the feedback capacitor C1 occurs, the step response waveform of the low-pass filtering circuit 3a has a very steep rising slope, as shown by the alternate long and short dash line in FIG. Because of the deviation, the rising curve itself is included in the allowable range, and the low-pass filtering circuit 3a
There is a problem that the abnormality of the feedback capacitor C1 cannot be found only by observing the step response waveform of.

The present invention has been made to solve the above-mentioned problems and uses the collision determination algorithm of a collision determination device for a virtual collision signal to determine whether or not there is an abnormality in a low-pass filtering circuit connected to an acceleration sensor for preventing aliasing distortion. The purpose is to diagnose.

[0011]

In order to achieve the above object, the present invention AD-converts the output of an acceleration sensor through a low-pass filtering circuit for eliminating aliasing distortion, and based on the AD-converted acceleration data. The generated calculated value is applied to a collision determination device that determines a threshold value to make a collision determination, issues a diagnostic command to virtually suddenly change the output of the acceleration sensor, and the calculated value is determined to be a threshold value at a predetermined time to determine the collision. It is characterized by diagnosing a low-pass filtering circuit.

Further, according to the present invention, the collision determining device integrates the absolute value circuit for calculating the absolute value of the acceleration data and the acceleration data over a relatively short interval up to the current value, and integrates the short interval. A short interval integrator for calculating a value, a short interval integrated value differentiator for differentiating the short interval integrated value with time to calculate a short interval integrated value differential value, and the acceleration over a relatively long interval up to the current value. And a long-range integrator for calculating a long-range integrated value, and the diagnosis has four upper limits of the absolute value, the short-range integrated value, the short-range integrated value, and the long-range integrated value. It is characterized in that it is determined whether or not it is within a predetermined threshold range sandwiched between a threshold and a lower limit threshold, and if even one value is outside the threshold range, it is determined that there is an abnormality.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to FIGS. FIG. 1 is a circuit configuration diagram showing an embodiment of a collision determination device to which the self-diagnosis method of the present invention is applied, and FIG. 2 is a flowchart for explaining a self-diagnosis operation by the self-diagnosis circuit shown in FIG. Three
FIG. 2 is a signal waveform diagram of each part of the circuit shown in FIG. 1 observed during self-diagnosis.

In the collision determination device 21 shown in FIG. 1, a self-diagnosis circuit 22 is provided in the digital signal processing section 1a of the conventional collision determination device 1 so that the low-pass filtering circuit 3a can be self-diagnosed each time the ignition key is operated. It is configured to be opened. The self-diagnosis circuit 22 starts self-diagnosis in response to the battery power being turned on in response to the operation of the ignition key, applies a pulse voltage as a diagnosis command to the acceleration sensor 2, and virtually outputs an acceleration signal of about 25 G, for example. The low-pass filtering circuit 3a is diagnosed by generating a threshold value of the response of the calculated value of each part in the digital signal processing section 1a to the sudden change of the output of the acceleration sensor 2 at a predetermined time point. Specifically, the self-diagnosis circuit 22 outputs the outputs of the absolute value circuit 6, the short interval integrator 4, the differentiator 16, and the long interval integrator 5 when a predetermined time τ has elapsed after issuing the diagnosis command. It is determined whether or not the four values of the absolute value DE, the short-term integrated value BV, the short-term integrated value differential value DB, and the long-term integrated value CV that are supplied are within predetermined threshold ranges sandwiched by the upper threshold and the lower threshold, respectively. If even one value is outside the threshold range, it is determined that there is an abnormality.

First, in step (101) of FIG. 2, the self-diagnosis circuit 22 applies the pulse voltage shown in FIG. 3A as a diagnosis command to the self-diagnosis input section of the acceleration sensor 2. In the following decision step (102),
It is determined whether or not the acceleration sensor 2 outputs the self-diagnosis signal shown in FIG. If the acceleration sensor 2 does not respond, it is determined in step (106) that there is an abnormality. When the self-diagnosis signal is output from the acceleration sensor 2, it is then determined in the determination step (103) whether the waveform of the filtered output of the low-pass filtering circuit 23a is disturbed or the filtered output itself is present. Also in this case, if the filtered output waveform is disturbed or the filtered output itself does not exist,
The process proceeds to step (106) to notify the occurrence of abnormality.

On the other hand, when the unfiltered filtered output shown in FIG. 3 (C) is obtained from the low-pass filtering circuit 3a, at the decision step (104), when a certain time τ has elapsed since the diagnostic command was issued. Then, four kinds of judgments, absolute value judgment, short section integration judgment, short section integration value differential judgment and long section integration judgment, are performed. The time τ that defines the timing for making these four types of determinations depends on the cutoff frequency of the low-pass filtering circuit 3a, but is required for the filtered output of the low-pass filtering circuit 3a shown in FIG. 3C to settle. It is set to half the time or a time close to it based on the experience value. Here, DE is an absolute value output, BV is a short interval integrated output, DB is a short interval integrated value differential output, and CV is a long interval integrated output. These four calculated values DE, BV, DB, The threshold range of CV is
As shown in FIGS. 3D to 3G, it is defined as a range sandwiched between the lower limit of the threshold value with the suffix L attached to each calculated value and the upper limit of the threshold value with the suffix H attached to each calculated value. is there.

Therefore, in decision step 104, D
EL <DE <DEH, BVL <BV <BVH, DBL <
When it is confirmed that DB <DBH and CVL <CV <CVH, it is found that the low-pass filtering circuit 3a has no abnormality, and it is concluded that there is no abnormality in step (105). Further, in the judgment step (104), if any one of the absolute value output DE, the short interval integrated output BV, the short interval integrated value differential output DB and the long interval integrated output CV deviates from the predetermined threshold range, the low range Since an abnormality is found in any of the circuit elements forming the filtering circuit 3a, the process proceeds to step (106), and an alarm lamp or the like is turned on to notify the abnormality. Further, the collision determination device itself suspends the collision determination until the abnormality is resolved.

In the low-pass filtering circuit 3a having the circuit configuration shown in FIG. 5, when the feedback capacitor C1 causes an open accident, it is not possible to detect an abnormality simply by observing the filtered output of the low-pass filtering circuit 3a. However, according to the accident diagnosis method of the present embodiment, when an open accident of this kind occurs, the output of the differentiator 16, that is, the value of the short-range integrated value differential output DB deviates from the threshold range, so that the abnormality immediately occurs. Can be found. Further, when the input resistance R1 of the low-pass filtering circuit 3a is short-circuited, the short-range integrated value BV first deviates from the threshold range, so that an abnormality can be immediately detected.

Thus, according to the above self-diagnosis method,
The output of the acceleration sensor 2 is AD-converted through the low-pass filtering circuit 3a for removing aliasing distortion, and applied to a collision determination device that makes a threshold determination by performing a threshold determination on a calculation value generated based on the AD-converted acceleration data. Then, a diagnosis command is issued to virtually change the output of the acceleration sensor 3a, and the calculated value is discriminated by a threshold value at a predetermined time to diagnose the low-pass filtering circuit. Abnormalities of circuit elements that cannot be found only by observing the response waveform can be indirectly determined using the algorithm used for collision determination, and most of the collision determination algorithms can be used. The self-diagnosis of the low-pass filtering circuit 3a can be performed simply and reliably by issuing a diagnostic command at the time of operating the ignition key.

Further, the self-diagnosis circuit 22 has a predetermined value in which the four values of the absolute value DE, the short interval integrated value BV, the short interval integrated value differential value DB, and the long interval integrated value CV are sandwiched between the upper limit threshold and the lower limit threshold, respectively. Since it is determined whether or not it is within the threshold range, and if even one value is out of the threshold range, it is determined that there is an abnormality, self-diagnosis can be performed in multiple directions based on four types of calculated values, and low range Abnormalities of circuit elements that are unlikely to appear in the step response waveform of the filtering circuit 3a can be accurately detected by observing the interval integral value of acceleration data and its time differential value.

In the above embodiment, the self-diagnosis circuit 2
Although two types of judgments, absolute value judgment, short-term integral judgment, short-term integral value differential judgment and long-term integral judgment, are made within 2, the digital signal processing unit 1a is already provided for collision judgment. The comparators 7, 14, 17, 9 are used, and Er, Vrs1, αr, Erl, etc., which are comparison criteria thereof, are appropriately switched and set for self-diagnosis, and the comparators 7, 14,
It is also possible to monitor the outputs of 17, 9 and make an abnormality determination.

[0022]

As described above, according to the present invention,
The output of the acceleration sensor is AD-converted through a low-pass filtering circuit for aliasing distortion removal, and is applied to a collision determination device that makes a threshold determination by performing a threshold determination on a calculation value generated based on the AD-converted acceleration data, Since the diagnosis command is issued to virtually suddenly change the output of the acceleration sensor, and the calculated value is discriminated by a threshold value at a predetermined time to diagnose the low-pass filtering circuit, the step response waveform of the low-pass filtering circuit is calculated. Abnormalities of circuit elements that cannot be detected only by observing can be indirectly determined using the algorithm used for collision determination, and most of the collision determination algorithms can be used. As long as the diagnostic command is issued at the time of operation, there is an excellent effect that the low-pass filtering circuit can be self-diagnosed easily and surely.

Further, the collision determination device integrates an absolute value circuit for calculating an absolute value of the acceleration data and a short value for integrating the acceleration data to a current value over a relatively short interval to calculate a short interval integrated value. An interval integrator, a short interval integral value differentiator that differentiates the short interval integral value with time to calculate a short interval integral value differential value, and the acceleration is integrated over a relatively long interval up to the current value. A long interval integrator for calculating an interval integrated value is provided, and in the diagnosis, four values of the absolute value, the short interval integrated value, the short interval integrated value differential value, and the long interval integrated value are set as an upper threshold value and a lower threshold value, respectively. Since it is determined whether or not it is within a predetermined threshold value range sandwiched, and if even one value is outside the threshold value range, it is determined that there is an abnormality. Therefore, self-diagnosis is multilaterally performed based on four types of calculated values. And appears in the step response waveform of the low pass filter. The abnormality of the pile circuit elements, an effect such as can be accurately found by observing the interval integral value and its time differential value of the acceleration data.

[Brief description of drawings]

FIG. 1 is a circuit configuration diagram showing an embodiment of a collision determination device to which a self-diagnosis method of the present invention is applied.

FIG. 2 is a flowchart for explaining a self-diagnosis operation by the self-diagnosis circuit shown in FIG.

FIG. 3 is a signal waveform diagram of each part of the circuit shown in FIG. 1 observed during self-diagnosis.

FIG. 4 is a circuit configuration diagram showing an example of a conventional collision determination device.

5 is a circuit diagram showing an example of a low-pass filter circuit for removing aliasing distortion shown in FIG.

6 is a diagram showing a step response waveform of the low pass filter circuit when the input resistance shown in FIG. 4 is short-circuited.

FIG. 7 is a diagram showing a step response waveform of the low pass filter circuit when the feedback capacitor shown in FIG. 4 is open.

[Explanation of symbols]

 1a Digital signal processing unit 2 Acceleration sensor 3 Bandpass filter 3a Low-pass filtering circuit 3b High-pass filtering circuit 4 Short-range integrator 5 Long-range integrator 6 Absolute value circuit 10 Judgment circuit 16 Differentiator 21 Collision judgment device 22 Self-diagnosis circuit

Claims (2)

[Claims] 1. A collision determination in which an output of an acceleration sensor is AD-converted through a low-pass filtering circuit for removing aliasing distortion, and a calculation value generated based on the AD-converted acceleration data is subjected to threshold value determination to make a collision determination. A collision determination device, which is applied to a device, issues a diagnostic command to virtually suddenly change the output of the acceleration sensor, and determines the threshold value of the calculated value at a predetermined time to diagnose the low-pass filtering circuit. Self-diagnosis method. 2. The collision determination device calculates an absolute value circuit for calculating an absolute value of the acceleration data, and integrates the acceleration data up to a current value over a relatively short interval to calculate a short interval integrated value. A short interval integrator, a short interval integrated value differentiator for calculating a short interval integrated value differential value by differentiating the short interval integrated value with time, and integrating the acceleration over a relatively long interval up to the current value, A long-range integrator for calculating a long-range integrated value is provided, and in the diagnosis, four values of the absolute value, the short-range integrated value, the short-range integrated value differential value, and the long-range integrated value are an upper threshold value and a lower threshold value, respectively. 2. The self-diagnosis method for a collision determination device according to claim 1, wherein it is determined whether or not it is within a predetermined threshold range sandwiched between the two, and if even one value is outside the threshold range, it is determined that there is an abnormality.