Method for Evaluating a Sensor Signal

The invention starts out from a method for evaluating a sensor signal. The same kind of method, whereby the integrated output signal from a sensor is compared to a threshold value, is known from the literature reference 1141 Ingenieurs de l'Automobile (1982) no. 6, 69-77. If the output signal from the sensor exceeds the specified threshold value, then restraining devices, such as air bags and/or belt tighteners are released to protect occupants. When conventional methods are used, the disadvantage is that the sensors have to be adjusted in a very costly procedure. In addition, during the lifetime of the motor vehicle, the once effected adjustment can change again, for example due to temperature influences and/or ageing processes that the components are subject to.

The advantage of the method according to the invention is that one can dispense with costly adjusting operations when the sensor is manufactured and installed in a control unit. Furthermore, any variations which occur later on in the form of temperature fluctuations and/or the ageing of components cannot have an adverse effect on the process of evaluating the sensor signals.

Advantageous further developments and refinements of the method of the present invention, and suitable devices for implementing the method will become apparent in view of the following detailed description and drawings .

Exemplified embodiments of the invention are depicted in the drawings and clarified in greater detail in the following description.

FIG. 1 depicts a first diagram representing a sensor signal as a function of time;

FIG. 2 a second diagram representing a sensor signal as a function of time;

FIG. 3 a block diagram of a first exemplified embodiment of a device for implementing the method according to the invention; and

FIG. 4 a block diagram of a second device.

In accordance with a first exemplified embodiment of FIG. 3, a device for implementing the method according to the invention comprises a sensor 1 having several outputs for its output signal. A first output of the sensor 1 is connected to the input terminal of a signal averager 2, whose output terminal is connected to the first input terminal of a first comparative circuit 4. A second output of the sensor 1 is directly connected to a second input terminal of the first comparative circuit 4. Also, a counting circuit 3 is connected to the signal averager 2. The output terminal of the first comparative circuit 4 is connected to a first input terminal of a second comparative circuit 6. A second input terminal of the second comparative circuit 6 is connected to the input terminal of a reference voltage source 5. The output terminal of the second comparative circuit 6 is connected to the input of an evaluation circuit 7, whose output is connected on its part to the input terminal of restraining devices 8.

In FIG. 4, a second exemplified embodiment of a device for implementing the method according to the invention differs still from the exemplified embodiment depicted in FIG. 3 in that the second output of sensor 1 is also not directly connected to an input terminal of a first comparative circuit 4, but rather to the input terminal of a second signal averager 9, whose output terminal is connected to the second input terminal of the first comparative circuit. In addition, a second counting circuit 10 is connected to the second signal averager 9.

The method according to the invention shall be clarified in the following, also with reference to the diagrams depicted in FIG. 1 and FIG. 2. In an idealized representation, FIG. 1 shows the output signal a from the sensor 1 as a function of time t. The sensor 1 is preferably an acceleration-sensitive sensor, for example a piezoelectric crystal, which emits an output signal that is proportional to the acceleration of the motor vehicle. To formulate an average value M from the output signal of sensor 1 by means of the first signal averager 2, according to the following formula ##EQU1## at specifiable time intervals Ti, n measured values ai, whereby i=1-n, are tapped off, added up, and divided by the number of measured values. By means of the counting circuit 3 controlling the first signal averager 2, the number of measured values ai to be tapped off can thereby be specified. The average value M formulated in the first signal averager 2 from the individual measured values ai is then fed to a first input terminal of a first comparative circuit 4, whose second input terminal is directly connected to an output of sensor 1. According to the following formula

a differential signal D is formed in the first comparative circuit 4. This differential signal D is fed from the output terminal of the first comparative circuit 4 to a first input terminal of a second comparative circuit 6. A reference voltage value from a reference voltage source 5 is directed to a second input terminal of the second comparative circuit 6. In this second comparative circuit 6, the differential signal D is compared to a specified threshold value S1 (according to the following inequality (3))

As long as the comparison shows that the differential signal D lies below the specified threshold S1, the average value M is assumed as an updated average value MA. However, if the comparison shows that the value of the differential signal D is greater than the specified threshold value S1, then the preceding value M is assumed as an updated average value MA. It follows from the representation according to FIG. 1 that the measured values a1, a2, a3 of the output signal from sensor 1 are tapped off at successive instants t1, t2, t3. On the one hand, these measured values a1, a2, a3 are each fed to the signal averager 2 and, on the other hand, directly to the first comparative circuit 4. In the first comparative circuit 4, the measured values a1, a2, a3 are compared to the updated average values M1, M2, M3, and the difference D1, D2, D3 is formed. When the difference between the average value M1 and the measured value al, which is acquired at the instant T1, is formed, a comparatively large differential signal value is revealed. According to the comparison in the second comparative stage 6, thus when compared to the specified threshold value S1, this differential signal value exceeds the threshold value S1. Accordingly, the average value M1 is not considered, but an average value M that had been acquired beforehand is retained as an average value M. The same applies to the depicted differential value D2. Only the differential value D3 lies below the specified threshold value S1 and, accordingly, one continues with M3 as an updated average value MA.

It is useful to formulate the average values M from a number of measured values ai of the sensor signal a, which correspond to a power of 2; preferably the number lies within the limits from 100 to about 300, in particular, about 256 measured values.

Compared to the idealized representation of FIG. 1, the diagram according to FIG. 2 shows a more realistic representation of the curve shape of the output signal from sensor 1. One can recognize relatively strong amplitude fluctuations around an imaginary center line of the idealized curve shape. As depicted at the instant t2, when a measured value a2 is acquired after the specified time cycle exactly at the point of an extreme amplitude climb, this measured value deviates drastically from a measured value ai acquired earlier or later on, and can considerably corrupt the expected result during the further processing of the signal. To suppress this disadvantageous effect of amplitude fluctuations, in a further exemplified embodiment of the method according to the invention, in order to acquire the differential signal D, the average value M is compared not only to a single measured value of the output signal a from the sensor 1, but rather to an average value of several measured values of this output signal.

For that, a number of measured values of the output signal a from the sensor 1 is effectively processed, corresponding to a power of 2. In particular 4 to 16, and preferably 8 measured values of the sensor signal a are combined to form an average value. This second average-value formation is implemented by a device according to FIG. 4, in that the output signal a from the sensor 1 is also fed to a second signal averager 9. This second signal averager 9 is controlled by a counting unit 10, which specifies the number of measured values to be processed to form the average value.

In a further advantageous refinement of the invention, the output signal a from the sensor 1 is initially integrated slowly, that is, over a relatively long period of time T1. The acquired integrated measured value I1 is then compared to the output signal a from the sensor 1. The acquired differential value D1, in turn, is compared to a specifiable threshold value S1. If the differential value D1 is greater than the threshold value S1, the output signal a from the sensor 1 is not retrieved for further integration. At the same time, the output signal a from the sensor 1 is integrated briefly, that is over a comparatively short period of time T2, and a floating average-value formation is carried out (integrated measured value I2). The integrated measured values I1, I2 are subsequently compared to one another. The differential value D3 resulting from this comparison is compared to a specifiable threshold value S3. If the differential value D3 exceeds the threshold value S3, then the passenger restraining system is released.