GB2530374A

GB2530374A - Device and method for sound-based environment detection

Info

Publication number: GB2530374A
Application number: GB1511594.2A
Authority: GB
Inventors: Karl Matthias
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 2014-07-07
Filing date: 2015-07-02
Publication date: 2016-03-23
Anticipated expiration: 2035-07-02
Also published as: DE102014213122A1; GB2530374B; GB201511594D0; FR3023622B1; FR3023622A1

Abstract

A device 20 and a method for sound-based environment detection by means of echoes are proposed. The method comprises the steps: injecting an electrical signal into a sound transducer 1, thereby causing an emission of an acoustic measuring signal 2 via the sound transducer 1, determining a beginning and an end, of a dying-out dominance region with the aid of stored quantities from a data store 15 which are assigned to the sound transducer, picking up a first electrical signal r(t) of the sound transducer within the dying-out dominance region, determining characteristic properties, in particular of an impulse response and /or of a corresponding transform of a transfer function, of the sound transducer 1 from the first electrical signal r(t). The method may allow a statement to be made about the state of the sound transducer and/ or the reliability of a system using the sound transducer. The steps may be carried out within a measuring cycle for environment detection. The transducer may be installed in a bumper 5 of a means of transport.

Description

Device and method for sound-based environment detection

Prior art

The present invention relates to a device and to a method for sound-based environment detection by means of echoes.

In particular, the present invention relates to improvements in the analysis of the dying-out region for investigating transducer properties and detecting echoes in the transducer signal.

In echo-based environment detection, electrical signals are converted via an electroacoustic transducer into acoustic signals and emitted into the environment. The echoes of the acoustic signals reflected by environment objects are converted into electrical signals via a second electroacoustic transducer and subseguently analysed to detect the environment object or its distance. Often the same transducers are used to emit the acoustic measuring signals and to receive the reflected echoes. After emission of the measuring signals, the diaphragm of the transmitting transducer gradually settles. The many-times lower strength of the echoes arriving at the transducer in the same time range usually reguire a timely dying-out cf the dying-out signals of the transducer and a suitable identification of the echoes with respect to the dying-out signals. If the transducer properties are known, the profile of the dying-out can be predicted. For example, by means of a difference method the component of the dying-out signal in the transducer signal can be eliminated, in order to be able to detect also objects in the vicinity of the transducer. In

the context of the following description, the term

"transducer properties" is understood as the sum of those features which influence the behaviour of the transducer.

This behaviour is influenced not only by the transducer itself, but also by its electrical wiring and the analogous components surrounding it, such as for example the transformer and acoustically relevant components, such as e.g. a mechanical periphery (bumper or the like) . As regards the signal processing, several forms of representation, convertible into one another, of the transducer properties are known to a person skilled in the art. These include in particular: -the impulse response of the transducer or its step response, -the transfer function of the transducer e.g. as the transform of the impulse response (particularly common transform methods in this regard are the Fourier transform, the Laplace transform, the Walsh transform etc.), -the equivalent parameters according to a model which describes the behaviour of the transducer (e.g. by means of coils, capacitors, resistors etc. or/and as a spring-mass system) -parameters of the impulse response or of a quantity derived from the latter, such as for example a pole-zero plot, -time signal e.g. as a sequence of sampled values.

If the transducer properties are known, the post-pulse oscillation threshold can additionally be appropriately tracked, as described in DE 102012221591 Al. A further aspect of the utilisation of the transducer properties is the monitoring of the transducer. Thus, it is known that dying-out which is too slow or too quick can be detected by means of a time measurement at a threshold value detector and taken into account in the assessment of the system reliability. Moreover, a more detailed determination of the transducer properties would be desirable, in order to be able to better assess the reliability of the transducer.

For example, it could be detected whether coverings of mud, ice or snow influence the detectability of the transducer and/or whether the transducer is damaged e.g. by stone impact.

The disadvantage of the methods customary today is that the evaluated signals cannot be clearly distinguished from the echo signals, since the measurements of the dying-out of the transducer take place in the same signal strength region in which echoes can also arrive. Thus, in the prior art, a longer dying-out is measured only in the cases in which no object can be situated permanently in front of the transducer. This is to be assumed, for example, when travelling at a certain minimum speed. However, it is desirable to determine the transducer properties immediately on functional startup of the transducer without time delay, in order to be able to ensure also measurements carried out early, with respect to their reliability. Thus, even when a vehicle is still stationary, the state of the sensors should be able to be analysed in order to optimise, for example, their proximity measurement capability, to detect damage of the sensor or the presence of coverings.

For this purpose, it is necessary to eliminate from the arriving signals within an echo cycle the region which is characterised predominantly by the transducer properties.

DE 10 2010 003 624 Al discloses a frequency measuring method for checking transducer properties and assessing its readiness for use.

Disclosure of the invention

A basic idea of the present invention consists in that, already after a transmitting excitation of a transmitting transducer has been switched off, the electrical transducer signal is or may be a superimposition of signals (echoes) picked up from the transducer environment and of the dying-out signal owing to the electrical excitation. For a certain time range, however, the strength of the dying-out, in a, according to the invention as "dying-out dominance region", is markedly greater than the strength of an echo arriving at the transducer in the same time range. The dying-out dominance region is that time range in the transducer signal (electrical signal at the electrical connections of the (sound) transducer) which lies between the end of the injecting of an electrical measuring signal to be emitted and the end of the dying-out dominance region. According to the invention, the latter is reached when the transducer signal is permanently lower than the loudest echo to be expected in real operation. The loudest echo to be expected in reality is typically that of a plane object which is a very good reflector (e.g. a wall) . The value of the loudest echo to be expected in reality is dependent on the echo propagation time and decreases continuously with increasing echo propagation time. Since the manufacturing and climate-related tolerances of the echo strength (e.g. echo amplitude, sound level, or the like) are markedly lower than the dynamic range of the dying-out, the transducer signal at least in the centre of the dying-out dominance region is dominated so strongiy by the profile of the dying-out signal that, based on the signal profile in the centre of the dying-out dominance, the characteristic features for determining the transducer properties (see above) can be determined already within a single echo cycle. It is clear to a person skilled in the art that the dying-out dominance region depends on the parameters of the chosen object which is a good reflector, and the boundary regions of the dying-out dominance region can be influenced by not insignificant echo conponents. In the context of the present invention, these boundary regions are therefore called "transition regions".

The insights described above are used according to the invention in a method for sound-based environment deteotion by means of echoes. Tn this method, in a first step an electrical signal is injected into a sound transducer, whereby an emission of an acoustic measuring signal takes place via the same sound transducer. To check the transducer properties, subseguently a beginning of a dying-out dominance region within the transducer signal is determined by reading out stored quantities from a data store whioh are assigned to the sound transduoer. In other words, information is stored in the data store at an earlier point in time (e.g. in the course of the manufacturing process), via which information the dying-out dominance region within the transducer signal can be identified. For example, the dying-out dominance region can be defined in dependence on the instant at which the excitation of the transducer is switched off. Alternatively or additionally, a strength of the transducer signal can be predefined for the limits of the dying-out dominance region, for example based on the time profile of the transducer signal strength during the echc cycle in particular with respect to the strength of the strongest echo to be expected in real operation. A transducer signal present within the dying-cut dominance region (first electrical signal of the sound transduoer) is subsequently picked up and used to determine characteristic properties of the sound transducer. The characteristic properties can be, in particular, a transfer function, an impulse/step response, a transform of the transfer function or the like.

Additionally or alternatively, equivalent parameters of a model of the transducer can be determined with respect to their values. For a person skilled in the art, descriptions as spring-mass system or as series resonant circuit are common as customary models, without however excluding alternative modelling forms. The time signal or its profile, its envelope or the like can also be used to describe the characteristic properties. According to the invention, an early investigation of the transducer properties of the sound transducer is made possible, in particular already within the current echo cycle. Merely an excitation of the sound transducer by means of an electrical signal, the reading-out from a data store and the reaction of the transducer within the dying-out dominance region are reguired to determine the characteristic properties of the sound transducer. In this way, defects can be detected early in an "operating cycle" of the sound transducer and thus measuring results can be validated. Furthermore, an evaluation of the transducer signals up to the end of the dying-out dominance region at least regarding the evaluation of any echoes present in the transducer signal can be suspended, whereby a false detection of echoes can be prevented and computing power saved.

The subclaims show preferred developments of the present invention.

Preferably, a model of a dying-out signal or of the sound transducer used can be produced based on the characteristic properties of the sound transducer and the injected electrical signal. In other words, with the aid of the transducer signal picked up during the dying-out dominance region, the "pure" dying-out signal without echoes can be inferred. Subsequently, the model of the component in the transducer signal caused by the dying-out can be used in the course of an echo detection in the real transducer signal. For this purpose, for example a real transducer signal can be reduced by the model of the dying-out signal of the sound transducer, in response to which the echo possibly present in the transducer signal remains as the result. In this way, a received echo does not necessarily need to have a higher strength than the dying-out signal.

Accordingly, the minimum reliably detectable distance to environment objects is reduced, whereby the detection reliability of systems for environment detection configured according to the invention is increased.

Preferably, a statement about a state of the sound

transducer and/or a reliability of a system utilising this sound transducer can be made, based on the characteristic properties of the transducer or with the aid of the model of the dying-out signal or of the transducer. Depending on the configuration of the model, for example value ranges for equivalent parameters used can be stored in the aforementioned data store and compared with the current model to assess the functionality of the sound transducer.

Alternatively or additionally, characteristic points or properties of the transfer function can be compared with stored values (e.g. by known methods of curve sketching) To determine the characteristic properties of the transducer, use may be made of a time profile of the transducer signal such as its strength and in the case of spectral analysis in particular its amplitude(s) and alternatively or additionally its phase(s) and alternatively or additionally its frequency(ies) of the signal of the transducer picked up within the dying-out dominance region. Alternatively or additionally, a strength, represented by an envelope, of the transducer signal can also be determined (e.g. by a rectification and lowpass filtering) and taken as a basis for the evaluation.

The use of transducer signals, which are available or present early, within a measuring cycle (period of time between two successively emitted measuring signals) enables the investigation of the transducer or its modelling and the use of the investigation results or the model within fewer measuring cycles, in particular within the same measuring cycle. In this way, extremely up-to-date findings regarding the functionality or the operating state of the transducer can be inferred in the environment detection. As a result, early assistance of the user is made possible by a system configured according to the invention.

The model of the transducer or of the dying-out signal produced on the basis of the characteristic properties can be a computationally represented or an electrically realised dying-out signal. Depending on the configuration of the system used for carrying out the invention, it is thus possible to produce a suitable reproduction of the dying-out signal which involves minimised outlay and is perfect in terms of tolerance. The use of an electrically realised signal model enables, for example, a comparison of the model with an actual transducer signal carried out by signal processing. If a digital or computational investigation or evaluation of the transducer signal is to be performed, a computationally represented model of the dying-out signal can entail advantages. In particular in the case where digital signal processing steps are taken into account for the echo evaluation or environment detection, a corresponding processing of the model may be advantageous.

The producing of the model of the dying-out behaviour e.g. in the form of the dying-out signal can comprise, in addition to the use of an electrical transducer signal received during the dying-out dominance region, also an electrical signal used to excite the sound transducer. In particular, a convolution of the excitation signal with a (previously known) impulse response of the sound transducer can in this case be used to predetermine certain model parameters, so that environment influences and any damage can be parameterised or identified within the model. For example, temperature dependences and other influences on the transducer function can be separately reproduced within the model and likewise separately updated at a later point in time. In some cases, this obviates the need for a comprehensive revision or redetermination of the model.

To refine the model determined as above, in a second, temporally following measuring cycle, an additional parameter or an additional characteristic property of the sound transducer can be determined or uncertainties of the determined model can be reduced or the model oan be updated. Based on this model, a second, refined model of the dying-out signal or the sound transducer can be produced and the model used for identification of the echo present in the electrical signal can be adapted using the second model. Furthermore, the model of the transducer properties can be improved and/or updated by the evaluation of the echo signals within an echo cycle and even more so over a plurality of echo cycles. If, for example, with the aid of the echo signals an object model and its movement behaviour is predicted and if this model is sufficiently validated, the remaining signal component characterised by the dying-out can be inferred. The validation of the characteristic properties determined in different ways can preferably be carried out with a guality criterion, such as e.g. the variance. If different models and/or sets of parameters are possible as a solution, the validity can be decided using quality criteria. Thus, the ime signal during the excitation of the transducer is also characterised by its characteristic properties, as described in DE 102012200743 Al. Likewise, the results of other methods, such as e.g. the method described in DL 102012221591 Al, in which the transducer properties can be detected only in specific operating states, can be taken into account as well in the validation of the transducer properties determined according to the invention. The use of information and findings obtained during following measuring cycles enables a more comprehensive (more time-consuming) investigation of the signals or the use of computationally less powerful hardware, which may be more cost-effective. In addition, as described above, parameters which have changed in the course of the operation can be i:i taken into account and represented in this way. It can thus be assumed that short-time changes of the signal at the transducer is the consequence of an object scene determining the echo signals, provided that the signal changes are similar to the expected echo signals or longer-lasting changes in the signal in particular while objects are moving are with higher probability the consequence of changed transducer properties.

According to a second aspect of the present invention, a device for sound-based environment detection is proposed.

This device comprises a sound transducer, a signal generator and an evaluation unit. The signal generator serves to generate an electrical signal, by means of which the scund transducer can be excited to emit an acoustic measuring signal. The evaluation unit serves for the above-described signal analysis, for which optionally also a data store can be provided in the device. Alternatively, the data store can be provided externally for access by the evaluation unit. The sound transducer is configured to transmit and receive sound signals into the environment of the device. In other words, the sound transducer can at a first moment be configured as a transmitter and at a second moment as a receiver (of an echo belonging to the emitted signal or echoes caused by other sound transducers) . The device according to the invention is configured by the above-described construction to carry out a method as described in detail in connection with the first-mentioned aspect of the invention.

Thus the device can include a threshold value which is provided by means of a data store and is applied to a filter output signal, in particular when the filter output is characterised by the signal strength, the preferably propagation-time-invariant threshold value being set so that it is intersected within the dying-cut dominance region by the envelope curve of the transducer signal, and the time duration of the intersection point(s) is a measure of the transducer properties, in particular of the time constant of the dying-cut or of the positions of the resonant freguencies in the case of multiple intersection within an echo cycle.

Thus the device can include a propagation-tine-dependent threshold value which is provided by means of a data store and is applied to an output signal of a receive filter and which is designed, with the aid of the transducer properties obtained, so that it is not exceeded in the exclusive presence of dying-out signals at least in certain sub-regions, but an additional occurrence of echoes in the signal causes the propagation-time-dependent threshold value of the output signal of the receive filter, which value is characterised by the transducer properties, to be exceeded.

Thus the device can include a propagation-time-dependent threshold value which is provided by means of a data store and is applied to an output signal of a receive filter and which is designed, with the aid of the transducer properties obtained, so that it is largely exceeded in the exclusive presence of dying-out signals at least in certain sub-regions, and with the aid of the tine response (duration/frequency) of possibly occurring short-term instances of falling below this threshold value, the characteristic properties of the transducer are determined, because, for example, the short-term instances of falling below the threshold value are caused by the frequency positions of the series resonanoe of the transduoer and the circuit surrounding the transducer and forming a parallel resonance.

Thus the device can analyse the transducer signal with regard to time profile of signal strength and dominant frequency component or phase component, in order to infer the transducer properties.

Thus the device can, with the aid of the transducer properties obtained, in subsequent echo cycles cause a reduction of the dying-out, for example by controlling an excitation of the transducer counteracting the dying-out and/or by synthesising a signal corresponding to the dying-out and subtracting it from the original signal.

The device can, for example, be a constituent of an environment sensor system for automotive use. In this regard, the transducer properties are, if not exclusively, at least partially, determined by the periphery of the sound transducer. This may, for example, be a covering of a bumper of the means of transportation configured according to the invention.

Brief description of the drawings

Exemplary embodiments of The invention are described in detail below with reference to the accompanying drawings.

In the drawings: Figure 1 is a schematic view of components of an exemplary erThodiment of a system for environment detection configured according to the invention; Figure 2 is a signal flow diagram of an exemplary embodiment of a system for environment detection configured according to the invention; Figure Ba is a basic representation of sections in a time signal of a sound transducer in an exemplary embodiment of a system for environment detection; Figure 3b is a practical realisation of an envelope of a dying-out transducer; Figure 4 is a time diagram of a simulation of a sound transducer signal in an exemplary embodiment of a system for environment detection according to the invention; Figure 5 is a time diagram of a simulation of a sound transducer signal in an exemplary embodiment of a system for environment detection containing an echo received from the environment; and Figure 6 shows a flow diagram illustrating steps of an exemplary embodiment of a method according to the invention.

Embodiments of the invention Figure 1 shows a system 20 for sound-based environment detection, which has an ultrasound transducer 1 as the sound transducer, installed in a bumper 5 of a means of transportation. The ultrasound transducer 1 is configured to emit a measuring signal 2 into the environment of the system 20. A wall W represents an object which is the best reflector in the real operation of the system 20 and which generates a strongest echo 2' at a given distance d between the wall W and the ultrasound transducer 1. Via a signal generator 3, the ultrasound transducer 1 can be excited to emit the measuring signal 2. The electrical time signals present at the ultrasound transducer 1 are received via a microprocessor 4 as the evaluation unit and compared with references stored in a data store 15. In addition, the microprocessor 4 can store in the data store 15 models and their parameters, in order to be able to use them at a later time.

Figure 2 shows a signal flow diagram of a system 20 for environment detection according to the invention, in which a signal generator 3 is supplied with a data-related representation s(t) of a measuring signal. The measuring signal is transmitted via the output of the signal generator 3, on the one hand, to the ultrasound transducer 1, and, on the other hand, as described below, is provided for different instances of the signal processing. The output signal r(t) of the ultrasound transducer 1 arrives at the input of an estimator 6, by which a model for the dying-out signal and thus for certain operating ranges of the ultrasound transducer 1 is created. For this, there can be provided -as represented by a dashed line -the representation s(t) of the measuring signal, whereby the determination of the transfer function of the ultrasound transducer 1 can be taken lntc account by the estimator 6.

The model Ml created by the estimator 6 is supplied to a synthesis unit 7, which additionally also receives the representation s(t) of the measuring signal. The synthesis unit 7 generates, through the aforementioned input quantities, an ideal dying-out signal r.1(t) which is supplied to a signal evaluation unit 8. By additionally also supplying to the signal evaluation unit 8 the real transducer output signal r(t), which optionally contains an echo of the measuring signal emitted by the ultrasound transducer 1, it is possible for a subtraction of the idealised transducer signal r(t) from the rca: transducer output signal r(t) to take place and for the difference to be detected as a signal which has arrived at the transducer independently of the dying-out signal or be subjected to further investigation. Tt is known to a person skilled in the art that, besides the representation, described here, of the transducer model Ml in the form of the time signal of the dying-out, it is also possible to use equivalently alternatives, such as e.g. a parameter range of a transfer function or a transformed quantity or its parameter or equivalent parameter of a model made of linear or nonlinear components. A wide range of forms of the signal evaluation are kncwn tc a person skilled in the art frcm the literature. Thus, the form, described here, of the application of the transducer properties constitutes only one of many different possibilities. For example, the signal evaluation 8 can be applied to a phase-sensitive or non-phase-sensitive filter output signal and accordingly the requirements and ways of representation of the model Ml synthesised by way of example or of the synthesised signal r4(t) can change.

In a preferred form, the dying-out signal is reconstructed as faithfully as possible to the actual dying-out and with correct signal strength and phase and is subtracted from the input signal r(t) in order that the remaining signal component oontains as exclusively as possible eoho signals.

In an alternative preferred form, the output signal of a receive filter, such as e.g. of a non-phase-sensitive matched filter for chirp signals, is flanked with threshold values such that the signal exceed or falls below these threshold values only on occurrence of signal components caused by reflected echoes and in this way the echo propagation time and thus the object distances can be inferred from the instants of exceeding or falling below.

Figure 3a shows the output signal of a receive filter such as that of a matched filter, in particular the stylised envelope 10 of a transducer signal and the stylised envelope 11 of a maximum environment echo occurring in reality. A time range I without electrical excitation of the electroacoustic sound transducer is followed by a time range II, in which the transducer is excited with an electrical signal for emiting a measuring signal. At the beginning of a third time range III, the excitation is changed, in particular greatly changed, for example switched off, in response to which the envelope of the transducer signal 10, considered logarithmically, falls linearly with respect to time (or with respect to the object distance d) . The dying-out region III is divided into three regions (IV, V. VI), of which a first is the optionally occurring excitation-side transition region IV to the second dying-out dominance region V identified according to the inventicn and following the latter a transition region VI to the echo dominance region VII. Tn this case, the dying-out dominance region V is characterised in that the strength of the dying-out signal is with certainty higher -for example at least by a factor of 2 -than the strength of the strongest expected received echo (of. envelope 11) . In a region around the point 9 at which identical values of the envelope 10 prevail for the dying-out signal and the envelope 11 of the echoes, the dying-out region III merges into the echo dominance region VII. Tn the echo dominance region, the strength of the weakest expected echo (not plotted) is greater than the strength of the dying-out signal 10 -for example at least by a factor of 2. The use, according to the invention, of the dying-out dominance region V and the optional modelling for using the findings obtained in this region enables an echo identification in principle even before the point 9 where the echoes definitely have lower strengths than the dying-our signal.

While Fig. 3a shows in the dying-out region ITI a continuously decreasing envelope curve, Fig. 3b shows, by way of example, an alternative envelope curve 10 of the transducer signal, caused by changed transducer properties.

The short-time dips, visible in the envelope curve of Fig. 3b, of the otherwise continuously decreasing envelope curve 10 may, for example, be the conseguence of particular transducer properties, such as for example the position of the series resonance frequency of the transducer and of the parallel resonance freguency of the circuit surrounding the transducer.

By way of example, an evaluation method by means of the threshold value S is shown in Fig. 3b. Within the dying-out dominance ITT, the threshoid value S runs somewhat abcve the strength of the time profile of the maximum echo II occurring in reality. Based on the last time when the filter cutput signal lOb falls below the threshold value S at the instant TN, the threshold value S from this instant onwards is converted to an echo detection threshold value, by the threshold value S from the instant N upwardly flanking the time profile of the dying-out. If an additively superimposed echo occurs, this threshold value S would be exceeded by the filter output signal 10 and based on the exceeding instant the echo propagation time characterised by the object distance could be inferred.

Figure 4 shows a real-measured rectified transducer signal 16, with the designation of the time (or distance) regions corresponding to Figure 3 being retained. The representation shows a saturation of the transducer signal 16 in the regions II and at the beginning of the dying-out region III, which also occurs (dynamically induced) in systems known in the prior art. In the dying-out region III, an idealised envelope 10 of the transducer signal 16 is therefore additionally drawn in for purposes of illustration. The saturation region is overcome only from an object distance of approx. 5 centimetres, so that here the rectified transducer signal 16 with respect to its strengths closely follows the idealised envelope 10. An envelope 11 of a maximum environment echo occurring in reality is likewise plotted. From this it becomes clear that in this case the envelope of the transducer signal 16 is dominated by the dying-out in the distance range from approx. 2 cm to approx. 7 cm. Despite the overdrlvlng at the beginning of region ITT, a transducer signal 16 is thus available in the dying-out dominance region. In this region, an essential transducer property, the time (or the equivalent distance) profile of the envelope of the dying-out, could thus be estimated. As a simple form of utilisation of the estimated transducer properties, the derivation of the profile of an eoho detection threshold 12 could subsequently be realised. Just above the idealised envelope 10, an echo detection threshold 12 is drawn in as a possible form of a signal evaluation 8, which threshold has to be exceeded by the transducer signal 16 for the signal-strength-based detection of echoes. Since in Figure 4 exclusively the profile of the transducer signal 16 is shown during the emission of the measuring signal and the subsequent dying-out, i.e. without the presence of an echo, in a following time region VII merely noise components 13 without environment echoes are present in the transducer signal 16.

Figure 5 shows the transducer signal 16 represented in Figure 4, in which signal an echo 14 of an environment object is additionally present. The reflecting surface of the environment object causes, owing to a corresponding echo, an additional occurrence of an increased strength of the transducer signal between 10 and 20 centimetres distance. Since the echo chosen by way of example is the strongest echo occurring in reality, the likewise drawn-in profile of the strength of the maximum echo occurring in reality intersects with the component of the transducer signal, caused by the echo, at approx. 15 cm. If one follows the profile of the echo strength of this strongest reflector at different reflector distances in an equivalent manner, it is thus possible to infer the profile of the maximum echo strength occurring in reality beyond the region III.

As already explained in connection with Figure 3, an echo detection threshold 12 can be derived as a simple realisation of the signal evaluation 8. Based on the exceeding of the echo detection threshold 12 by the transducer signal, the existence of a reflecting surface can be subsequently inferred. Approximately at ten centimetres, the echo exceeds the detection threshold 12.

At earlier instants, the echo signal dips into the dying-out signal 16. By subtracting the dying-out signal 16 in Figure 5, however, a transducer signal already increasing before the ten-centimetre mark would, according to the invention, enable an echo detection. In this way, information about the environment object could be obtained and evaluated at an earlier instant.

For the sake of completeness, it should be mentioned that, besides the direct utilisation of the trans!ducer model obtained in the dying-out dominance region 5 shown here, further forms of evaluation of the echo signal which are known to a person skilled in the art are also common. By way of example, mention may be made here of the evaluation by means of a matched filter. For this purpose, to determine the echo propagation time, mostly an incoherent filtering, i.e. not taking account of the phase position, is carried out. In this case there results, in an equivalent manner, at the filter output a signal which is, equivalently to the signal strength, a measure of the similarity of the respective transducer signal to the signal expected by the filter.

Also worth pointing out in this regard, however, is a mixed form of the signal analysis in which the model of the transducer properties is effected by coherent evaluation, i.e. evaluation sensitive to the phase position, while the subsequent utilisation of the transducer model Mx, for example in detection of the echo propagation time, nowadays mostly makes do with the incoherent filtering which is not sensitive to the phase position.

Figure 6 shows steps of an exemplary embodiment of a method, according to the invention, for sound-based environment detection by means of echoes. In this method, in step 100 an electrical signal is injected into a sound transducer, thereby causing an emission of an acoustic measuring signal via the sound transducer. Tn step 200, a beginning, in particular also an end, of a dying-out dominance region is determined with the aid of stored quantities from a data store. In step 300, a first electrical signal is picked off at the electrical connections of the sound transducer within the identified dying-out dominance region. Since the dying-out dominance region has low susceptibility to external influences, in step 400 characteristic properties of the sound transducer can be determined from the first electrical signal. These serve, for example, to produce a transfer function or corresponding information characterising the sound transducer. In step 500, there is produced a model of a dying-out signal of the sound transducer based on the characteristic properties of the sound transducer and the injected electrical signal. This enables in step 600 the use of the model of the dying-out signal in the identification of an echo present in the electrical signal of the transducer. In a further (subseguent) measuring cycle, in step 700 additional parameters of the transfer function of the sound transducer are determined, in step 800 a second model of the dying-out signal of the sound transducer based on the additional parameters is produced and in step 900 the model of the dying-out signal used for the identification of an echo present in the electrical signal is adapted by means of the second model. In this way, a more time-consuming evaluation of the transducer signal or a more time-consuming production of the model can be used. In addition, parameters which vary with time or any defects which have occurred and other findings can be taken into account in the second model.

In the context of the present invention, a threshold value 11 which is constant or decreases with progressing echo propagation time can be generated and for detection of an echo a check can subsequently be made as to when the actual transducer signal falls below this threshold value U the last time or exceeds the threshold value 11 the first time.

In doing so, it is assumed based on preliminary investigations that within the first-time exceeding or last-time falling below the transducer signal, the transducer signal is as far as possible independent of external influences, such as e.g. echoes, but is determined substantially -taking into account the exciting measuring signal -by the properties of the transducer itself. The threshold value 11 which is constant or decreases with progressing echo propagation time is known already at the beginning of the echo cycle (e.g. as a time profile stored in a data store) . It can preferably be formed from the profile of the echo peak strengths of the object which is the best reflector in reality at different object distances d (for example as 1.5-times the envelope curve of the maximun echo occurring in reality on the respective excitation by the measuring signal) . For this purpose, in the preliminary investigations the profile of the maximum echo strength occurring in reality is determined, for example by experiment within the echo dominance VII. By means of a model, such as e.g. the simple model R1(d) = by determining the model parameters K and n, the profile of the threshold value ii within the region III can be subsequently inferred. It can, however, be measured specifically and stored in a data store 15 during manufacture. Alternatively, experimental values which have shown good results irrespective of manufacturing tolerances can be stored in a data store 15.

A person skilled in the art knows that the characteristic properties of a transmission system (e.g. transducer including periphery) can be represented in different ways equivalent to one another. Examples of this are the time profile of the impuise response or parameters of an equation describing the impulse response, and also the transform of the impulse response. A transform of the transfer function or equivaient parameters of an eiectrical network representing the transmission system can also be used to formulate the characteristic properties. A person skilled in the art also knows how the concrete reaction of a transmission system to input signals, such as e.g. to differently injected excitation signals, can be determined by means of the characteristic properties. Merely by way of example there is mentioned, in this regard, the convolution of the impulse response with the time profile of an excitation signal or with the time profile of an injected measuring signal pulse, in order to determine the transducer signal during the excitation and the subsequent dying-out. A person skiiied in the art also knows that an electrical connection is aiways characterised by two conjugate guantities, such as e.g. current and voltage.

Thus, the system response of a transducer having characteristic properties, at both electrical connections of which a current is injected for excitation to emit a measuring signal, is for example preferably the time profile of the voltage of the electrical connections.