WO2015192156A1 - Ultrasound sensor and method for locating objects - Google Patents

Abstract

The invention relates to an ultrasound sensor (26, 37, 41) and to a method for locating objects (7, 8, 17) in space by means of such an ultrasound sensor (26, 37, 41), which ultrasound sensor comprises at least four ultrasound receivers (27, 28, 29, 30), wherein an ultrasound signal (42 or 9) is emitted by an ultrasound transmitter (43 or 41) that is external or that is comprised by the ultrasound sensor (26) and an ultrasound signal (11) reflected at at least one associated reflection point (33, 34, 35, 36) is received by the ultrasound receivers (27, 28, 29, 30) of the ultrasound sensor (26) and is evaluated in order to obtain at least one object position, characterized in that, in the case of a plurality of object positions obtained in parallel, the object positions are selected in accordance with a selection criterion, which is defined in that the at least four reflection points (33, 34, 35, 36) each associated with the object position lie approximately in a common plane in space.

Description

 Ultrasonic sensor and method for locating objects

The invention relates to a method for locating objects in space with the aid of an ultrasound sensor and a corresponding ultrasound sensor having at least four ultrasound receivers, an ultrasound signal being radiated from an external transmitter or ultrasound sensor and each reflected by the ultrasound receivers in at least one associated reflection point Received ^¬ ultrasonic signal and evaluated to obtain at least one object position in an evaluation unit, which is connected to the Ultraschallemp ^¬ catchers.

In particular, the invention relates to an ultrasound sensor or a method implemented by the ultrasound sensor for reliable location of three-dimensional object positions (SD localization) relative to the position of the ultrasound sensor on the basis of three-dimensional reflection points in space by means of ultrasound. The invention aims primarily at any non-concave objects. Reflection point here is a point on an object to be located, at which the ultrasound signal emitted by the ultrasound transmitter is reflected, before it reaches one of the ultrasound receivers. A reflection point is therefore the point on the surface of the object in which the reflection ^¬ onsbedingung for a particular ultrasonic transmitter and a particular, namely the associated ultrasonic receiver (ie, a particular transmitter-receiver pair) is met. Therefore, in ERAL ^¬ nen the recipients associated reflection points are not identical and there are so many per object reflection points as receiver. The ultrasonic sensor is designed to measure the transit times of an ultrasound signal emitted by the ultrasound transmitter and reflected in the reflection points and received by the ultrasound receivers. From the measured transit times, the distance between the ultrasound sensor and the reflection points, or, strictly speaking, the sum of the distances between the ultrasound transmitter and the reflection point and the reflection point and ultrasound receiver, can be derived at a known sound velocity. The ultrasonic sensor can basically be used as an active sensor device with its own ultrasound sensor. traschallsender or be designed as a passive sensor device without its own ultrasonic transmitter. The ultrasonic receiver or the ultrasonic transmitter are arranged not only for structural reasons, but also to achieve a corresponding spatial resolution, at intervals from each other. In general, different reflection points are therefore present in the case of several ultrasound transmitters and / or a plurality of ultrasound receivers for the same or the same object. Each reflection point is associated with a specific transmitter-receiver pair, wherein for the sake of simplicity in a single transmitter can be spoken by an assignment to the receivers and vice versa.

It is well known, by means of ultrasonic sensors, the propagation time of a first reflected ultrasound signal (nachfol ^¬ quietly also called short echo) to be measured and therefrom under An ^¬ assumption that the origin of the echo is situated on an acoustic axis of the sensor, this origin, which corresponds to the Reflection point of the echo corresponds to locate. However, such sensors are unsuitable for a true three-dimensional positioning, as no infor ^¬ mation is obtained about the direction of the echo. In addition, these sensors typically to a very narrow field of view are turned ^¬ limits (that are received echo only from a narrow solid angle) to any errors in the assumption that the origin of the echo is situated on an acoustic axis of the sensor to re ^¬ reduce.

In addition, sensor devices for the 3D location of objects with multiple transmitters and / or receivers have already been proposed. To obtain the necessary direction information, at least three running time measurements are required. Therefore, for example, sensor devices for gesture recognition by means of ultrasound, as have been proposed for mobile terminals, as a rule have a plurality of transmitters. For example, the shows

WO 2011/048433 AI a system with multiple transmitters and receivers to determine multiple echoes and to calculate the 3D position information of a hand movement. Furthermore, Kanaki et al. (Kaniak, G., Schweinzer, H., "A 3D Airborne Ultrasound Sensor for High-Precision Location Data Estimation and Conjunction," Instrumentation and Measurement Technology Conference Proceedings, 2008. IMTC 2008. IEEE, pp. 842, 847, 12-15 May 2008, doi: 10.1109 / IMTC.2008.4547154) already described a sensor device with four ultrasonic receivers (microphones) for locating the reflection points.

The known devices and methods have in common that they with multiple echoes and / or superimposed echoes, as well as with situations in which the ektpositionen significantly differ from the used in the location, ie in the calculation of the ektpositionen, models, not in an appropriate manner to be able to deal with. In particular, difficulties arise when the running time differences between two echoes on different objects are small. For more than one object, therefore, it can often not be clearly determined which echoes belong together and must be combined for the calculation of the object position. Those objects that are farther apart than the maximum distance between two receivers may still be relatively easily separated, since the time differences between the echoes for a single object may not be greater than predetermined by the distance between the receivers. If the objects are together but closer and have a very similar distance from the sensor device, it is difficult to determine which echoes zusammengehö ^¬ ren, that have their origin or point of reflection on the same object. To circumvent this problem, elaborate methods have already been proposed in which, for example, an assignment is to be reconstructed on the basis of the phase and amplitude information of the echo. For simpler methods, only those combinations are rejected for which there is no geometric solution for a corresponding object position. However, this does not resolve all ambiguities, especially if more than three run time measurements are available or can be combined.

As long as the ultrasound receivers are relatively close together angeord net and all reflection points are on the same object, the object position based on the reflection points can be relatively ge ^¬ accurately determined or made by the use of different reflection points error - depending on given Accuracy requirements and geometry of the objects (optimal are objects with flat or convex surfaces, whereas in concave surfaces with relatively larger errors must be expected) - neglected. However, if the Reflection ^¬ onspunkte are on different objects, this can lead to incorrect object positions. This is because the individual run time measurements are just distance measurements and do not contain directional information. The direction information can only be derived from the deviations between the individual transit time measurements.

It is an object to resolve the observed in the prior art disadvantages and provides an ultrasonic sensor and a method for operation of such an ultrasonic sensor to sheep ^¬ fen, which is accurate, reliable three-dimensional Ortungsda ^¬ th (in the form of object positions) of the invention, and thus the Processing and ultimately the practical use of the obtained location data. In particular, it is an object of the invention, with a single ultrasonic sensor parallel positioning of multiple objects in the widest possible solid angle while at the same time largely obj ektpositions- artefacts (incorrectly recognized object positions), which are due in particular to discontinuities in the object surfaces.

This object is achieved in a method of the initially mentioned type according to the invention, that in several object positions obtained in parallel these are selected according to a selection criterion, which is defined by the fact that the object position respectively associated at least four reflection points are approximately in a common plane in space. The approximation is particularly true when the deviations of the reflection points from a common plane are less than 5 mm, preferably less than 1 mm, in particular less than 400 μπι, less than 200 μπι, or less than 100 μπι. While the deviations for plane surfaces are independent of the distance, distance-dependent limit values can also be used for previously known object geometries (eg in the case of a sphere, the farther the sphere is from the sensor is, all the reflection points closer together, since only a very small angle is necessary to reach the distance of the receiver / transmitter at a greater distance; accordingly, the deviation of the reflection points from a plane decreases with increasing distance). It is obvious to the person skilled in the art that, depending on the requirements of the particular application and accuracy of the components used and, if appropriate, on the distance and on the sensor dimensions, a correspondingly adapted tolerance range for the approximate criterion is selected. The reflection points are in turn assigned to different transmitter-receiver pairs or-in the case of a single transmitter-to different receivers. For selection, the reflection points themselves are not determined as the three-dimensional space ^¬ dots, but it is only checked whether for the reflection points corresponding measured transit times and a solution exists for the corresponding distances and taking into account the reflection condition, in which the characterized by the distances Reflection points lie in a plane. Since the selection criterion for moving objects (or a moving sensor) is generally met only at one point in time, it is essential, at least in these cases, for the measurements to be carried out at a time, ie simultaneously and simul ^taneously , so that the received and reflected for reconstructing an object position

Ultrasound signals go back to us the same radiated ultrasonic ^¬ sound signal. The requirement of the minimum number of four measurements arises naturally, since with only three measurements it is always possible to find a plane which satisfies the reflection condition for the three corresponding reflection points. The advantage and success of the proposed Se ^¬ lesson criterion is based on the fact that this is the one for recognizable objects no significant limitation because the reflection points due to the sensor size on the Ob ^¬ ject are geometrically so close together that even a partially approximately flat surface to Satisfying the criterion. It is also highly unlikely ^¬ scheinlich On the other hand, that the reflection points of different objects are to ^¬ due in a common plane. For these reasons, for example, by means of the specified selection criterion Recognize artifacts, which result from an association of reflection points on different objects and thus avoid. As a result, the positioning method becomes more robust as a whole than distortions in the form of false or only apparently plausible object positions.

Accordingly, the object stated above is according to the invention in a Ul ^¬ traschallsensor the initially defined type as ^¬ solved by that the evaluation unit comprises a selection module which is set at more gained parallel Objektpositio ^¬ nen for determining that object positions in which the object position respectively associated at least four reflection points are approximately in a common plane in space. The selection module implements the top of the to ^¬ connexion with the inventive method in detail explained selection criterion. The fact that the evaluation unit is set up to obtain at least one object position from the received ultrasound signals means, among others, that in addition to the selection module, all elements ^necessary for evaluating the received ultrasound signals, such as typically a computing unit, a Memory and - in the case of a LaufTeitmessung - a timer includes. The selection module may be realized and for example also be retrofitted to an appropriate sensor device or its evaluation as a separate Hardwareein ^¬ integral or, preferably, as a software module. In principle, any type of device for the ultrasonic transmitter or the ultrasonic receiver can be used. For example, a broadband electrostatic transducer as an ultrasonic transmitter and four or more small non-directional MEMS micro ^¬ phone can be used as an ultrasonic ^receiver .

In general, the present method does not aim to realize a flat surface, but on tune the position of the object or the position of an object corresponding to the equivalent reflection point on the surface of the object to be ^¬. For this reason, the condition that all

Reflection points in a plane, even if they are only approximately fulfilled, since in practice they only apply to certain object geometries. metric applies:

• in the case of a flat surface, it is clearly strictly met;

In the case of a spherical surface, it is generally not exact, but only approximately fulfilled, with the (virtual) plane intersecting the sphere; if the ball converges to a point, the condition is also fulfilled;

• in the case of a cylinder, two points may behave as if they were at one level, given their orientation. The other two reflection points behave like a sphere with a corresponding radius; Approximately, the condition is met again here as well.

In order to increase the robustness of the present method even further, it has been found to be favorable if the Ul ^¬ traschallsensor at least five ultrasound receiver, wherein the selection criterion is accordingly defined by the fact that the Whether ektposition each case at least associated five reflection points substantially in a common plane in space. Only with a fifth sensor can artifacts, which are not recognizable at certain symmetries with four sensors (for example, if in each case two reflection points lie on two different objects with parallel surfaces), be detected and avoided. Accordingly, it is also favorable in the present ultrasonic sensor if it has at least five ultrasonic receivers which are connected to the evaluation unit.

In order to facilitate further processing of the acquired location data or object positions, it is advantageous if object positions for which no common level can be found are recognized as artifacts and discarded. Those Objektpo ^¬ sitions where no common ground for the (hypothetical ^¬ rule) reflection points can be found to meet the

Selection criterion not and it is therefore advantageous to filter them early, for example, still in the ultrasonic sensor and not - like the other location data - eg output via a data interface.

A simple, computationally efficient and thus particularly rapid selection can be achieved if an even number of ultrasound receivers are arranged in one plane and in at least two mirror-symmetrical pairs with respect to a common center, wherein the selection criterion is defined by the averaged flight time of the receiver pairs in the Wesent ^¬ union matches. This criterion is of course equivalent to a comparison of the sums of maturities. The simplified in this case, a so-called mid-criterion selection criterion can therefore be checked by a simple means ^¬ value calculation of the running time for each receiver pair and then comparing the resulting averages. Accordingly, it is particularly advantageous in connection with the vorlie ^¬ constricting ultrasonic sensor when at least a portion of the ultrasonic receivers in a plane and in at least two relative to a common midpoint spiegelsymme ^¬ trical pairs is arranged. In contrast, has a

asymmetric arrangement of the receiver has the advantage that a He ^¬ recognition of artifacts in more situations than in a symmetrical construction is possible.

In this context, it has proved to be advantageous if, after checking the selection criterion based on the receiver pairs, a plane is determined based on the transit times measured by the receiver pairs and checked for a second part of ultrasound receivers which need not be arranged in mirror-symmetrical pairs whether these Ul ^¬ traschallempfängern associated reflection points are located in the determined plane, otherwise the selection criterion is generally not fulfilled. In the event that further receivers are provided other than the sym ^¬ trically arranged receiver pairs, the selection criterion can thus be evaluated in a two-step process, wherein the above-described and efficient verifiable center criterion is turned on ^¬ in a first step and then only ene combinations of Reflection ^¬ tion points are taken into account, which are the center criterion fulfill. The comparatively rchenaufwendigere review, even the measured by the non-symmetrically arranged Ultraschallemp scavengers with the determined level consis tent, takes place only for a comparatively small, because already preselected number of combinations. This method can be used in particular when using a fifth ultrasonic receiver together with four symmetrically arranged Ul traschallempfängern used.

A particularly efficient method for checking the Selekti ^¬ onskriteriums can be used when the receiver pairs have the same distance between the ultrasonic transmitter and ultrasonic receivers, with in particular the receiver in one of the number n of the ultrasonic receiver corresponding n-fold rotational symmetry are arranged, wherein an equivalent Reflection ^¬ onspunkt is determined for all the receivers on the basis of values measured by the Empfängerpaa ^¬ maturities and is checked for a second portion of ultrasonic receivers, which may not be symmetrical, whether these ultrasonic ^¬ receivers associated reflection points are consistent with the equivalent point of reflection, wherein otherwise the Selekti ^¬ onskriterium is generally not fulfilled. With the distance d Zvi ^¬ rule the ultrasound transmitter and the individual

 For ultrasonic receivers, the position of the equivalent reflection point can be calculated as follows:

 2 2 2

z = V r -x -y where Ii is in each case the distance corresponding to a transit time measured by the ultrasound receiver i. There being the equivalent reflection point is known, the expected Laufzei ^¬ th for the one or more ultrasound receiver can be calculated immediately. That is, the reflection points of the other receivers are consistent if and only if their distance is approximately equal to the distance of the equivalent reflection point.

The present method is particularly suitable for ene cases in which the ektpositionen of multiple objects from a single radiated ultrasonic signal (eg, from an ultrasonic sensor with a single ultrasonic transmitter) are determined. Such scene analysis are particularly affected by the problem of artifacts, so that the pre ^¬ troubled selection criterion enables significant improvements here.

Due to the high significance of the illustrated above Selekti ^¬ onskriteriums can during its use in receiving more than a reflected ultrasonic signal in a receiver, the object positions are determined by forming all possible combinations of reflection points, each reflection point is associated with a different ultrasonic receiver. The re ^¬ inflection points are not usually known as dreidimensiona ^¬ le points, but characterized only by their (measured) distance from the ultrasonic sensor. The naturally large number of combinations - which significantly increases with the number of existing objects - alone can be reduced by the Se ^¬ lesson criterion on the valid combinations, ie, additional, much more complex Selekti ^¬ mecha- nisms, such as taking into account the directivity or the amplitude and / or phase of the received

Echoes, omitted.

In general, and particularly in the context of combinatorial all reflection points, it is favorable if Objektpositio ^¬ NEN, in which the maximum distance between two respectively associated reflection points exceeds the maximum distance between two ultrasonic receivers are discarded prior to selection. Since this criterion is mathematically even easier to implement than even the central criterion shown above, it should be checked if necessary before all other selection ^steps . The invention will be explained below with reference to particularly preferred embodiments, to which it should not be limited, and with reference to the drawings. The drawings show in detail:

 1A shows an example of an active ultrasonic sensor according to the prior art with only three ultrasonic receivers and FIG. 1B shows a time diagram of the ultrasonic signals detected by the three receivers;

 FIG. 2 shows an application of an ultrasonic sensor according to FIG. 1 with objects arranged slightly offset; FIG.

 3 shows an active ultrasonic sensor with four ultrasonic receivers in an application according to FIG. 2;

 4A shows an active ultrasonic sensor with four ultrasonic receivers in an application according to FIG. 1A and FIG. 4B the associated time diagram, comparable to FIG. 1B;

 5 shows an active ultrasonic sensor with five ultrasonic receivers in an application according to FIG. 2;

 6 shows schematically an application of an active ultrasonic sensor with five ultrasonic receivers in an environment with retroreflectors;

 7 schematically shows an application of a passive ultrasonic sensor with five ultrasonic receivers in an environment provided with external ultrasonic transmitters;

Fig. 8 is a schematic plan view of an active ^¬ Ultra sound sensor of FIG. 5; and

9 is a schematic block diagram of an ultrasonic sensor ^¬ .

In Fig. 1A and 1B, the fundamental difficulty to a ^¬ order of the echoes from two different objects is illustrated. 1A shows schematically an active ultrasonic sensor 1 with three ultrasonic receivers 2, 3, 4 and with an ultrasonic transmitter 5, wherein the ultrasonic receivers 2, 3, 4 and the ultrasonic transmitter 5 are arranged in a common plane 6. The ultrasonic transmitter 5 is for locating the objects 7, 8 a short broadband ultrasonic signal 9, for example, a linearly frequency-modulated signal with a length of a few hundred microseconds and tuned to the transmitter bandwidth, eg 30 kHz from. The emitted ultrasonic signal 9 is reflected on objects 7, 8 in the vicinity of the ultrasonic sensor 1. Ektoberfläche those points a Ob, which satisfy the Re ^¬ flexionsbedingung for a transmitter-receiver pair of the ultrasonic sensor 1, are referred to as reflection points. That is, the radiated from the ultrasonic transmitter 5 and emitted Ultra ^¬ sound signal 9 of the ultrasonic receiver 2, 3, 4 is reflected in the reflection points in each case a. The here in the vicinity of the ultrasonic sensor 1 and at a similar distance from the ultrasonic sensor 1 arranged objects 7, 8 are spherical or approximately point-shaped, so that the three reflection points in this case substantially coincide in one point and are therefore not shown separately. A light emitted from the ultra-sonic transmitter ^¬ 5 9 ultrasonic signal is reflected at the two objects 7, 8 are each substantially in the entire the ultrasonic emitter 5 facing half-space. Thus reach each ultrasonic receiver 2, 3, 4, respectively, both a light reflected from the first object 7 ultrasound signal 10 and an image reflected from the second object 8 Ultraschallsi ^¬ gnal 11. The timing diagram 12 in Fig. 1B has three separate time axes 13, 14 , 15 for each ultrasonic receiver 2, 3, 4 corresponding to the three receiver channels. On the time axes 13, 14, 15 are each the two time points ti, t _2, t ₃ of the reception of the reflected ultrasonic signals 10, located 11, wherein the first object 7 reflected Ul ^¬ traschallsignal 10 as a solid vertical line and from the second Object 8 reflected ultrasonic signal 11 are shown as gestri ^¬ smiled vertical line. The relevant for the location tracking time information of the individual ultrasonic receiver 2, 3, 4 and the corresponding distance information is obtained by correlation of the individual receiver channels based on ^{vorge ¬} said or learned correlation patterns. In this case, the measured ultrasonic signals are grouped, each group 16 each consisting of a running time measurement t _± from each receiver channel. In the case of n ultrasonic receivers or microphones, a group 16 consists of exactly n running time measurements t _± , where t _{± represents} a running time measurement of the channel i. In this case, only valid groups 16 are preferably formed, a group 16 then being valid only if the condition (t.-tj-c <d ₁₃ V 1 <i, j is met <n (2) where c is the speed of sound of the ultrasonic receivers in the vicinity of the ultrasonic sensor 1 and d _{± j} is the physical distance Zvi ^¬ rule i and j . in this case, a certain delay time measurements t _± generally more groups belong. a selection of a group used for the positioning of an object 16 may for example take place in that closely neighboring time signals are preferred. This assignment is based basically on the assumption that different objects are arranged at a different distance from the ultrasonic sensor 1. However, since this does not apply in the case shown (see Fig. Fig. 1A), the assignment to the group 16 is false and the reconstructed from the mutually associated signals object positions are therefore also in error. the actual Po ^¬ Objects of the objects 7, 8 can not be recognized in such an assignment.

FIG. 2 shows a further potential error case which can not be detected or solved with the aid of only three ultrasonic receivers 2, 3, 4 on the basis of the measured transit times. In this case 17 has an object to a discontinuity 18 in the form ei ^¬ ner stage. The transition at the point of discontinuity 18, ie in this case the height of the step, is smaller than the ^distance between the ultrasonic receivers 2, 3, 4, so that the condition (2) is fulfilled. The individual ultrasonic receivers 2, 3, 4 associated reflection points 19, 20, 21 lying on both sides of the discontinuity 18, and therefore in various ^¬ which sections 22, 23 of the surface of the object 17. In the case shown, is a reflection point 21 further from

Ultrasound sensor 1 is removed as the other two reflection points 19, 20. It is assumed that for locating the object 17 only the measured transit times corresponding to the three reflection points 19, 20, 21 are available. From the three transit times, that point in space can be calculated by trilateration, for which the distance from the three transmitter-receiver pairs formed with the ultrasound transmitter 5 and with one of the ultrasound receivers 2, 3, 4 corresponds to the measured transit times

corresponds (see for example Manolakis, DE, "Efficient Solution and Performance analysis of 3-D position estimation by trilateration, "Aerospace and Electronic Systems, IEEE Transactions on, vol.32, no.4, pp1239, 1248, Oct. 1996 doi: 10.1109 / 7.543845) illustrate this calculation with the fact that for three imaginary spheres whose radii are the respective distances corresponding to the measured transit times, a common tangent plane (not shown) is determined, and that the sought Ob ektposition that point in the tangent plane is assumed, which the ultrasonic transmitter 5 is closest. Basically, this method of reconstruction would be sufficient for a 3D positioning, but it is not robust to measurement errors and opposite errors due to the implicit assumptions regarding the object geometry. in particular impli ^¬ grace this calculation that all the reflection points lie on the same object and that the can vernachläs ^¬ shape or type of the object are SIGt. in the in Fig. 2 g However ezeigten situation suggests this assumption fails: due to the different distances of the three reflection points 19, 20, 21 in Figure 2, the attached ^¬ made tangential plane an inclination, which does not meet the rea ^¬ len circumstances, and the object position 25 determined is erroneous. , The present invention relates insbeson ^¬ particular, the detection and elimination of this problem by the use of at least one ultrasonic receiver in the fourth Kom ^¬ bination with a specially tailored to ultrasonic sensors having four or more receivers locating method.

In Fig. 3, an ultrasonic sensor 26 with four ultrasonic receivers 27, 28, 29, 30 in a similar situation as in

2, ie an attempt is made to locate the object 17 with the point of discontinuity 18. Again, the result will be incorrect if one tries to determine an object position from the measured Laufzei ^¬ th with the known methods. To avoid such mistakes, therefore, proposes the

present invention provides, firstly that the Refle ^¬ xionspunkte lie on the same item to verify the assumption. This check can be carried out in particular by determining whether the (hypothetical) reflection points which correspond to the measured transit times lie in one plane. Based on this criterion, those groups of running time measurements t _± , which withstand the review, are selected and the remaining groups can be discarded. In the case shown in FIG. 3, the ultrasonic receivers 27, 28, 29, 30 are arranged in pairs 27, 28 and 29, 30 and mirror-symmetrically with respect to an intersection of the connecting lines of the individual pairs 27, 28 and 29, 30. The connecting lines therefore intersect exactly in their respective center. The ultrasonic transmitter 31 is also in the center of the receiver, that is arranged exactly at the common intersection of the connecting lines. In this construction, the ultrasonic sensor 26 can be shown that the transit time measurements t ^¬ _± must meet for the pairs of ultrasonic receivers 27, 28, 29, 30 following condition:

The inaccuracy of an ultrasonic sensor with a receiver ^¬ distance of 80 mm to a distance of 1 m is less than 1 mm, so that all the groups which do not satisfy the condition (3) can be recognized as invalid and discarded. According to this method, the measurement shown in Fig. 3 can be recognized as invalid, so that at least no erroneous Whether ^¬ jektposition is detected.

In addition, the principle described above can also be used in the assignment of the running time measurements, as will be explained with reference to FIGS. 4A and 4B. As in FIGS. 1A and 1B, two objects 7, 8 are shown here, which are responsible for two reflected ultrasonic signals 10, 11 or the corresponding running time measurements t _± or t _± 'of the here four receiver channels. By checking condition (3) for all combinatorially possible groups, the two correct assignments which correspond to the average transit times t == (ti + t ₂ ) / 2 * = (t ₃ + t ₄ ) / 2 or t '== (t ₁ '+ t ₂ ') / 2 * = (t ₃ '+ t ₄ ') / 2 ent ^¬ speaking, determined and calculated on the basis of the grouped measurements, the two corresponding object positions.

However, special cases exist in which a ^¬ Ultra sound sensor with four symmetrically arranged Ultraschallemp- catch a discontinuity 18 can not recognize, as shown in Fig. 5. In this case, the ultrasonic sensor 37 with respect to the discontinuity 18 of the object 17 is randomly symmetrical ^¬ assigns, so that for each pair of receivers 27, 28 and 29, 30 each have a reflection point on each side of the discontinuity 18 and thus satisfies the condition (3) is, although each of the four running time measurements of the two receiver pairs 27, 28 and 29, 30 determined object position would be erroneous for the same reason, as already elucidated with reference to FIGS. 2 and 3 in detail ^¬ tert. To cover this special case, the

Ultrasonic sensor 37 to a fifth ultrasonic receiver 38, so that the distance between the ultrasonic sensor 37 and a fifth reflection point 39 on the object 17 can be measured by means of ultrasound. Thus, the error can be recognized as follows: First, a provisional object position is determined based on the running time measurements of the four ultrasonic receivers 27, 28, 29 and 30, and in a second step it is checked whether the running time measurement of the fifth ultrasonic receiver 38, i. over a distance from the ultrasonic transmitter 31 to the reflection point 39 and back to the ultrasonic receiver 38, is consistent with this preliminary object position, e.g. the distance corresponding to the provisional object position is compared with the separately measured distance from the fifth ultrasonic receiver 38.

From Fig. 6 to 8 further advantages in connection with the use of the ultrasonic sensor 37 for locating can be seen. Based on the additional location information, which from the measurements of the fourth and - optionally - the fifth

Ultrasonic receiver can be obtained, it is possible ent ^¬ speaking reference points 40 in space, not only the position of the ultrasonic sensor 37, but also the

To determine alignment. This can be done, for example, by means of passive orientation points, preferably in the form of retroreflectors 40 (see FIG. It can be demonstrated that the selection criteria described above are also valid for ultrasound signals reflected on retroreflectors 40 (multiply). If three such landmarks are present, their positions in space are known, the position and orientation of the. Can be based on ent ^¬ speaking running time measurements Ultrasonic sensor 37 are calculated in a global coordinate system. The advantage of a retro-reflector 40 as Orien ^¬ tierungspunkt is that the "visible" for the ultrasonic sensor 37 position of the retroreflector 40 is independent of the incident angle A of the ultrasonic signal ^¬ 9 and the distance from the ultra-sonic sensor ^¬ 37th

In addition, the present invention can be applied not only to active ultrasonic sensors 26 and 37 having their own ultrasonic transmitters 31, but also to passive ultrasonic sensors 41 (see Fig. 7). In such passive ultrasonic sensors 41, the ultrasonic signals 42 are radiated by external signal generators 43, which are arranged at known positions in space and therefore - as described above - can also be used as orientation points. The ektpositionen Whether etwai ^¬ ger objects 44 can thereby using the measured in space

Running times of the ultrasonic signals 42 and the known Positio ^¬ nen the signal generator 43 are reconstructed. In this case, the described selection criterion can additionally be used for the assignment of a plurality of running time measurements to an ultrasonic transmitter, ie in this example a signal generator 43.

Compared with passive sensors 41, an active ultrasonic sensor 37 has the advantage that the fixed distance d between the Ultra ^¬ sound transmitter 31 and the ultrasonic receivers 27, 28, 29, 30, 38, for example, can be used for temperature compensation (see. Fig. 8). In this case, the sound velocity c at the current ambient conditions is preferably measured on the basis of several running time measurements t and the known distance d, and the ultrasonic sensor 37 is calibrated in this way. The accuracy of the measurement is limited only by the time resolution of the transit time measurement and the distance between transmitter 31 and receivers 27, 28, 29, 30, 38. At a distance of 4 cm between the transmitter 31 and the receivers 27, 28, 29, 30, 38 could

Sound velocity can be determined, for example, with an accuracy of about ± 2 m / s.

FIG. 9 schematically shows an ultrasound sensor 45 with an ultrasound transmitter 46 and two (or more) ultrasound receivers 47 shown. The ultrasonic transmitter 46 receives an ultrasonic signal to be transmitted from a drive unit 48, the signal being adjusted in an amplifier 49 to the requirements of the ultrasonic transmitter 46. The ultrasound receivers 47 are set up to receive the reflected ultrasound signals and are connected to correlator units 50, which can correlate a received re flected ultrasonic signal with the transmitted Ultraschallallsi ^¬ signal. In order to determine the correct temporal correlation, the correlator units 50 are connected and synchronized with the drive unit 48 via synchronization connections 51 (dashed lines). Based on the calculated result of the correlations run times are those in a Vorselektionsmodul 52 on the condition (2), ie, pre-selected on basis of the sensor dimensions and discarded ene combinations which can be physically out ^¬ closed due to sensor geometry. Subsequently, in a further selection module 53, the middle criterion (3) is applied to the measured and preselected transit times and thus the large part of artifacts is removed. Finally, the Whether ^¬ jektortung is carried out based on the adjusted times, checked with a possible additional receiver for plausibility and output in the positive case in a locating module 54, which is connected to the selection module 53rd This means that after the object location, the criterion - if available - is used with the additional (eg fifth) receiver and output after a positive check; object location is based on the measured transit times of the other receivers. Together, the ultrasound receivers 47 with the correlator units 50, the preselection module 52, the selection module 53 and the locating module 54 form an evaluation unit 55. In particular, the prescience module 52, the selection module 53 and the locating module 5 can be used in practice as software modules on a common, in Ul ^¬ traschallsensor 45 integrated microprocessor platform (not shown) may be implemented.

Claims

claims 1. A method for locating objects (7, 8, 17) in space by means of an ultrasonic sensor (26) (27, 28, 29, 30), wherein (from an external or by the ultrasonic sensor at least four Ultra ^¬ sound receiver 26) Ultrasound transmitter (43 or 31) an ultrasonic signal (42 or 9) is emitted and from the ultrasonic receivers (27, 28, 29, 30) of the ultrasonic sensor (26) in each case one in at least one associated reflection point (33, 34, 35, 36 is received) reflected Ul ^¬ traschallsignal (11) and evaluated for obtaining at least one object position, characterized in that at more gained parallel object positions, these are selected according to a selection criterion, which is defined in that the object position associated in each case at least four reflection points (33 , 34, 35, 36) are approximately in a common plane in space. 2. The method according to claim 1, characterized in that the ultrasonic sensor (37) has at least five ultrasonic receivers (27, 28, 29, 30, 38), wherein the selection criterion is accordingly defined by the fact that the object position depending ^¬ Weil assigned at least five Reflection points (33, 34, 35, 36, 39) lie approximately in a common plane in space. 3. The method of claim 1 or 2, characterized in that object positions for which no common plane gefun ^¬ can be recognized as artifacts and discarded. 4. The method according to any one of claims 1 to 3, characterized in that an even number of ultrasonic receivers (27, 28, 29, 30) in a plane and in at least two relative to a common center mirror-symmetrical pairs (27, 28 and 29, 30), wherein the selection criterion is defined by the fact that the average transit times of the receiver pairs (27, 28 or 29, 30) are substantially identical. 5. The method according to claim 4, characterized in that after the checking of the selection criterion using the recipient ^¬ couple (27, 28, 29, 30) has a plane on the basis of the receiver pairs (27, 28, 29, 30) measured transit times is determined and for a second part of ultrasound receivers (38), which need not be arranged in mirror-symmetrical pairs, it is checked whether the latter Ultrasound receivers (38) associated reflection points (39) lie in the determined plane, otherwise the selection ^¬ criterion is not met in total. 6. The method according to claim 4, characterized in that the receiver pairs (27, 28 or 29, 30) have the same distance between the ultrasonic transmitter (31) and ultrasonic receivers (27, 28, 29, 30), wherein in particular the ultrasonic receiver (27 , 28, 29, 30) are arranged in one of the number n of ultrasonic receivers (27, 28, 29, 30) corresponding n-fold rotational symmetry, wherein an equivalent reflection point for all Ul ^¬ traschallempfänger (27, 28, 29, 30) on Base of the Receiver pairs (27, 28 and 29, 30) measured transit times is determined and for a second part of ultrasonic receivers (38), which need not be arranged symmetrically, it is checked whether the reflection points (39) associated with these ultrasonic receivers (38) with the equivalent reflection ^¬ point are consistent, otherwise the selection criterion is not met in total. 7. The method according to any one of claims 1 to 6, characterized in that the ob ektpositionen of several objects (7, 8) are determined starting from a single radiated ultrasound signal (9). 8. The method according to any one of claims 1 to 7, characterized in that upon receipt of more than one reflected ultrasonic signal (10, 11) in a receiver (27, 28, 29, 30), the object positions by forming all possible combinations of reflection points (33, 34, 35, 36), each reflection point (33, 34, 35, 36) being associated with another ultrasonic receiver (27, 28, 29, 30). 9. The method according to any one of claims 1 to 8, characterized in that ektpositionen in which the maximum distance between two respective associated reflection points (33, 34, 35, 36) the maximum distance between two ultrasonic receivers (27, 28, 29, 30) is discarded prior to selection. 10. Ultrasonic sensor (26, 37, 41, 45) for locating objects (7, 8, 17, 44) in the room with at least four ultrasonic receivers (27, 28, 29, 30, 47) which are arranged one of a external or to the ultrasonic sensor (26, 37, 45) associated with the ultrasonic transmitter (31, 43, 46) radiated and in at least one reflection point (33, 34, 35, 36) reflected Ultra ^¬ sound signal (9, 42 or 10, 11) received, and with an evaluation unit (55), which is connected to the ultrasonic receivers (27, 28, 29, 30, 47) and adapted to gain at least one object position of the received ultrasonic signals (10, 11), characterized in that the Evaluation unit (55) comprises a selection module (53) which is set up at several par ^¬ allel object positions for determining those object positions in which the object position respectively associated at least four reflection points (33, 34, 35, 36) approximately in a common plane in the rough m lie. 11. Ultrasonic sensor (37, 41) according to claim 10, characterized in that at least five ultrasonic receivers (27, 28, 29, 30, 38, 47) are connected to the evaluation unit (55). 12. Ultrasonic sensor (26, 37, 41) according to claim 10 or 11, characterized in that at least a part of the ultrasonic ^¬ receivers (27, 28, 29, 30) in a plane and in at least two with respect to a common center mirror-symmetrical pairs ( 27, 28 and 29, 30) is arranged. 13. Ultrasonic sensor (26, 37, 41) according to claim 12, characterized in that the receiver pairs (27, 28 and 29, 30) have the same distance between the ultrasonic transmitter (31) and ultrasonic receivers (27, 28, 29, 30) in particular the ultrasonic receivers (27, 28, 29, 30) are arranged in one of the number n of the ultrasonic receivers (27, 28, 29, 30) corresponding n-zäh- ligen rotational symmetry.