Ultrasonic measuring unit having an evaluation unit with an ai module for simulating multilateration calculations on the basis of detected echoes of ultrasonic signals

Abstract

The invention relates to an ultrasonic measuring unit (1), in particular for use in a vehicle (2), comprising a control unit (3) and an evaluation unit (4), wherein the control unit (3) is designed to control first ultrasonic transducers (11) for transmitting ultrasonic signals and, on the basis of reception signals generated with the aid of second ultrasonic transducers (12), to detect echoes produced owing to signal paths of the ultrasonic signals, and the evaluation unit (4) has a trained AI module (5) and is designed to process the reception signals in order to detect an object (6; 60) causing the echoes, wherein the processing involves at least generating feature vectors (111, 112, 113) on the basis of the reception signals with the aid of the AI module (5), and the control unit (3) or the evaluation unit (4) is configured to determine, on the basis of the reception signals, a signal propagation time or signal propagation distance corresponding to the signal path for the associated echo, and the AI module (5) is designed to determine the feature vectors (111, 112, 113) on the basis of the signal propagation times or the signal propagation distances.

Description

Ultrasound measuring unit with an evaluation unit with an AI module for simulating multi-calculations based on detected echoes of ultrasound signals AREA OF TECHNOLOGY

[0001] The invention relates to an ultrasonic measuring unit, particularly for use in a vehicle, comprising a control unit and an evaluation unit. The invention further relates to a method for operating an ultrasonic measuring unit. STATE OF THE ART

[0002] It is known to use ultrasonic sensors to detect objects in the vicinity of a vehicle. For example, DE 10 2016 218064 A1 describes an ultrasonic sensor system for detecting the environment of a vehicle using ultrasound. The ultrasonic sensor system described therein comprises an ultrasonic transmitter, an ultrasonic receiver, and a control unit for controlling the operation of the ultrasonic sensor system. The control unit is configured to execute an operating procedure in which ultrasonic signals are emitted at a plurality of successive time points. Furthermore, the operating procedure provides for receiving ultrasonic echo signals received from one or more objects in the environment, whereby a track is assigned to each emitted and received ultrasonic signal, and a distance value is determined for each existing ultrasonic echo.In this operating procedure, the ultrasound echoes are grouped into tracks of multiple temporally successive ultrasound signals if the distance values ​​of the ultrasound echoes follow a predetermined pattern. Ultrasound echoes that do not belong to any group are classified as interference.

[0003] It is an object of the invention to design an improved ultrasonic measuring unit and a method for operating an ultrasonic measuring unit. The objects underlying the invention are solved by the features of the independent claims.

[0004] An ultrasonic measuring unit, particularly for use in a vehicle, is proposed. The ultrasonic measuring unit comprises a control unit and an evaluation unit. The control unit is configured to drive first ultrasonic transducers to transmit ultrasonic signals. Furthermore, the control unit is configured to detect echoes generated by the ultrasonic signals based on the signal paths of the received signals from second ultrasonic transducers. The evaluation unit has a trained AI module and is configured to process the received signals to detect an object causing the echoes. This processing includes at least the generation of feature vectors using the AI ​​module based on the received signals. The feature vectors describe the shape, position, direction of movement, and / or velocity of the object.The respective feature vector can, in particular, contain values ​​to specify a point and / or a line that describe the shape and / or position of the object. These values ​​can, for example, specify the coordinates of points or lines in a space located in front of the ultrasonic measuring unit.

[0005] Furthermore, the control unit or evaluation unit is configured to determine a signal propagation time or signal path for each echo, corresponding to the signal path, depending on the received signals. The Kl module is designed to determine the feature vectors as a function of the signal propagation times or signal paths. Since the signal propagation times or signal paths are determined as a function of the received signals, the Kl module is designed to determine the feature vectors, particularly indirectly, as a function of the received signals.

[0006] The AI ​​module, in particular, comprises mathematical functions with parameters. The values ​​of the parameters are specifically adapted to training datasets through training of the AI ​​module. The trained AI module is specifically defined by the fact that the values ​​of the parameters of the mathematical functions are adapted to the training datasets. The mathematical functions can, for example, include a chain of artificial neurons arranged in several hidden layers. Connection weights between the neurons, especially between neurons from different hidden layers, The neurons can represent the values ​​of the parameters of mathematical functions. Collectively, they can form an artificial neural network of the AI ​​module. This neural network can be, for example, a multi-layer perceptron network, a radial basis function (RBF) network, or a convolutional neural network (CNN).

[0007] Because the AI ​​module is configured to determine feature vectors based on signal propagation times or signal paths, calculations necessary for multilateration using these propagation times or paths could be replicated or simulated by executing the AI ​​module's mathematical functions. In other words, multilateration could be performed by executing the AI ​​module's functions. This would reduce the computational effort required to determine feature vectors. Consequently, the ultrasonic measurement unit could enhance vehicle safety, as the proposed ultrasonic measurement unit would enable faster object detection, particularly in real time.

[0008] The echoes are caused in particular by at least a first portion of the ultrasound waves emitted by at least one of the first ultrasound transducers, which carry the ultrasound signals, being reflected by the object and reaching the second ultrasound transducers as the first reflected ultrasound waves. Depending on the application, the object's position, direction of movement, and / or velocity includes, in particular, its relative position, direction of movement, and / or velocity with respect to one of the first and second ultrasound transducers. It is assumed that the relative positions of the first and second ultrasound transducers are known. In principle, a second portion of the ultrasound waves can be reflected by other objects, such as a road surface, and reach the second ultrasound transducers as second reflected ultrasound waves.

[0009] The received signals generated by the second ultrasound transducer are primarily in the form of electrical signals, such as amplitude signals of an electrical voltage or current. Each of the second ultrasound transducers has at least one ultrasound diaphragm to convert the reflected ultrasound waves into the received signals. It is possible that the second ultrasound transducer receives the signals based on the reception of the first and / or second transducer. reflected ultrasound waves generate the first received signals. These signals, generated by the interaction of the ultrasound membrane with the first reflected ultrasound waves, are referred to as useful signals. Capturing the echoes can thus be considered capturing the useful signals. Second received signals, generated by the interaction of the ultrasound membrane with the second reflected ultrasound waves, are referred to as interference signals.

[0010] According to one variant, the control unit is configured to capture the echoes in the form of the useful signals. Specifically, the evaluation unit is configured to detect the useful signals as a function of the received signals. In this variant, the evaluation unit, and in particular a first additional control module of the evaluation unit, can be configured to determine the signal propagation times or signal paths as a function of the useful signals. To detect the useful signals as a function of the received signals, the evaluation unit can be configured to perform filtering of the received signals. Filtering of the received signals includes, in particular, separating the useful signals from the interference signals.

[0011] For example, the evaluation unit can separate the desired signals from the interference signals by comparing the amplitude of the received signals with a predefined first threshold. The evaluation unit can then assign those received signals whose amplitude exceeds the first threshold to the desired signals. This assignment can be achieved, for example, by sorting, marking, and / or separately storing those received signals whose amplitude is greater than the first threshold. Conveniently, the evaluation unit is configured to perform the filtering in such a way that, as a result of the filtering, the desired signals are stored separately from the interference signals in a memory within the evaluation unit.

[0012] According to one possible embodiment, the evaluation unit can have a second additional Kl module, wherein the second additional Kl module is configured to detect the useful signals depending on the received signals, in particular to separate them from the interference signals and in particular to make them available for separate processing.

[0013] To determine the signal transit times or signal travel distances, the evaluation unit or the control unit can be configured to record a specific time of transmission of the respective ultrasound signal from a respective first ultrasound transducer of the first The ultrasound transducers, hereinafter referred to as the respective transmission time, are used to detect the transmission time. The ultrasound measuring unit can have a communication link between the control unit and the evaluation unit for detecting the respective transmission time. The first ultrasound transducers transmit the ultrasound signals, particularly in pulses and with a time offset from each other.

[0014] Furthermore, the evaluation unit or the control unit can be configured to record a set of reception times of the useful signals for each transmitted ultrasound signal. This is based on the assumption that in most applications, the respective ultrasound signal can be received by two or more of the secondary ultrasound transducers at the different reception times of the respective set of reception times. This is primarily due to the fact that the secondary ultrasound transducers are located at different distances from the object.

[0015] In particular, the evaluation unit or the control unit is configured to determine, for each ultrasound signal, especially for each useful signal, a set of signal propagation times from the difference between the respective reception time of each set of reception times and the respective transmission time. For example, with three first and three second ultrasound transducers, the evaluation unit could calculate three sets of three signal propagation times each, assuming the transmitted ultrasound signals are received by the three receivers. If at least one of the ultrasound signals is reflected more than once and received twice by at least one of the second ultrasound transducers, more than nine signal propagation times can result.

[0016] According to one possible embodiment, the evaluation unit or the control unit is configured to calculate the signal propagation distances based on the signal propagation times, taking the speed of sound into account. The evaluation unit can calculate the speed of sound as a function of a measured ambient temperature. For most applications, the speed of sound can be assumed to be constant. In these cases, the terms "signal propagation time" or "signal propagation distance" can express a similar level of information. For this reason, the phrase "signal propagation times" or "signal propagation distances" is used in this disclosure.

[0017] In particular, the evaluation unit is configured to calculate the respective feature vector using the Kl module and the respective set of signal propagation times. For this purpose, the evaluation unit can send the respective set of signal propagation times or signal paths to the Kl module as an input data set. The Kl module can, in particular, be configured to calculate the respective feature vector as an output data set, depending on the respective set of signal propagation times or signal paths.

[0018] To differentiate the ultrasound signals, they can be transmitted sequentially, such that a subsequent signal is emitted after a previous signal has been received. Alternatively or additionally, the ultrasound signals can also have different characteristics. For example, they can differ in their pulse sequences and / or frequencies.

[0019] Within the scope of this disclosure, the first ultrasonic transducers are considered as transmitters of the ultrasonic waves and the second ultrasonic transducers as receivers of the ultrasonic waves. In this context, one of the transmitters and one of the receivers may be designed as a single ultrasonic sensor. In this case, the ultrasonic sensor can operate as a transmitter in a first time interval and as a receiver in a second time interval.

[0020] In one possible configuration, the AI ​​module is trained using training datasets, each of which comprises a training input dataset and a corresponding training output dataset. Specifically, each training input dataset contains training signal propagation times or distances of a training ultrasound signal reflected from a training object. Each training output dataset contains at least one training feature vector. This training feature vector is generated, in particular, by multilateration using the relative positions of the first and second ultrasound transducers to each other and as a function of the training signal propagation times or distances of the corresponding training input dataset.The training feature vectors can, in particular, contain values ​​to specify training points and / or training lines that describe the shape and / or position of the training object.

[0021] The training signal propagation times can be determined analogously to the signal propagation times described above. For this purpose, the transmitters can send out the respective training ultrasound signal at a specific training transmission time. The receivers can receive the training ultrasound signals reflected from the training object. The evaluation unit or the control unit can determine a set of training reception times for each transmitted training ultrasound signal, in particular by separating training signal signals from training interference signals. Furthermore, depending on the training transmission time and the training reception times, the evaluation unit or the control unit can determine a set of training signal propagation times and / or training signal propagation distances for each training ultrasound signal.

[0022] To determine the training feature vectors, a multilateration procedure can be performed using the relative positions of the first and second ultrasound transducers to each other and as a function of the training signal transit times or training signal path lengths. The following describes the functions of a multilateration unit configured to perform the multilateration procedure. The multilateration unit can be part of the evaluation unit or a separate, particularly external, evaluation unit, especially an external computer.

[0023] According to one possible configuration, the multilateration unit is set up to determine values ​​of ellipse parameters for the respective training signal transit time or distance, describing a specific ellipse. The respective ellipse is specifically assigned to the respective training signal transit time or distance of the respective set of training reception times associated with the respective transmitted training ultrasound signal. Thus, the multilateration unit can calculate a respective set of ellipse parameter values, hereinafter referred to as the respective ellipse parameter set, for each transmitted training ultrasound signal. The ellipses describe, in particular, possible locations of the training object in relation to at least one of the first and second ultrasound transducers. In principle, the possible locations can also be described by hyperbolas.For the sake of simplicity, this revelation refers only to ellipses. It is understood, therefore, that multilateration can also be performed using hyperbolas.

[0024] The values ​​of the ellipse parameters characterize the ellipses. For example, the values ​​of the respective ellipse parameter set can specify a measure of a major axis and a measure of a minor axis of the respective ellipse, and in particular a position of the foci of the respective ellipse.

[0025] Each ellipse is characterized in particular by the fact that the sum of the distances between any point on the ellipse and its foci is equal to the product of the respective training signal propagation time and the speed of sound. The speed of sound is present in the space between the ultrasound measuring unit and the training object. A first focal point of each ellipse is determined by the position of the first ultrasound transducer from which the respective training ultrasound signal is transmitted. A second focal point of each ellipse is determined by the position of the second ultrasound transducer from which the respective training ultrasound signal is received.

[0026] According to one variant, the multilateration unit can be configured to determine the training feature vectors, in particular the values, especially coordinates and / or angles, for describing the training points and / or training lines by calculating the intersection points of the ellipses. For each emitted training ultrasound signal, an intersection point can be determined. The respective intersection point of the ellipses describes, in particular, a respective reflection point at which the respective training ultrasound signal was reflected by the training object, and thus a respective training point among the training points. According to one variant, the multilateration unit can determine the training lines based on at least two of the training points.

[0027] In principle, the multilateration unit can be configured to approximate the reflection point relative to the first and / or second ultrasound transducers, depending on the ellipse parameter sets for the respective training ultrasound signal. In particular, the multilateration unit is configured to minimize distances between the approximated reflection point, i.e., the approximated training point, and a selected set of more than two ellipses.

[0028] Alternatively or additionally, the multilateration unit can approximate the position of a given locus by minimizing the distances of the selected ellipses to the locus. The locus represents, in particular, points of the training object that extend along a direction on a surface of the training object, for example, if the training object is a wall. In one variant, the multilateration unit can be configured to determine the training lines in the form of approximated loci.

[0029] Minimizing the distances between each approximated training point and / or the approximated locus and the selected ellipses can be achieved through an iterative process using a termination criterion. If the termination criterion is met, the last approximated training point or locus can represent the respective training point or training line.

[0030] In the iterative process, the multilateration unit changes the position of the approximated training point or the approximated locus line in each new iteration step and recalculates the distances. The position can be changed, for example, depending on the derivatives of the distances according to the dimensions of the space between the ultrasound measurement unit and the training object.

[0031] Because the training feature vectors are generated using multilateration, the trained AI module, which is trained using the training datasets, can replicate, and in particular simulate, the multilateration process, especially individual steps of the multilateration process. This allows the trained AI module to determine feature vectors as a function of the signal propagation times or signal paths, as if these feature vectors had been determined using the multilateration process itself.

[0032] The following describes a possible training method for the AI ​​module. As described above, the AI ​​module's training datasets are used to adapt the AI ​​module's parameter values ​​to the training datasets. Specifically, the training input datasets are applied to an input of the AI ​​module; that is, the training signal propagation times or signal paths of the respective training input dataset are read in using the AI ​​module's inputs. Furthermore, approximate training feature vectors can be calculated using the AI ​​module's mathematical functions. Here, the respective approximate The training feature vector is calculated as a function of the training input dataset of the respective training dataset using mathematical functions. Furthermore, a deviation of the respective approximated feature vector from the training feature vector belonging to the training output dataset corresponding to the respective training input dataset used to calculate the respective approximated feature vector can be calculated.

[0033] For each training dataset, a respective deviation between the approximated feature vector and the respective training feature vector can be determined in this way. Conveniently, changes in the values ​​of the mathematical function parameters can be determined based on these deviations. For example, the deviations, especially the squared deviations, can be summed, and the resulting sum can be derived with respect to the parameters of the AI ​​module. The change in each parameter can then be calculated, in particular, as a function of the derivative of the sum with respect to that parameter. In this way, the parameters can be significantly altered if the derivative of the sum with respect to the respective parameter is relatively large. Such a calculation, especially the adaptation of the AI ​​module parameters to the training datasets, is similar to the backpropagation method.It goes without saying that other possible learning methods can be used to train the AI ​​module. Furthermore, the AI ​​module is not limited to a neural network. In principle, the mathematical functions can represent polynomial functions of the AI ​​module.

[0034] In a training module, the AI ​​module is configured to determine the feature vectors based on the signal strengths of the echoes acquired by the second ultrasound transducer. This allows the echoes to be weighted according to their signal strength when determining the feature vectors, thereby increasing the accuracy of the ultrasound measurement unit. Specifically, the evaluation unit can be configured to determine the signal strength of the respective input signal. In this case, the AI ​​module can be configured to calculate the feature vectors based on the signal strengths of the input signals. The signal strength of the respective input signal can, for example, be a time-averaged amplitude of the input signal. In this training module, the training input data sets specifically include signal strengths of the training ultrasound signals. Specifically, the training input data sets can include a signal strength of a respective input signal from the received Training ultrasound signals are included. In particular, the respective training data set can include a time-averaged amplitude of the useful signal of the respective training ultrasound signal.

[0035] In a further training module, the K-module is configured to determine the feature vectors as a function of the relative positions of the first and second ultrasound transducers. The relative positions of the first and second ultrasound transducers include, in particular, two-dimensional coordinates of the first and second ultrasound transducers within a coordinate system in a plane of the ultrasound measuring unit on which the first and second ultrasound transducers are arranged. Specifically, the evaluation unit is configured to provide, for each ultrasound signal, the coordinates of the respective transmitter that emitted the respective ultrasound signal and the respective receiver that received the respective ultrasound signal as input data for the K-module. These coordinates, in particular, specify the foci of the ellipses.Since the Kl module is specifically designed to replicate the multilateration process, the Kl module could replicate multilateration more accurately if the coordinates of the foci of the ellipses could be used as input data for the Kl module.

[0036] In a training course, the AI ​​module is configured to determine feature vectors based on timestamps of the ultrasound signals generated by the control unit. Each timestamp can indicate the transmission time of the respective ultrasound signal. This training could enable the AI ​​module to replicate a multilateration method that takes into account relative movement between the ultrasound measurement unit and the object. The evaluation unit can determine the ultrasound signal timestamps, for example, by capturing maxima of the received signals. These maxima can, in particular, represent the useful signals.

[0037] A further training course could include configuring the AI ​​module to determine feature vectors based on an assignment of echoes to the respective second ultrasound transducers that detected the echo and / or an assignment of echoes to the respective first ultrasound transducers that emitted the echo. This training could eliminate the need to present the relative positions of the first and second ultrasound transducers to the AI ​​module as input data for calculating the feature vectors. This could reduce the computational effort required for calculating the feature vectors.

[0038] The assignment of echoes to transmitters and / or receivers can be achieved through coding. According to one possible implementation of this training, the AI ​​module can calculate the feature vectors based on information about which transmitter sent out the respective ultrasound signal and which receiver received it. In this training, the training input datasets also contain information about which transmitter sent out the respective training ultrasound signal and which receiver received it. If the training described above is performed with such configured training input datasets, the relative positions of the transmitters and receivers could be represented during the AI ​​module's training by adjusting the parameter values ​​of the AI ​​module's functions.In this case, physically adjacent first and / or second ultrasound transducers could induce a locally enhanced connection between neurons of the KNN, which may be particularly noticeable through comparatively high values ​​of connection weights.

[0039] In a further embodiment, the control unit can be configured to control the first ultrasound transducers such that they emit the ultrasound signals at different time intervals. In this embodiment, the evaluation unit is configured to associate the respective portions of the signal transit times or signal paths with the respective emitted ultrasound signal and to group them into the respective set of signal transit times or signal paths described above. The evaluation unit can perform this association, and in particular the grouping, of the respective portions of the signal transit times or signal paths, for example, based on the different wavelengths at which the first ultrasound transducers emit the ultrasound signals.Furthermore, the evaluation unit can be configured to select a subset of signal propagation times or signal paths depending on differences between the corresponding signal propagation times or signal paths of the different sets. In this configuration, the K-module can be configured to determine the feature vectors depending on the selected subset of signal propagation times or signal paths. By selecting the subset, signal propagation times calculated from the noise signals could, in particular, be discarded for the calculation of the feature vectors. This could improve the accuracy of the... The ultrasound measuring unit can be increased.

[0040] In particular, the evaluation unit can be configured to track a potential reflection area of ​​the object over time by selecting a subset of signal propagation times or signal path lengths, i.e., to perform so-called "tracking" of the potential reflection area. For this purpose, the evaluation unit can determine a relative velocity of the potential reflection area to the ultrasonic measuring unit based on the differences between the corresponding signal propagation times or path lengths of two of the different sets, in particular a first and a second set of signal propagation times. Using this relative velocity, the evaluation unit can then check whether the signal propagation times determined based on the last emitted ultrasonic signals (hereinafter referred to as last signal propagation times) were reflected at the potential reflection area.For such a check, the evaluation unit is specifically designed to approximate the values ​​of the last signal propagation times based on the relative velocity and to compare these approximations with the actual values. The evaluation unit can, for example, approximate the last signal propagation times based on the propagation times calculated for the ultrasound signals emitted before the last ultrasound signal. If the difference between the approximated last signal propagation times and the actual values ​​is below a predefined second threshold, the evaluation unit can include the last signal propagation times in the subset of propagation times. Otherwise, the evaluation unit can discard the last signal propagation times for the calculation of the feature vectors.

[0041] According to a possible further development of this configuration, the evaluation unit has a third additional logic module (KL module) which is configured to determine the subset of signal propagation times as a function of the signal propagation times. In this further development, the logic module can be configured to determine the feature vectors as a function of the selected subset of signal propagation times or the associated signal paths.

[0042] In a further embodiment, the second ultrasonic transducers can be configured to generate the received signals at different time intervals. In particular, the K-module is configured to determine the feature vectors during these different time intervals, depending on the received signals. Furthermore, the evaluation unit can be configured to generate an image of the vehicle's surroundings based on the feature vectors determined at the different time intervals. The received signals The second ultrasound transducer can be used to generate different time intervals for two reasons, among others. Firstly, the ultrasound waves can be reflected by different surfaces, resulting in different signal paths and thus different signal propagation times and therefore different reception times. Secondly, the ultrasound signals can be emitted at different time intervals, as described above. By determining the feature vectors during these different time intervals based on the received signals generated during those intervals, the feature vectors can be successively generated and stored in the ultrasound measurement unit, particularly in the evaluation unit.The evaluation unit is specifically designed to generate an image of the vehicle's surroundings in the form of a union of the points and lines described by the feature vectors.

[0043] In general, the time intervals at which receivers receive the first reflected ultrasound signals emitted by a single transmitter are significantly smaller than the time intervals at which the transmitters emit the various ultrasound signals. The different time intervals during which the AI ​​module generates the feature vectors as a function of the received signals are, in particular, larger than the time intervals at which receivers receive the first reflected ultrasound signals emitted by a single transmitter. This allows the AI ​​module to process multiple signal propagation times or signal paths within the respective time interval, specifically the aforementioned set of signal propagation times, in order to simulate the aforementioned multilateration.

[0044] In a further development of this configuration, the evaluation unit can be configured to determine the probability that a portion of the received signals are caused by reflections of the ultrasound signals on a road surface, hereinafter referred to as road surface reflection signals. In this further development, the KL module is specifically configured to determine the feature vectors as a function of this probability. By using the probability, the influence of the road surface reflection signals on the calculation of the feature vectors could be reduced. This could increase the accuracy of the ultrasound measuring unit. Furthermore, it is possible for the evaluation unit to calculate the selected portion of the signal propagation times or signal travel distances. The dependency of the probability is determined. In this case, the Kl module can be configured to determine the feature vectors depending on the selected portion of the signal propagation times or signal paths, which is determined using the probability.

[0045] According to one variant, the evaluation unit can be configured to determine the probability using a frequency distribution as a function of at least one signal profile, in particular an amplitude profile, of the received signals. The frequency distribution can, in particular, indicate how the amplitude of possible received signals is distributed over time. Specifically, the frequency distribution is generated by processing ultrasound signals reflected from a surface under investigation. For this purpose, transmitters can send the ultrasound signals toward the surface under investigation, and receivers can generate received signals based on the reflected ultrasound signals received. The evaluation unit can, for example, generate a received spectrum, in particular the frequency distribution, from the received signals.

[0046] Furthermore, a method for operating an ultrasonic measuring unit, particularly for use in a vehicle, is proposed. The ultrasonic measuring unit comprises a control unit and an evaluation unit with a trained AI module and can be configured according to one of the variants described above. The method comprises the following steps. In one step, the control unit activates first ultrasonic transducers to transmit ultrasonic signals. In another step, echoes generated by the ultrasonic signal paths are detected based on received signals generated by second ultrasonic transducers. In a further step, the received signals are processed to detect an object causing the echoes. This processing includes at least the generation of feature vectors using the AI ​​module based on the received signals.The feature vectors describe the shape, position, direction of movement, and / or velocity of the object. In a further step, a signal propagation time or signal path for each echo is determined based on the received signals using the control unit or the evaluation unit. In a further step, the feature vectors are determined based on the signal propagation times or signal paths using the K-module.

[0047] The procedure can further include the following additional steps. In a further step, training input data sets are generated. Each training input data set contains training signal propagation times or training signal path lengths of a training ultrasound signal reflected from a training object. In a further step, a corresponding training output data set is generated for each training input data set. Each training output data set contains at least one training feature vector.Generating the respective training output dataset can, in particular, involve generating the respective training feature vector using multilateration, taking into account the relative positions of the first and second ultrasound transducers and their respective positions, and depending on the training signal propagation times or distances of the corresponding training input dataset. In a further step, the AI ​​module is trained using training datasets. Each training dataset comprises the respective training input dataset and the corresponding training output dataset.

[0048] In one aspect a system is revealed which... Stet clita kasd gubergren, no sea takimata sanctus est Lorem ipsum dolor sit amet. Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam erat, sed diam voluptua. At vero eos et accusam et justo duo dolores et ea rebum.

[0049] It is understood that one or more of the aforementioned embodiments can be combined with each other, as long as the embodiments do not exclude each other. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The following examples are explained in more detail using the drawings. They show:

[0051] Fig. 1 shows an object and a vehicle with an ultrasonic measuring unit, comprising a control unit and an evaluation unit with a KL module;

[0052] Fig. 2a shows a first amplitude signal generated with the aid of a first ultrasound receiver of the ultrasound measuring unit shown in Fig. 1;

[0053] Fig. 2b shows the course of a second amplitude signal generated with the aid of a second ultrasound receiver of the ultrasound measuring unit shown in Fig. 1;

[0054] Fig. 2c shows a third amplitude signal generated using a third ultrasound receiver of the ultrasound measuring unit shown in Fig. 1;

[0055] Fig. 3 shows a first feature vector generated using the Kl module;

[0056] Fig. 4 shows a second feature vector generated using the Kl module;

[0057] Fig. 5 shows a third feature vector generated using the Kl module;

[0058] Fig. 6 shows the vehicle shown in Fig. 1 and another object;

[0059] Fig. 7 shows an input data set and an output data set of the Kl module. DETAILED DESCRIPTION

[0060] In the following, similar elements are marked with the same reference symbols.

[0061] Fig. 1 shows an ultrasonic measuring unit 1 for use in a vehicle 2. The ultrasonic measuring unit 1 comprises a control unit 3 and an evaluation unit 4. The control unit 3 is configured to control first ultrasonic transducers, such as a first ultrasonic transmitter 11.1, a second ultrasonic transmitter 11.2, and a third ultrasonic transmitter 11.3, to transmit ultrasonic signals. For example, the control unit 3 can control the first ultrasonic transducers such that the first ultrasonic transmitter 11.1 emits first ultrasonic waves 10.1 to transmit a first ultrasonic signal, the second ultrasonic transmitter 11.2 emits second ultrasonic waves 10.2 to transmit a second ultrasonic signal, and the third ultrasonic transmitter 11.3 emits third ultrasonic waves 10.3 to transmit a third ultrasonic signal. According to one embodiment, the ultrasonic waves 10.1, 10.2, and 10.3 have different frequencies.

[0062] Furthermore, the control unit 3 is designed to detect echoes generated based on signal paths of the ultrasound signals, depending on the received signals generated by second ultrasound transducers, for example by a first ultrasound receiver 12.1, a second ultrasound receiver 12.2 and a third ultrasound receiver 12.3. The received signals can be in the form of electrical signals, in particular amplitude signals of a voltage or a current. The received signals can be transmitted from the second ultrasonic transducers 12.1, 12.2, 12.3 to the control unit 3 via lines shown as solid lines in Fig. 1.

[0063] The evaluation unit 4 has a trained AI module 5 and is configured to process the received signals to detect an object 6 that causes the echoes. The processing includes at least the generation of feature vectors using the AI ​​module 5 as a function of the received signals. The feature vectors describe the shape, position, direction of movement, and / or velocity of the object 6. In the application of the ultrasonic measuring unit shown in Fig. 1, the object 6 can be in the form of a post.

[0064] Furthermore, the control unit 3 or the evaluation unit 4 is configured to determine a signal propagation time or signal path corresponding to the signal path for the respective echo, depending on the received signals. The K module 5 is configured to determine the feature vectors as a function of the signal propagation times or signal paths.

[0065] According to one possible embodiment, the received signals from the control unit 3 can be sent to the evaluation unit 4 in the form of amplitude signals over time, as shown in Fig. 2. The following describes an example of how the evaluation unit 4 can determine the signal propagation times, and in particular the signal paths, as a function of the received signals.

[0066] Fig. 2a shows an example of a first waveform 20.1 of a first amplitude signal 21.1 over a time axis 22. The first waveform 20.1 can be generated, for example, by the control unit 3 using first electrical signals generated by the first ultrasound receiver 12.1. The first waveform 20.1 exhibits a maximum at a first reception time 22.1, which is generated by receiving the first ultrasound signal reflected from the first object 6 using the first ultrasound receiver 12.1. The maximum of the first waveform 20.1 specifically represents a first echo of the first ultrasound signal.

[0067] Fig. 2b shows a second waveform 20.2 of a second amplitude signal 21.2 over the Time axis 22. The second course 20.2 can be, for example, controlled using control unit 3 under The second waveform 20.2 is generated using second electrical signals produced by the second ultrasound receiver 12.2. At a second reception time 22.2, the second waveform 20.2 exhibits a maximum, which is generated by receiving the first ultrasound signal reflected from the first object 6 using the second ultrasound receiver 12.2. The maximum of the second waveform 20.2 specifically represents a second echo of the first ultrasound signal.

[0068] Fig. 2c shows a third waveform 20.3 of a third amplitude signal 21.3 over the time axis 22. The third waveform 20.3 can, for example, be generated by the control unit 3 using third electrical signals generated by the third ultrasound receiver 12.3. The third waveform 20.3 exhibits a maximum at a third reception time 22.3, which is generated by receiving the first ultrasound signal reflected from the first object 6 using the third ultrasound receiver 12.3. The maximum of the third waveform 20.3 specifically represents a third echo of the first ultrasound signal.

[0069] As can be seen in Figures 2a, 2b, and 2c, the first ultrasound waves 10.1 are received later by the second ultrasound receiver 12.2 than by the first ultrasound receiver 12.1. Likewise, the first ultrasound waves 10.1 reflected by the object 6 are received later by the third ultrasound receiver 12.3 than by the second ultrasound receiver 12.2. This is because the second ultrasound receiver 12.2 is located farther from the first ultrasound transmitter 11.1 than the first ultrasound receiver 12.1 is from the first ultrasound transmitter. The same applies to the second ultrasound receiver 12.2 compared to the third ultrasound receiver 12.3.

[0070] The evaluation unit 4 can preferably, within the framework of an evaluation of the echoes, which in the application shown in Figs. 1 and 2 comprise the first, second, and third echoes, calculate a first set 101 of signal transit times, hereinafter referred to simply as transit times, for the first ultrasound signal. The respective transit time of the first set 101 is the time that elapses between the emission of the first ultrasound signal, i.e., the emission of the first ultrasound waves 10.1, and the reception of the reflected first ultrasound signal by the respective ultrasound receiver 12.1, 12.2, 12.3.

[0071] For example, the control unit 3 can send a start signal to the evaluation unit 4 to measure the transit times when the control unit 3 detects the first ultrasound transmitter 11.1. to emit the first ultrasound wave 10.1 at a first transmission time. The evaluation unit 4 can, in particular, determine a first transit time t.1:1 of the first set 101, which requires the first ultrasound signal to travel from the first ultrasound transmitter 11.1 to the first ultrasound receiver 12.1 via reflection from the object 6, is calculated in the form of a difference between the first reception time 22.1 and the first transmission time. Similarly, the evaluation unit 4 can calculate a second travel time t 1 2 of the first sentence 101 based on the first transmission time and the second reception time 22.2. In the same way, the evaluation unit 4 can determine a third transit time t. 1 3 of the first sentence 101 based on the first transmission time and the third reception time 22.3.

[0072] Similarly, the evaluation unit 4 can determine a second set 102 of transit times for the second ultrasound signal and a third set 103 of transit times for the third ultrasound signal, depending on a second transmission time of the second ultrasound signal and a third transmission time of the third ultrasound signal and further reception times.

[0073] For the second and third ultrasound signals received by the ultrasound receivers 12.1, 12.2, 12.3, a corresponding further first time course of a further first amplitude signal is obtained, as well as a further second time course of a further second amplitude signal and a further third time course of a further third amplitude signal. The further reception times can be determined analogously from the maxima of these further time courses. The aforementioned first further Kl module can be configured to measure the propagation times of the first set 101, the second set 102, and / or the third set 103 as a function of the received signals from receivers 12.1, 12.2, 12.3.3, in particular depending on their amplitude signals, especially depending on the maxima of the amplitude signals, and the transmission times of the ultrasonic waves, such as the first, second, and third transmission times. In particular, the aforementioned second further Kl module can be configured to detect useful signals in the form of the maxima of the amplitude signals, such as the first maximum of the first amplitude signal 21.1, depending on the received signals, such as the first amplitude signal 21.1.

[0074] According to one variant, transmitters 11.1, 11.2, and 11.3 can emit the first, second, and third ultrasound signals, respectively, within a time interval. This can be a Increase the accuracy in detecting object 6, especially if object 6 is not moving in relation to the ultrasonic measuring unit 1.

[0075] According to another variant, the first transmitter 11.1 can transmit the first ultrasound signal in a first time interval, the second transmitter 11.2 the second ultrasound signal in a second time interval, and the third transmitter 11.3 the third ultrasound signal in a third time interval. In this variant, the receivers 12.1, 12.2, and 12.3 can detect the first, second, and third echoes, respectively, within the first time interval. Echoes of the second ultrasound signal can be received by the receivers 12.1, 12.2, and 12.3 within the second time interval, and echoes of the third ultrasound signal can be received by the receivers 12.1, 12.2, and 12.3 within the third time interval. This could particularly increase the accuracy in detecting object 6 if object 6 is moving relative to the ultrasound measuring unit 1.

[0076] In particular, the relative velocity of object 6 to the ultrasonic measuring unit 1 can be calculated in this case. By comparing the reception times at which the first, second, and third ultrasonic signals are received by a single ultrasonic receiver, for example, the second ultrasonic receiver 12.2, the evaluation unit 4 can determine changes in transit times. Based on these changes in transit times, the evaluation unit 4 can calculate the relative velocity. Alternatively or additionally, the evaluation unit 4 can be configured to track the aforementioned possible reflection area of ​​object 6 over time, based on the recorded echoes at the different time intervals, as in the "tracking" described above.

[0077] According to a possible embodiment of the Kl module 5 shown in Fig. 3, the Kl module 5 can be configured to calculate a first feature vector 111 as a function of the first set 101 of transit times. The Kl module 5 can generate the first feature vector, in particular, within the first time interval. The first feature vector 111 comprises values ​​that describe the shape, position, direction of motion, and / or velocity of the object 6 and, for the sake of simplicity, are each represented by a rectangle in Fig. 3. For example, the first feature vector 111 can comprise a value of an x-coordinate and a value of a y-coordinate of a coordinate system 100 shown in Fig. 1. These values ​​can represent the position of a first reflection point 121 approximated by the Kl module 5 within a drawing plane of the Specify in Fig. 1 in the coordinate system 100 within the first time interval.

[0078] Similarly, as shown in Fig. 4, the Kl module 5 can calculate a second feature vector 112 within the second time interval as a function of the second set 102 of runtimes. Analogous to the first set 101, the second set 102 of runtimes can determine a first runtime t. 2 1 , a second runtime t2.2 ur| d a third runtime t 2 3 The second feature vector 112 can be determined analogously to the transit times of the first set 101, with receivers 12.1, 12.2, 12.3 receiving the second ultrasound signal instead of the first. Similarly, the second feature vector 112 can include the x-coordinate and y-coordinate values ​​of the first reflection point 121 approximated using the Kl module 5, thus specifying the position of the first reflection point 121 in the coordinate system 100 within the second time interval.

[0079] Similarly, as shown in Fig. 5, the Kl module 5 can calculate a third feature vector 113 within the third time interval as a function of the third set 103 of runtimes. Analogous to the first set 101, the third set 103 of runtimes can determine a first runtime t. 3 1 , a second term t 3 2 and a third term t 3 3 The third feature vector 113 can be determined analogously to the transit times of the first set 101, with receivers 12.1, 12.2, 12.3 receiving the third ultrasound signal instead of the first. Similarly, the third feature vector 113 can include the x-coordinate and y-coordinate values ​​of the first reflection point 121 approximated using the Kl module 5, thus specifying the position of the first reflection point 121 in the coordinate system 100 within the third time interval.

[0080] Fig. 6 shows another application of the ultrasonic measuring unit 1, in which another object 60 is located in front of the ultrasonic measuring unit 1. The other object 60 can be, for example, another vehicle or a person. In this application, the first, second, and third ultrasonic waves 10.1, 10.2, 10.3 are reflected at a surface 61 of the other object 60 facing the ultrasonic measuring unit 1. In this application, the values ​​of the feature vectors 111, 112, 113 can each specify a slope and a position of a straight line 62 in the coordinate system 100.

[0081] Fig. 7 shows in an abstract form how the Kl module 5 processes the transit times and / or distances calculated using the transit times and the speed of sound. The Kl module 5 can generally be configured to calculate an output data set 72 based on an input data set 71. In this process, the Kl module 5 specifically determines values ​​of the output data set 72 based on values ​​of the input data set 71 using the mathematical functions of the Kl module 5 described above. The values ​​of the input data set 71 can, in particular, include the first set 101, the second set 102, and / or the third set 103 of the runtimes.

[0082] The output data set 72 can comprise the first, second, and / or third feature vector 111, 112, 113. The control unit 3 is specifically configured to repeatedly trigger the transmitters 11.1, 11.2, 11.3 to emit the first, second, and third ultrasound signals, respectively, over several time intervals. These multiple time intervals can, for example, be several hundred, and can include, in particular, the aforementioned first, second, and third time intervals. The evaluation unit 4 is specifically configured to generate a corresponding input data set for each time interval, analogous to the input data set 71 shown in Fig. 7, and to transmit it to the control module 5. Each input data set includes, in particular, a set of transit times or distances, such as the first set 101.The AI ​​module 5 is specifically configured to calculate a respective output data set, which can be configured analogously to the output data set 72 shown in Fig. 7, depending on the respective input data set. The respective output data set can, in particular, contain a respective feature vector, which can be configured analogously to the first feature vector 111. This makes it possible, with the aid of the AI ​​module 5, to determine a respective set of values ​​for the respective time interval to describe the shape, position, direction of movement, and / or speed of object 6 or further object 60. The evaluation unit 4 is specifically configured to store the respective feature vector and, depending on a stored set of feature vectors, to generate an image of the environment 13 of the vehicle 2. The image includes, in particular, points and lines that can be specified based on the stored set of feature vectors.

[0083] According to one possible embodiment, the evaluation unit 4 can temporarily store the run signals or run distances over several time intervals and generate a summary set of run times depending on the stored run times. In this embodiment, the input data set 71 can comprise the summary set of run times or run distances. Furthermore, it can be provided that the Kl module 5, depending on The input data set 71, with its summary set of transit times or distances, is used to calculate the output data set 72 such that it contains values ​​describing a point and a line, or multiple points, or multiple lines for object 6 or further object 60. In this configuration, the Kl module 5 is specifically trained to output several of the feature vectors using the output data set 72, depending on transit times or distances generated by reflections of the ultrasound waves at object 6 or further object 60 at different time intervals.

[0084] Alternatively or additionally, the evaluation unit 4 can be configured to generate the input data set 71 in such a way that it contains the aforementioned selected subset of signal propagation times or signal paths. This would allow, in particular, signal propagation times originating from interference signals to be excluded from the calculation of the feature vectors.

[0085] Alternatively or additionally, the evaluation unit 4 can be configured to generate the input data set 71 such that the input data set 71 includes signal strengths of the echoes detected by the second ultrasonic transducers 12.1, 12.2, 12.3. The evaluation unit 4 can calculate the signal strengths, for example, in the form of a maximum value of the respective waveform 20.1, 20.2, 20.3 of the corresponding amplitude signal 21.1, 21.2, 21.3 of the respective receiver 12.1, 12.2, 12.3.

[0086] Alternatively or additionally, the input data set 71 can contain relative positions of the transmitters 11.1, 11.2, 11.3 and the receivers 12.1, 12.2, 12.3 to each other. The relative positions can be specified, for example, in the form of coordinates of the transmitters 11.1, 11.2, 11.3 and the receivers 12.1, 12.2, 12.3 within the coordinate system 100.

[0087] Alternatively or additionally, input data set 71 can contain timestamps of the ultrasound signals generated by control unit 3. The timestamps can indicate the first transmission time of the first ultrasound signal, the second transmission time of the second ultrasound signal, and / or the third transmission time of the third ultrasound signal.

[0088] Alternatively or additionally, input data set 71 can specify an assignment of the echoes to the respective second ultrasound transducers 12.1, 12.2, 12.3 that detect the echoes. The assignment of the echoes to the respective second ultrasound transducers 12.1, 12.2, 12.3 can be specified, for example, by the input data set 71 including information about which receiver 12.1, 12.2, 12.3 determines the respective transit time of the corresponding received ultrasound signal, in particular the first, second or third ultrasound signal.

[0089] Alternatively or additionally, input data set 71 can include an assignment of the echoes to the respective first ultrasound transducers 11.1, 11.2, 11.3 that emitted the echo. The assignment of the echoes to the respective transmitters 11.1, 11.2, 11.3 can, for example, be specified by input data set 71 including information about which transmitter 11.1, 11.2, 11.3 emitted the first, second, or third ultrasound signal. This can be particularly useful if the transmitters 11.1, 11.2, 11.3 emit the ultrasound signals at the same frequency but at different time intervals.

[0090] Furthermore, it may be provided that the input data set 71 includes the probability described above, with which a portion of the received signals are caused by reflections of the ultrasound signals on the roadway.

[0091] In a further embodiment, input data set 71 can include information about a result of the signal propagation tracking described above. For example, input data set 71 can indicate whether the respective signal propagation time or signal path is included in the subset of selected signal propagation times. Alternatively or additionally, the information about the tracking result can also include how often the potential reflection area was tracked.

[0092] The initial data set 72 can further include a probability value indicating the probability that the respective feature vector is correct. Furthermore, the initial data set can include information about signal paths corresponding to the respective signal propagation time or signal travel distance. Additionally, the respective feature vector can include information about a relative velocity between the ultrasonic measuring unit 1 and the object 6 or the further object 60. The relative velocity can be determined according to one of the variants described above. In particular, the information about the relative velocity can include information about a direction and a height of the relative velocity in the coordinate system 100.

[0093] Furthermore, the output data set 72 may include information about the height of at least part of object 6 or the further object 60, which is described by the respective feature vector. In particular, the output data set 72 may indicate in binary form whether the respective part or the entire object 6 or the entire further object 60 is located in a lower area, for example in an area between 0 and 50 cm, or in an upper area, for example in an area of ​​1 m or more above the roadway.

[0094] In addition to the AI ​​module 5, the evaluation unit 4 can include the aforementioned first and second AI modules. Alternatively or additionally, the evaluation unit 4 can include a third AI module, which is configured to approximate dimensions and / or the position of object 6, object 60, and / or a free space for parking in the vicinity 13, depending on the feature vectors, in particular the first feature vector 111, the second feature vector 112, and / or the third feature vector 113. Specifically, the third AI module can be configured to detect the shape of object 6, object 60, and / or the free space based on the feature vectors.Alternatively or additionally, the third further AI module can be designed to determine the dimensions and / or the position of object 6, further object 60 and / or free space depending on the detected shape of object 6, further object 60 and / or free space.

[0095] Those skilled in the art will understand that aspects of the present invention may be implemented as a device, a method, a computer program, or a computer program product. Accordingly, aspects of the present invention may take the form of a purely hardware embodiment, a purely software embodiment (including firmware, software in memory, microcode, etc.), or an embodiment combining software and hardware aspects, all of which may be generally referred to herein as a "circuit," "module," or "system." Furthermore, aspects of the present invention may take the form of a computer program product, which is carried by a computer-readable medium or by several computer-readable media in the form of computer-executable code. A computer program also comprises computer-executable code. "Computer-executable code" may also be referred to as "computer program instructions." TI Any combination of one or more computer-readable media may be used. The computer-readable medium may be a computer-readable signaling medium or a computer-readable storage medium. A "computer-readable storage medium," as used herein, comprises a physical storage medium capable of storing instructions executable by a processor of a computer device. The computer-readable storage medium may be referred to as a computer-readable non-volatile storage medium. The computer-readable storage medium may also be referred to as a tangible computer-readable medium. In some embodiments, a computer-readable storage medium may also be capable of storing data that enables access to it by the processor of the computer device.Examples of computer-readable storage media include, but are not limited to: a floppy disk, a magnetic hard disk, a solid-state hard disk, flash memory, a USB flash drive, random access memory (RAM), read-only memory (ROM), an optical disk, a magneto-optical disk, and the processor's register file. Examples of optical disks include compact discs (CDs) and digital versatile disks (DVDs), for example, CD-ROM, CD-RW, CD-R, DVD-ROM, DVD-RW, or DVD-R discs. The term computer-readable storage medium also refers to various types of recording media suitable for being accessed by the computer device via a network or communication link. For example, data can be accessed via a modem, the internet, or a local area network.Computer-executable code running on a computer-readable medium may be transmitted via any suitable medium, including but not limited to wireless, wired, fiber optic, RF, etc., or any suitable combination of the foregoing media.

[0096] A computer-readable signal medium can contain a propagated data signal that includes the computer-readable program code, for example, in a baseband signal or as part of a carrier signal (carrier wave). Such a propagation signal can be in any form, including, but not limited to, an electromagnetic form, an optical form, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium that is not a computer-readable storage medium and that can transmit, propagate, or transport a program for use by or in conjunction with a system, device, or apparatus for executing instructions.

[0097] "Computer memory" or "memory" is an example of a computer-readable storage medium. Computer memory is any memory that is directly accessible to a processor.

[0098] "Computer data storage" or "data storage" is another example of a computer-readable storage medium. Computer data storage is any non-volatile, computer-readable storage medium. In some embodiments, computer memory can also be computer data storage, or vice versa.

[0099] A "processor," as used herein, comprises an electronic component capable of executing a programmatically or machine-executable instruction or computer-executable code. References to the computing device comprising a "processor" should be interpreted to mean that it may include more than one processor or processing cores. The processor may, for example, be a multi-core processor. A processor may also refer to a collection of processors within a single computer system or distributed across multiple computer systems. The term computing device or computer should also be interpreted to mean, possibly, a collection or network of computing devices or computers, each comprising a processor or processors.The computer executable code can be executed by multiple processors, which may be located within the same computer device or even distributed across multiple computers.

[0100] Computer-executable code may comprise machine-executable instructions or a program that causes a processor to perform an aspect of the present invention. Computer-executable code for performing operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object-oriented programming language such as Java, Smalltalk, C++, or similar languages, and conventional procedural programming languages ​​such as the programming language "C" or similar programming languages, and translated into machine-executable instructions. In some cases, the computer-executable code may be in the form of a higher-level programming language or in a pre-translated form, and used in conjunction with an interpreter that generates the machine-executable instructions. The computer executable code can run entirely on a user's computer, partially on the user's computer as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, via the internet using an internet service provider).

[0101] Computer program instructions can be executed on one processor or on multiple processors. In the case of multiple processors, these can be distributed across several different entities (e.g., clients, servers). Each processor could execute a portion of the instructions intended for its respective entity. Therefore, when referring to a system or procedure that encompasses multiple entities, the computer program instructions are understood to be adapted to be executed by a processor assigned to or associated with each entity.

[0102] Aspects of the present invention are described with reference to flowchart representations and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the invention. It is noted that each block or parts of the blocks of the flowcharts, representations, and / or block diagrams can be executed by computer program instructions, optionally in the form of computer-executable code. It is further noted that combinations of blocks in different flowcharts, representations, and / or block diagrams can be combined, provided they are not mutually exclusive.These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, or other programmable data processing device to create a device such that the instructions executed through the processor of the computer or other programmable data processing device generate means for performing the functions / steps specified in the block or blocks of the flowcharts and / or block diagrams.

[0103] These computer program instructions can also be written on a computer-readable format. The instructions stored on the computer-readable medium must be capable of controlling a computer or other programmable data processing equipment or other devices to function in a particular manner, such that the instructions stored on the computer-readable medium produce a manufactured product, including instructions that implement the function / step specified in the block(s) of the flowcharts and / or block diagrams.

[0104] The computer program instructions can also be stored on a computer, other programmable data processing devices, or other devices to cause the execution of a series of process steps on the computer, other programmable data processing devices, or other devices to generate a process executed on a computer, such that the instructions executed on the computer or other programmable devices generate procedures for implementing the functions / steps specified in the block or blocks of the flowcharts and / or block diagrams.

Claims

REQUIREMENTS 1. Ultrasonic measuring unit (1), in particular for use in a vehicle (2), comprising a control unit (3) and an evaluation unit (4), wherein — the control unit (3) is designed to control first ultrasound transducers (11) to send ultrasound signals and, depending on received signals generated by second ultrasound transducers (12), to detect echoes generated due to signal paths of the ultrasound signals, — the evaluation unit (4) has a trained AI module (5) and is configured to process the received signals to detect an object (6; 60) causing the echoes, wherein the processing includes at least generating feature vectors (111, 112, 113) using the AI ​​module (5) depending on the received signals, wherein the feature vectors (111, 112, 113) describe a shape, position, direction of movement and / or speed of the object (6; 60), and — the control unit (3) or the evaluation unit (4) is configured to determine a signal propagation time or signal path corresponding to the signal path for the respective echo, depending on the received signals, and the Kl module (5) is configured to determine the feature vectors (111, 112, 113) depending on the signal propagation times or signal paths.

2. Ultrasound measuring unit (1) according to claim 1, wherein the Kl module (5) is configured to determine the feature vectors as a function of — Signal strengths of the echoes detected using the second ultrasound transducer and / or — relative positions of the first ultrasound transducer (11) and the second ultrasound transducer (12) to each other and / or — timestamps of the ultrasound signals generated by the control unit (3), wherein the timestamps each indicate a transmission time of the respective ultrasound signal, and / or — an assignment of the echoes to the respective second ultrasound transducers (12) that detect the echo and / or — an assignment of the echoes to the respective first ultrasound transducers emitting the echo (11).

3. Ultrasound measurement unit (1) according to claim 1 or 2, wherein the Kl module (5) is trained using training data sets and the training data sets each comprise a training input data set and a training output data set corresponding to the training input data set, wherein the respective training input data set comprises training signal transit times or training signal travel distances of a training ultrasound signal reflected at a training object, and the respective training output data set comprises at least one training feature vector which is generated by multilateration using relative positions of the first ultrasound transducers (11) and the second ultrasound transducers (12) to each other and depending on the training signal transit times or training signal travel distances of the respective training input data set corresponding to the respective training output data set.

4. Ultrasound measuring unit (1) according to one of the preceding claims, wherein the control unit (3) is configured to detect the echoes in the form of useful signals and the evaluation unit (4) is configured to detect the useful signals as a function of the received signals and to determine the signal transit times or signal travel distances as a function of the useful signals.

5. Ultrasound measuring unit (1) according to one of the preceding claims, wherein the control unit (3) is configured to control the first ultrasound transducers (11) such that the first ultrasound transducers (11) emit the ultrasound signals at different times, and the evaluation unit (4) is configured to associate respective parts of the signal transit times or signal trajectories with the respective emitted ultrasound signal and to group them in a respective set of signal transit times or signal trajectories and to select a subset of the signal transit times or signal trajectories depending on differences between the signal transit times or signal trajectories of the different sets, and the K-module (5) is configured to determine the feature vectors (111, 112, 113) depending on the selected subset of the signal transit times or signal trajectories.

6. Ultrasound measuring unit (1) according to one of the preceding claims, wherein the second ultrasound transducers (12) are configured to generate the received signals at different time intervals, and the Kl module (5) is configured to during the different time intervals to determine the feature vectors (111, 112, 113) depending on the received signals, and the evaluation unit (4) is set up to generate an image of the vehicle's environment depending on the feature vectors (111, 112, 113) determined at the different time intervals.

7. Ultrasound measuring unit (1) according to one of the preceding claims, wherein the evaluation unit (4) is configured to determine a probability that part of the received signals are caused by reflections of the ultrasound signals on a roadway, and the Kl module (5) is configured to determine the feature vectors (111, 112, 113) depending on the probability.

8. Ultrasound measuring unit (1) according to one of the preceding claims, wherein the Kl module (5) comprises at least one artificial neural network.

9. Method for operating an ultrasonic measuring unit (1), in particular for use in a vehicle, wherein the ultrasonic measuring unit (1) comprises a control unit (3) and an evaluation unit (4) with a trained AI module (5), the method comprising the following steps: — Controlling first ultrasound transducers (11) to send ultrasound signals using the control unit (3); — Detection of echoes generated due to signal paths of the ultrasound signals depending on received signals generated using second ultrasound transducers (12); — Processing the received signals to detect an object causing the echoes, wherein the processing includes at least generating feature vectors (111, 112, 113) using the Kl module (5) depending on the received signals, wherein the feature vectors (111, 112, 113) describe a shape, position, direction of movement and / or speed of the object (6; 60); — Determining a signal propagation time or signal path for the respective echo depending on the received signals using the control unit (3) or the evaluation unit (4); — Determining the feature vectors (111, 112, 113) as a function of the signal propagation times or signal paths using the Kl module (5).

10. The method of claim 9, the method further comprising the following additional steps: — Generating training input data sets, wherein each training input data set contains training signal transit times or training signal travel distances of a training ultrasound signal reflected from a training object; — Generating a training output data set corresponding to the respective training input data set, wherein the respective training output data set has at least one training feature vector, the generation of the respective training output data set comprising generating the respective training feature vector by means of multilateration using relative positions of the first ultrasound transducers and the second ultrasound transducers to each other and depending on the training signal transit times or training signal travel distances of the respective training input data set corresponding to the respective training output data set; — Training the Kl module (5) using training data sets, wherein each training data set comprises the respective training input data set and the respective training output data set corresponding to the respective training input data set.

11. Computer program product comprising instructions executable by a processor, wherein the execution of the instructions causes the processor to perform the method according to claim 9 or 10.

Images