Operation method of a detection device that uses interference processing with an artificial neural network.

Abstract

The present invention relates to a method (44) for operating a detection device, particularly a detection device (14) for a vehicle (10), a detection device, a driver assistance system, and a vehicle. In the method (44), at least one electromagnetic beam (26) is emitted by the detection device into a monitoring area of ​​the detection device. The detection device receives the at least one electromagnetic beam originating from the monitoring area and converts it into at least one detection variable (40) that can be processed by at least one evaluation device. Based on the at least one detection variable (40), at least one interference process is performed using at least one artificial neural network (CNN). In the at least one interference process, at least one interference analysis (46) is performed, in which the at least one detection variable (40) is examined for known interference patterns (52) of interference variables using the at least one artificial neural network (CNN). If at least one known interference pattern (52) is recognized, the at least one detection variable (40) is corrected by the interference variable belonging to the at least one recognized interference pattern (52).

Description

Technical Field

[0001] The present invention relates to a method for operating a detection device, particularly a detection device for a vehicle, at least one electromagnetic beam is transmitted into the monitoring area of the detection device using the detection device, at least one electromagnetic beam originating from the monitoring area is received using the detection device and converted into at least one capture variable that can be processed using at least one evaluation device, at least one interference processing process is executed based on at least one capture variable using at least one artificial neural network.

[0002] The present invention also relates to a detection device, particularly a detection device for a vehicle, the detection device comprising at least one transmission device for transmitting an electromagnetic beam into the monitoring area of the detection device, receiving an electromagnetic beam originating from the monitoring area and having at least one means for determining a capture variable from the received electromagnetic beam, having at least one means for executing an interference processing process based on the capture variable, wherein at least one means comprises at least one artificial neural network.

[0003] Furthermore, the present invention relates to a driver assistance system having at least one detection device, particularly a driver assistance system for a vehicle, the at least one detection device comprising at least one transmission device for transmitting an electromagnetic beam into the monitoring area of the at least one detection device, and receiving an electromagnetic beam originating from the monitoring area and having at least one means for determining a capture variable from the received electromagnetic beam, and at least one means for executing an interference processing process based on the capture variable, the at least one means comprising at least one artificial neural network.

[0004] Furthermore, the present invention relates to a vehicle having at least one detection device, wherein the at least one detection device is At least one transmitting device for transmitting an electromagnetic beam into the monitoring area of ​​at least one detection device, A means for receiving an electromagnetic beam originating from a monitoring area and determining capture variables from the received electromagnetic beam, A means for performing an interference processing process based on captured variables, comprising at least one means having at least one artificial neural network. [Background technology]

[0005] A method for operating a radar system having at least two radar sensors is known from German Patent Application Publication No. 102020107372(A1). In this case, in particular, the following steps are specified to be performed sequentially, preferably in the order described or in any order, and each and / or all of the steps can be repeated: The step of performing signal emission in the radar sensor in each case to emit at least one radar signal (via the radar sensor), preferably via at least one transmitting antenna of each radar sensor, in particular in the form of an electromagnetic signal radiated into the environment outside the radar sensor. The step of performing signal processing in each radar sensor to determine a capture variable specific to the radar signal emitted by each radar sensor, in particular, a capture variable specific to the emitted radar signal that is reflected by a target, delayed by the signal propagation time, and can be received, for example, by at least one receiving antenna of the radar sensor. - A step of performing an interference evaluation to detect at least one interference in a radar sensor in each case based on each capture variable, wherein the interference evaluation can preferably be performed centrally for all of the capture variables in each radar sensor, or individually for each capture variable. - A step of providing at least one, at least two, at least four, or at least six adaptive options to avoid at least one detected interference by adapting the signal emission. • A step of performing an evaluation of at least one adaptive option for each radar sensor, in particular for each radar sensor. • A step of performing adjustments to adaptive options between various radar sensors based on evaluation. The step of performing signal emission adaptation according to at least one adaptive option based on adjustment, in particular only if the adaptive option causes interference reduction for the majority of radar sensors, and / or by selecting the adaptive option that causes interference reduction for the majority of radar sensors.

[0006] The present invention is based on the objective of designing methods, detection devices, driver assistance systems, and vehicles of the type described above that enable improved determination of capture variables. In particular, the intent is to improve the signal-to-noise ratio of the capture variables. In particular, the intent is to improve the determination of capture variables, either alternatively or additionally, with respect to cost, especially with respect to the cost of materials, the cost of components and / or installation costs, and / or with respect to the effectiveness of the capture variables. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] German Patent Application Publication No. 102020107372(A1) Specification [Overview of the project]

[0008] According to the present invention, the objective is achieved by the fact that in this method, at least one interference analysis is performed during at least one interference processing process, in which at least one capture variable is examined for known interference patterns of interference variables using at least one artificial neural network, and if at least one known interference pattern is recognized, interference variables belonging to at least one recognized interference pattern are purged from at least one capture variable.

[0009] According to the present invention, at least one electromagnetic beam is transmitted to a monitoring area. At least one electromagnetic beam originating from the monitoring area is received using a detection device and converted into at least one capture variable.

[0010] Electromagnetic beams originating from the monitoring area can be favorably converted into capture variables in the form of an electrically received signal using means of a detection device, particularly using at least one receiving device which may have at least one antenna. The electrically received signal can be processed using electrical means, in particular electrical control and / or evaluation devices.

[0011] The electromagnetic beam that can be received by the detection device may have an electromagnetic echo beam, or may consist of an electromagnetic echo beam. The electromagnetic echo beam may originate from an electromagnetic beam transmitted using the detection device and reflected by at least one object. The capture variables determined from the echo beam are specific to the reflecting object. For better identification, capture variables that originate solely from the echo beam may also be called "echo-received variables."

[0012] Alternatively or additionally, the received electromagnetic beam may have, or consist of, an interference beam from an interference source. The interference beam received using a detection device is converted into a corresponding received variable, similar to an echo beam. For better identification, a received variable that originates solely from an interference variable may also be called an "interference variable."

[0013] The capture variable may be a superposition of any echo-receiving variable and any interference variable. If the interfering beam is not captured, the capture variable consists only of the echo variable, if present. If the echo beam is not captured, the capture variable consists only of the interference variable, if present.

[0014] According to the present invention, in order to reduce the influence of any interference sources and corresponding interference variables on the determination of information about the monitoring area, particularly object information about objects within the monitoring area, at least one interference processing process is performed based on at least one capture variable using at least one artificial neural network.

[0015] The object information may also include distance variables, direction variables, and / or velocity variables that characterize the distance, direction, and / or velocity of the object relative to the detection device or a corresponding reference point or reference system.

[0016] At least one interferometry analysis is performed during at least one interferometry process. During at least one interferometry analysis, at least one capture variable is examined for known interference patterns. Known interference patterns originate from interference variables known before performing the interferometry analysis. If a known interference pattern is recognized, the corresponding interference variable is purged from the capture variable.

[0017] Known interference sources may be external interference sources in particular. External interference sources may be other radiation sources, particularly radar sources, that transmit electromagnetic beams in the same wavelength range or overlapping wavelength range as the detection device according to the present invention.

[0018] Interfering beams from known interference sources can induce characteristic interference patterns in corresponding interference variables determined using a detection device. In particular, corresponding noise patterns from known interference sources can be identified in the determined capture variables. Therefore, when a known interference pattern is recognized, the corresponding interference variable can be purged from the capture variables. Thus, the overall signal-to-noise ratio of the capture variables, especially the echo-receiving variables contained within them, can be improved.

[0019] According to the present invention, machine learning is used to analyze the captured variables. Depending on the type, material, shape, X-Y-Z coordinates of the object, environmental conditions (e.g., rain), external noise, the dynamics of the environment or the object, and other factors, the shape of the received electromagnetic beam is different and also depends on the transmitted electromagnetic beam.

[0020] Interference analysis uses at least one artificial neural network. For this purpose, a suitable multi-layer neural network (deep neural network) having an input layer, several intermediate layers (hidden layers), and an output layer can be specified. For training, a detection device can be used to pre-record several different scenarios, which can be used to train the neural network. Different classes can be used depending on the type of application and the degree of automation, for example, from SAE level 0 to SAE level 4.

[0021] Using at least one interference analysis to purge at least one captured variable also makes it possible to determine sufficiently good data using a detection device with less accurate components. It is also possible to use components that may themselves be associated with more noise. Thus, simpler and less expensive components can be used throughout a detection device that can still be used to determine sufficiently good captured variables for an application and, in some cases, the corresponding degree of automation. Thus, the performance of the detection device can be improved by correcting the measured values using at least one interference analysis.

[0022] By improving the signal-to-noise ratio of the captured variables, the validity of the data determined therefrom can be improved. Thus, a higher safety level can be achieved using the detection device according to the present invention. The detection device according to the present invention can be used to determine data compliant with the automation levels SAE0 to 4 required for autonomous or semi-autonomous driving.

[0023] The present invention can be advantageously used in detection devices for vehicles, particularly automobiles. Advantageously, the present invention can be used in detection devices for land vehicles, particularly automobiles, trucks, buses, motorcycles, etc., aircraft, particularly drones, and / or ships. The present invention can also be used in detection devices for vehicles that can operate autonomously or at least semi-autonomously. However, the present invention is not limited to detection devices for vehicles. It can also be used in detection devices for steady-state operation in robots and / or machinery, particularly construction or transport machinery such as cranes and excavators.

[0024] The detection device can be advantageously connected to, or be part of, at least one electronic control device of a vehicle or machine, particularly a driver assistance system and / or chassis control system and / or driver information device and / or parking assistance system and / or gesture recognition system. In this way, at least part of the functions of the vehicle or machine can be performed autonomously or semi-autonomously using information acquired by the detection device.

[0025] The detection device can be used to capture stationary or moving objects, especially vehicles, people, animals, plants, obstacles, uneven road surfaces, especially potholes or rocks, road boundaries, traffic signs, open spaces, especially parking spaces, rainfall, and / or movement and / or gestures.

[0026] In one advantageous configuration of this method, at least one interference analysis can be repeatedly performed, and the purged capture variables determined from each interference analysis can be combined to form at least one combined capture variable. This makes it possible to further improve the signal-to-noise ratio of the capture variable.

[0027] At least one interferometry analysis can be advantageously performed 2 to 10 times, especially 4 times. The signal-to-noise ratio is further improved with each pass of at least one interferometry analysis. Performing the analysis 4 times improves the signal-to-noise ratio by a factor of 2.

[0028] A more advantageous configuration of this method allows multiple different electromagnetic beams to be transmitted to the same scene in the monitoring area, and the capture variables for each can be determined.

[0029] In this way, at least one interferometry analysis can be performed on at least some of the multiple capture variables determined, and the respective purged capture variables can be determined for at least some of the different electromagnetic beams transmitted.

[0030] At least some of the multiple purged capture variables determined in this way can be combined to form at least one combined capture variable.

[0031] This makes it possible to further improve the signal-to-noise ratio of the captured variable.

[0032] To capture the same scene, different electromagnetic beams can be transmitted within correspondingly smaller time windows.

[0033] Advantageously, multiple different electromagnetic beams can be transmitted sequentially to the same scene in the monitoring area. This avoids interference between the transmitted electromagnetic beams.

[0034] Four different electromagnetic beams can be advantageously transmitted to the same scene, their respective capture variables can be determined, and interference analysis can be performed for each. This allows for smaller time windows to be executed accordingly, and therefore, changes in the captured scene are kept as small as possible.

[0035] Different electromagnetic beams can differ in terms of shape, wavelength, pulse duration, transmission duration, transmission power, coding, and other characteristics. Corresponding echo-received variables can be better distinguished from arbitrary interference variables by varying the transmitted electromagnetic beam. Therefore, interference variables can be better identified and eliminated.

[0036] A more advantageous configuration of this method allows the use of artificial convolutional neural networks as at least one artificial neural network. Artificial convolutional neural networks (CNNs) are a machine learning concept inspired by biology, aimed at feature extraction. Since noise has patterns different from regular signals, particularly echo beam patterns, particularly object patterns, especially interference patterns, the captured variables that are captured after recognizing interference patterns can be reduced accordingly. In this case, it is even possible to change the electromagnetic beam transmitted to the monitoring area using a detector for scanning. In this way, at least one interference analysis can be used to determine which electromagnetic beam originates from the beam transmitted by the detector. Interference noise can be identified from this recognition and therefore can be removed computationally.

[0037] In a more advantageous configuration of this method, During at least one interferometry analysis, at least one capture variable can be initially examined for known object patterns caused by electromagnetic echo beams reflected by known objects. If at least one known object pattern is recognized, the echo capture variable corresponding to at least one known object pattern can be removed from at least one capture variable. Next, the at least one capture variable, from which at least one recognized echo capture variable has been removed, can be examined for known interference patterns of the interference variable using at least one artificial neural network. If at least one known interference pattern is recognized, the interference variables belonging to at least one recognized interference pattern can be purged from the original at least one capture variable, which could contain at least one echo capture variable. This can further improve the signal-to-noise ratio.

[0038] As a result of the fact that at least one captured variable is first removed using a known object pattern, the corresponding interference pattern can be identified more effectively. Therefore, the removal of the original captured variable from the interference variable can be improved.

[0039] To a certain extent, object patterns caused by objects whose object patterns are already known can be first removed from at least one original capture variable. The capture variable from which the object patterns have been removed can then undergo at least one further interference analysis, during which interference variables with known interference patterns can be identified. The identified interference variables can then be removed from at least one original capture variable, and thus this purged capture variable ideally contains only the echo variables of the objects, provided all interference variables have been identified.

[0040] In a more advantageous configuration of this method, A predefined interference pattern and / or possibly an object pattern can be used in at least one interference analysis. and / or Interference patterns and / or object patterns learned during the operation of the detection device can be used for at least one interference analysis. In this way, the method can be more flexibly accessed with a greater number of known interference patterns and / or known object patterns.

[0041] Predefined interference patterns and / or object patterns can be used in at least one interference analysis. These patterns can be pre-learned, particularly under laboratory conditions, and stored in a corresponding storage medium, especially in the storage medium of the detection device. In this way, the corresponding interference patterns and / or object patterns can be accessed more quickly when performing the method.

[0042] Alternatively or additionally, interference patterns and / or object patterns learned during operation can be used. This makes it possible to continuously increase the number of known interference patterns and / or object patterns. This also makes it possible to continuously improve the method.

[0043] In a more advantageous configuration of this method, The signal converted from the electromagnetic beam using the receiving device of the detection device, particularly the electrically received signal, can be used as a capture variable. and / or Object information about objects captured during measurement by the detection device can be used as capture variables, and the object information is determined from the received signal, particularly the electrical received signal, which is converted from the electromagnetic beam using the receiving device of the detection device. In this way, at least one interferometry analysis can be performed at an appropriate processing level.

[0044] The received signal can be advantageously used as a capture variable. Therefore, interferometry can be performed directly using the received signal at a lower processing level. In this way, the interfering variable can be removed very quickly.

[0045] Electrical received signals, particularly voltage variables, can be advantageously used as capture variables. Electrical received signals are generated during the conversion of an electromagnetic beam using means of a detection device, especially a receiving device. Electrical received signals can be processed using electrical means, especially an electrical evaluation device.

[0046] The received signal may be an echo-received signal originating from the echo beam, an interference signal originating from the interference beam, or a superposition of the echo-received signal and the interference signal.

[0047] Alternatively or additionally, object information can be used as capture variables. This allows for interferometry to be performed at a higher processing level. Therefore, the effectiveness of images containing object information, particularly depth images, can be improved.

[0048] In a more advantageous configuration of this method, it can be used to operate a detection device in the form of a radar sensor used to transmit an electromagnetic beam in the form of a radar beam.

[0049] Radar sensors are highly variable with respect to the transmitted radar beam. This also allows for changing the capture variables to improve the distinction between object patterns and interference patterns. Therefore, the same scene can be scanned, especially continuously, using different radar beams. Thus, the signal-to-noise ratio can be improved overall with purged capture variables.

[0050] At least one receiving device of a detection device, particularly a radar sensor, can be advantageously configured to receive an electromagnetic beam of the same type as the electromagnetic beam transmitted using the detection device, particularly a radar beam.

[0051] The wavelength range in which at least one receiving device can receive an electromagnetic beam can favorably include the wavelength range in which the electromagnetic beam, particularly a radar beam, is emitted using a detection device. This ensures that at least the echo of the transmitted electromagnetic beam, particularly the radar beam, can be reliably received.

[0052] A further advantageous configuration of this method allows for the further processing of at least one purged capture variable, and in some cases at least one purged combined capture variable, which can be used particularly for image processing. and / or At least one transmitting device and / or at least one receiving device of the detection device can be adjusted based on at least one purged capture variable, and in particular, possibly at least one purged combined capture variable.

[0053] At least one purged capture variable, and in some cases at least one purged combined capture variable, can be handled more favorably. This allows for further information to be obtained regarding the monitoring area.

[0054] At least one item of object information, in particular at least one distance variable, at least one direction variable, and / or at least one velocity variable, can be advantageously determined from at least one purged capture variable, and in particular possibly at least one purged combined capture variable. This allows for a more accurate characterization of the captured scene.

[0055] At least one purged capture variable, and in some cases at least one purged combined capture variable, can be advantageously subjected to image processing. This allows for the removal of further interference effects.

[0056] Alternatively or additionally, at least one transmitting device and / or at least one receiving device of the detection device can be tuned based on at least one purged capture variable, and in particular, possibly at least one purged combined capture variable. This makes it possible to adapt the performance of the detection device to general conditions.

[0057] Furthermore, according to the present invention, this objective is achieved by the fact that, in the case of a detection device, the detection device has at least some of the means for carrying out the method according to the present invention.

[0058] According to the present invention, the detection device has at least one interferometry means that can be used to perform interferometry according to the present invention.

[0059] The detection device may advantageously have at least one artificial neural network, particularly an artificial convolutional neural network. The capture variables can be inspected for known interference patterns of the interference variables using the artificial neural network when performing interference analysis, and if a known interference pattern is recognized, the interference variables belonging to the recognized interference pattern can be purged from the capture variables.

[0060] Interference patterns can be recognized even more effectively using artificial convolutional neural networks.

[0061] In one advantageous embodiment, the detection device may be a radar sensor. The monitoring area can be non-contactively monitored for objects using the radar sensor. The radar sensor can be variably adjusted based on the emitted radar beam. Thus, the shape of the radar beam, pulse duration, length, and / or coding can be changed, in particular. By transmitting different radar beams to the same scene, the radar sensor can thus be used to determine more capture variables for the same captured scene. Thus, the identification of interference patterns can be further improved.

[0062] This objective is further achieved, according to the present invention, by the fact that, in the case of a driver assistance system, the driver assistance system has at least some of the means for carrying out the method according to the present invention.

[0063] According to the present invention, the driver assistance system comprises at least one detection device and at least a portion of means for performing a method according to the present invention for operating the at least one detection device.

[0064] The vehicle can operate autonomously or semi-autonomously using the driver assistance system.

[0065] At least one monitoring area within the vehicle's environment and / or interior can be used to monitor objects using a detection device. Distance variables, direction variables, and / or velocity variables characterizing the distance, direction, and / or velocity of a captured object can be determined using at least one detection device. Information obtained using at least one detection device can be used in conjunction with a driver assistance system to operate the vehicle autonomously or semi-autonomously.

[0066] According to the present invention, a driver assistance system has at least a portion of the means for carrying out the method according to the present invention. At least one detection device of the driver assistance system may, advantageously, have at least a portion of the means for carrying out the method according to the present invention. Therefore, if at least one detection device is part of a driver assistance system, then a portion of the means of the at least one detection device for carrying out the method according to the present invention is also part of the driver assistance system, i.e., a portion of the means of the driver assistance system for carrying out the method according to the present invention. Thus, this applies to the means of a vehicle having at least one driver assistance system and / or at least one detection device.

[0067] This objective is further achieved, according to the present invention, by the fact that, in the case of a vehicle, the vehicle has at least some of the means for carrying out the method according to the present invention.

[0068] The vehicle may, advantageously, have at least one driver assistance system, in particular at least one driver assistance system according to the present invention. The driver assistance system can be used to operate the vehicle autonomously or semi-autonomously.

[0069] Alternatively or additionally, a vehicle may have at least one detection device, in particular at least one detection device according to the present invention. The environment of the vehicle and / or at least one monitoring area inside the vehicle may be used to monitor for objects.

[0070] At least one detection device, in particular at least one detection device according to the present invention, may be advantageously connected to or part of a driver assistance system, in particular at least one driver assistance system according to the present invention. In this way, information obtained using the at least one detection device can be used by a driver assistance system to operate a vehicle autonomously or semi-autonomously.

[0071] Furthermore, the features and advantages shown in relation to the method, detection device, driver assistance system, and vehicle according to the present invention, as well as their respective advantageous configurations, are applied in a mutually corresponding manner, and vice versa. Of course, individual features and advantages can be combined with each other, in which case even more advantageous effects may be obtained than the sum of the individual effects. [Brief explanation of the drawing]

[0072] Further advantages, features, and details of the present invention will become apparent from the following description, in which exemplary embodiments of the invention will be described in more detail with reference to the drawings. Those skilled in the art will also find it preferable to individually consider the features disclosed in combination in the drawings, specification, and claims, and to combine them to form further suitable combinations. The drawings are schematic.

[0073] [Figure 1] This is a front view of a vehicle equipped with a driver assistance system that includes a radar sensor. [Figure 2] Figure 1 shows a functional diagram of a driver assistance system equipped with a radar sensor. [Figure 3] Figures 1 and 2 show the time profiles of the electro-received signals determined from the radar echo signal and electromagnetic interference beam using the radar sensor receiving device, as well as the corresponding time profiles of the electro-echo received signal and electro-interference signal. [Figure 4] The time profile of the raw received signal from Figure 3 is shown. [Figure 5] The time profile of the echo received signal from Figure 3 is shown. [Figure 6] A flowchart illustrating the method for operating the radar sensor from Figures 1 and 2 is shown.

[0074] In the diagram, identical parts are denoted by the same reference numeral. [Modes for carrying out the invention]

[0075] Figure 1 shows a front view of a vehicle 10 in the form of an automobile. The vehicle 10 has a driver assistance system 12. The vehicle 10 can operate autonomously or semi-autonomously using the driver assistance system 12. Figure 2 shows a functional diagram of the driver assistance system 12.

[0076] The driver assistance system 12 includes a detection device in the form of a radar sensor 14. The driver assistance system 12 also has a central processing unit 16.

[0077] The radar sensor 14 is positioned, for example, on the front fender of the vehicle 10 and directed within the monitoring area 18 in the forward direction of travel of the vehicle 10. The radar sensor 14 may also be positioned at different locations on the vehicle 10 and oriented in different ways. The driver assistance system 12 may also have multiple radar sensors 14 that can be positioned at different locations on the vehicle 10 and oriented in different ways. Furthermore, the driver assistance system 12 may have different detection devices.

[0078] The present invention is illustrated by example using one radar sensor 14 shown in Figures 1 and 2. However, the present invention can also be used with other radar sensors or other detection devices that use an electromagnetic beam to monitor a corresponding area.

[0079] The radar sensor 14 includes, for example, a transmitting device 20 having a transmitting antenna Tx, for example, a receiving device 22 having a receiving antenna Rx, and a control / evaluation device 24 such as an electronic control and evaluation device.

[0080] The transmitting device 20 and the receiving device 22 are functionally connected to the control / evaluation device 24. This allows information to be exchanged between the transmitting device 20, the receiving device 22, and the control / evaluation device 24.

[0081] The control and evaluation device 24 is connected to the central processing unit 16 of the driver assistance system 12. This allows for the exchange of information between the radar sensor 14 or the control and evaluation device 24 and the central processing unit 16.

[0082] Furthermore, the radar sensor 14 may include multiple transmitting antennas Tx and multiple receiving antennas Rx. The radar sensor 14 may also be in the form of a multiple-input multiple-output (MIMO) radar sensor.

[0083] The transmitting device 20 can be used to generate an electrically scanned signal that can be transmitted into the monitoring area 18 using the transmitting antenna Tx, for example, as an electromagnetic scanning beam in the form of a radar signal 26. The radar signal 26 may be transmitted, for example, as a radar pulse in the form of a chirp. The transmitting device 20 can be used to modify the transmitted radar signal 26. For example, the shape, pulse duration, signal duration, and / or encoding of the radar signal 26 can be changed.

[0084] The radar signal 26 may be reflected by an object 28 located within the monitoring area 18.

[0085] The radar sensor 14 can be used to detect stationary or moving objects 28 such as vehicles, people, animals, plants, obstacles, uneven road surfaces such as depressions or rocks, road boundaries, traffic signs, empty spaces such as parking spaces, rainfall, and / or motion and / or gestures.

[0086] The radar signal 26 reflected by object 28 in the direction of the radar sensor 14 can be received as an electromagnetic beam in the form of a radar echo signal 30 using the receiving antenna Rx of the receiving device 22.

[0087] The received radar echo signal 30 can be converted into a capture variable in the form of an electrical echo received signal 38 using the receiving device 22. Figures 3 and 5 show the time profile of the exemplary electrical echo received signal 38 resulting from the radar echo signal 30 of the exemplary radar signal 26.

[0088] Object information about the captured object 28 can be determined based on the propagation time of the transmitted radar signal 26 until the corresponding radar echo signal 30 is received. For example, distance variables 32, direction variables, and velocity variables characterizing the distance, direction, and velocity of the captured object 28 in a reference system for the radar sensor 14 can be determined. Indirect or direct propagation time methods can be used. When using a MIMO radar sensor 14, the distance variable 32 can be determined from the phase difference between the electrical scanning signal used to generate the radar signal 26 and the electrical echo received signal 38 of the captured radar echo signal 30.

[0089] Object information is determined by the control and evaluation device 24.

[0090] In addition to the radar echo signal 30 originating from the captured object 28, the receiving antenna Rx of the receiving device 22 is also used to receive an electromagnetic interference beam 34 originating from, for example, an external interference source 42. The electromagnetic interference beam 34 is converted into an electrical interference signal 36 using the receiving device 22.

[0091] The interference source 42 may be, for example, another radar sensor that emits an interference beam 34 in the form of a radar beam. Figure 2 shows three interference sources 42 as an example, and their reference numerals are denoted by indices 1, 2, and 3 for better distinction. Accordingly, the reference numerals of the corresponding electrical interference signals 36, whose time profiles are shown in Figure 3, are indicated using indices 1, 2, and 3.

[0092] The electrical echo received signal 38 and the electrical interference signal 36, originating from the echo signal 30, are superimposed to form capture variables in the form of the electrical raw received signal 40. Figures 3 and 4 show, as an example, the time profiles of the raw received signal 40 for the scene shown in Figure 2 with three interference sources 421, 421 and 423.

[0093] The raw received signal 40 depends on the type, material, shape, and spatial position of the reflective object 28, for example, its position in a defined reference frame. Furthermore, the raw received signal 40 depends on environmental conditions, such as general precipitation, external noise, the environment, or the dynamics of the captured object 28. In addition, the raw received signal 40 depends on the radar signal 26 used.

[0094] For comparison, Figure 3 shows the time profiles of an exemplary electro-raw received signal 40, the corresponding echo received signal 38, and three electro-interference signals 361, 362, and 363.

[0095] The electrical interference signals 361, 362, and 363 originate from three interference sources 421, 422, and 423, each emitting electromagnetic interference beams 341, 342, and 343, respectively.

[0096] Figure 4 shows only the time profile of the raw received signal 40 from Figure 3. Figure 5 shows the time profile of the electrical echo received signal 38 from Figure 3 after the interference processing process, where interference signals 361, 362 and 363 were removed according to the method described in more detail below.

[0097] The interference signal 36 impairs the signal-to-noise ratio of the echo received signal 38. Consequently, the accuracy of the determined object information regarding the object 28 detected by the radar sensor 14 is impaired.

[0098] To enable the determination of the most accurate possible object information regarding object 28, such as the precise distance variable 32, precise direction variable, and / or precise velocity variable for object 28, it is necessary to improve the signal-to-noise ratio.

[0099] For this purpose, an interference processing process is performed in method 44 for operating the radar sensor 14. Method 44 is shown as a flowchart in Figure 6.

[0100] During the interference processing, the interference analysis 46 is performed using an artificial neural network. The neural network is implemented as, for example, a convolutional neural network (CNN).

[0101] For example, in method 44, four interference analyses 46 are performed. More or fewer interference analyses 46 can also be performed. The signal-to-noise ratio is improved by the number of interference analyses 46 performed.

[0102] For each interferometry analysis 46, a radar signal 26 is transmitted, a corresponding echo signal 30 is captured, and it is converted into a raw received signal 40. The four interferometry analyses 46 are performed at short time intervals on the same scene within the monitoring area 18. Each interferometry analysis 46 uses a different variation of the radar signal 26, and as a result, the four interferometry analyses 46 use four different variations of the radar signal 26. For ease of distinction, the reference codes for the four different variations of the radar signal 26 are given the following indices 1, 2, 3, and 4.

[0103] For clarity, the four interferometry analyses 46 are shown at the same level in the flowchart of Figure 6. The interferometry analyses 46 and the corresponding radar measurements are performed sequentially in time. The sequence and principle of the four interferometry analyses 46 are identical. Therefore, the same reference numerals are used in the figure. The interferometry analysis 46 of the radar signal 261 on the left side of Figure 6 will be described in more detail below using the example scene shown in Figure 2, in a manner that represents all four interferometry analyses 46.

[0104] In measurement step 48, radar measurements are performed using the radar signal 261. The corresponding echo signal 30 and interference beam 34 from the interference source 42, shown as an example in Figure 2, are received using the receiving antenna Rx of the receiving device 22 and converted into an electro-raw received signal 40. The time profile of the raw received signal 40 is shown in Figures 3 and 4.

[0105] The raw received signal 40 is sent to the neural network CNN.

[0106] Furthermore, known interference patterns 52 of known electrical interference signals and known object patterns 54 of known objects 28 are transmitted from pattern memory 50 to the neural network CNN. Pattern memory 50 is, for example, part of a control and evaluation device 24.

[0107] The interference pattern 52 is characterized by the time profile of the electrical interference signal 36. A known interference pattern 52 may be a pattern of interference signal 36 that normally occurs when the vehicle 10 is in operation. For example, a known interference signal 36 may originate from an interference beam 34 transmitted by a radar sensor of another vehicle.

[0108] The object pattern 54 is characterized by the time profile of the electrical echo received signal 38. A known object pattern 54 may be, for example, a pattern of the echo received signal 38 of an object 28 that normally occurs during the operation of the vehicle 10. A known object 28 may be, for example, a vehicle, a person, an animal, a plant, an obstacle, an uneven road surface such as a pothole or rock, a road boundary, a traffic sign, an empty space such as a parking space, etc.

[0109] Known interference patterns 52 and known object patterns 54 are predetermined by reference measurements using known interference sources 42 or known objects 28, for example, at the end of the production line, and stored in pattern memory 50. Reference measurements can be performed, for example, under laboratory conditions. Alternatively or additionally, known interference patterns 52 and / or known object patterns 54 may also be recorded, for example, “learned” during the normal operating conditions of the vehicle 10.

[0110] In the exemplary embodiment described, it is assumed that for the interference signal 36 from the scene in Figure 2, a corresponding known interference pattern 52 is stored in the pattern memory 50. It is also assumed that for the echo reception signal 38 of the object 28 shown therein, such as a road sign, a corresponding known object pattern 54 is stored in the pattern memory 50.

[0111] The raw received signal 40 is compared with known object patterns 54 in the neural network during the object purging step 56. A pattern recognition method can be performed for this purpose, for example. If a match is identified with the known object pattern 54, in this case the object pattern 54 of a road sign, the raw received signal 40 is reduced by the identified echo received signal 38 of the known object pattern 54, i.e., the road sign, and supplied to the interferometry step 60 as the reduced received signal 58.

[0112] In the interference analysis step 60, the reduced received signal 58 is compared with a known interference pattern 52. A pattern recognition method can be performed for this purpose, for example. If a match with a known interference pattern 52 is recognized, the original raw received signal 40 is reduced in the purging step 62 by the interference signal 36 of the corresponding known interference pattern 52. In the illustrated exemplary embodiment, the pattern of the interference signal 36 is caused by interference beams 341, 341, and 343 from three interference sources 421, 421, and 423 shown in the scene of Figure 2, and matches, for example, the corresponding known interference pattern 52 stored in the pattern memory 50. Thus, the original raw received signal 40 is reduced by the interference signals 361, 361, and 363.

[0113] After removing the effects of the recognized interference signals 361, 361 and 363, if all interference signals 361, 361 and 363 occurring during the measurement are identified using the known interference pattern 52, the interference is purged, and only the echo received signal 38 originating from the reflective object 28, i.e., the road sign, remains.

[0114] The purged echo-received signals 38 determined in each of the four exemplary interferometry analyses 46 are combined in the superposition step 64 to form a combined echo-received signal 66.

[0115] In the information determination step 68, object information about the captured object 28, such as the distance variable 32, the direction variable and / or velocity, is determined from the combined echo received signal 66.

[0116] Optionally, object information can undergo further processing, such as image processing.

[0117] Object information, such as the distance variable 32, is transmitted to the central processing unit 16 of the driver assistance system 12.

[0118] Optionally, the settings of the transmitting device 20 and / or receiving device 22 can be adapted to the current scene based on the combined echo received signal 66.

[0119] Instead of being performed based on the raw received signal 40 as capture variables, the interferometry 46 may also be performed based on object information, such as a distance variable 32, a direction variable, and / or a velocity variable, as capture variables. In this case, the object information is predetermined based on the corresponding raw received signal 40.

Claims

1. A method (44) for operating a detection device (14), particularly a detection device (14) for a vehicle (10), At least one electromagnetic beam (26) is transmitted to the monitoring area (18) of the detection device (14) using the detection device (14), At least one electromagnetic beam (30, 34) originating from the aforementioned monitoring area (18) 1 ,34 2 ,34 3 ) is received using the detection device (14) and converted into at least one capture variable (40) that can be processed using at least one evaluation device (24), At least one interference processing process is performed using at least one artificial neural network (CNN) based on at least one capture variable (40). At least one interference analysis (46) is performed during the at least one interference processing process, wherein in the interference analysis, at least one capture variable (40) is examined for known interference patterns (52) of the interference variable (36) using at least one artificial neural network (CNN), and if at least one known interference pattern (52) is recognized, the interference variable (36) belonging to at least one of the recognized interference patterns (52) is purged from the at least one capture variable (40), Furthermore, during the at least one interferometry analysis (46), the at least one capture variable (40) is first examined for known object patterns (54) caused by an electromagnetic echo beam (30) reflected by a known object (28), If at least one known object pattern (54) is recognized, the echo capture variable (38) corresponding to at least one known object pattern (54) is removed from at least one capture variable (40). Next, the at least one capture variable (58) from which the recognized echo capture variable (38) has been removed is examined for known interference patterns (52) of the interference variable (36) using at least one artificial neural network (CNN), and if at least one known interference pattern (52) is recognized, the interference variable (36) belonging to the at least one recognized interference pattern (52) is purged from the original at least one capture variable (40) which could contain the at least one echo capture variable (38). method.

2. The method according to claim 1, characterized in that the at least one interferometry (46) is repeatedly performed, and the purged capture variables (38) determined from each of the interferometry (46) are combined to form at least one combined capture variable (66).

3. Multiple different electromagnetic beams (26 1 , 26 2 , 26 3 , 26 4 ) is transmitted to the same scene in the monitoring area (18), and each capture variable (40) is determined. At least one interference analysis (46) is respectively executed for at least a part of the plurality of the captured variables (40) determined in this way, and a respective purged captured variable (38) is determined for at least a part of the transmitted different electromagnetic beams (26 1 , 26 2 , 26 3 , 26 4 ). The method according to claim 1, characterized in that at least some of the plurality of purged capture variables (38) determined in this manner are combined to form at least one combined capture variable (66).

4. The method according to claim 1, characterized in that an artificial convolutional neural network is used as at least one artificial neural network (CNN).

5. A predefined interference pattern (52) and / or object pattern (54) is used in the at least one interference analysis (46). and / or The method according to claim 1, characterized in that the interference pattern (52) and / or object pattern (54) learned during the operation of the detection device (14) are used in the at least one interference analysis (46).

6. The receiving device (22) of the detection device (14) is used to use the electromagnetic beam (30, 34 1 ,34 2 ,34 3 The received signal converted from ), particularly the electrical received signal, is used as the capture variable (40). and / or The method according to claim 1, characterized in that object information (32) relating to an object (28) captured during measurement by the detection device (14) is used as a capture variable (40), and the object information is determined from a received signal (38), particularly an electrical received signal, converted from an electromagnetic beam (30) using a receiving device (22) of the detection device (14).

7. The method according to claim 1, characterized in that the method (44) is used to operate a detection device (14) in the form of a radar sensor used to transmit an electromagnetic beam (26) in the form of a radar beam.

8. The method according to claim 1, wherein at least one purged capture variable (40) or at least one purged combined capture variable (66) is further processed.

9. The method according to claim 1, characterized in that at least one transmitting device (20) and / or at least one receiving device (22) of the detection device (14) is adjusted based on at least one purged capture variable (38) or at least one purged combined capture variable (66).

10. A detection device (14), particularly a detection device (14) for a vehicle (10), The device has at least one transmitting device (20) for transmitting an electromagnetic beam (26) into the monitoring area (18) of the detection device (14), Electromagnetic beams (30, 34) originating from the aforementioned monitoring area (18) 1 ,34 2 ,34 3 ) receives the received electromagnetic beam (30, 34 1 ,34 2 ,34 3 ) has at least one means (22) for determining the capture variable (40), The system has at least one means (24) for performing an interference processing process based on captured variables (40), wherein the at least one means (24) has at least one artificial neural network (CNN), The detection device (14) is characterized in that it has means for performing the method described in any one of claims 1 to 9.

11. The detection device (14) according to claim 10, characterized in that the detection device (14) is a radar sensor.

12. A driver assistance system (12) having at least one detection device (14), in particular a driver assistance system (12) for a vehicle (10), wherein the at least one detection device (14) is At least one transmitting device (20) for transmitting an electromagnetic beam (26) into the monitoring area (18) of the at least one detection device (14), Electromagnetic beams (30, 34) originating from the aforementioned monitoring area (18) 1 ,34 2 ,34 3 ) receives the received electromagnetic beam (30, 34 1 ,34 2 ,34 3 ) and at least one means (22) for determining the capture variable (40), A means (24) for performing an interference processing process based on capture variables (40), comprising at least one means (24) having at least one artificial neural network (CNN), A driver assistance system (12) characterized in that the driver assistance system (12) has means for performing the method described in any one of claims 1 to 9.

13. A vehicle (10) having at least one detection device (14), wherein the at least one detection device (14) At least one transmitting device (20) for transmitting an electromagnetic beam (26) into the monitoring area (18) of the at least one detection device (14), Electromagnetic beams (30, 34) originating from the aforementioned monitoring area (18) 1 ,34 2 ,34 3 ) receives the received electromagnetic beam (30, 34 1 ,34 2 ,34 3 ) and at least one means (22) for determining the capture variable (40), A means (24) for performing an interference processing process based on capture variables (40), comprising at least one means (24) having at least one artificial neural network (CNN), A vehicle (10) characterized in that it has means for performing the method described in any one of claims 1 to 9.

Images