EP3966743A1 - Überwachung eines ki-moduls einer fahrfunktion eines fahrzeugs - Google Patents
Überwachung eines ki-moduls einer fahrfunktion eines fahrzeugsInfo
- Publication number
- EP3966743A1 EP3966743A1 EP20720772.1A EP20720772A EP3966743A1 EP 3966743 A1 EP3966743 A1 EP 3966743A1 EP 20720772 A EP20720772 A EP 20720772A EP 3966743 A1 EP3966743 A1 EP 3966743A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- module
- monitoring
- data stream
- training
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W50/04—Monitoring the functioning of the control system
- B60W50/045—Monitoring control system parameters
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W60/00—Drive control systems specially adapted for autonomous road vehicles
- B60W60/001—Planning or execution of driving tasks
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0475—Generative networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/09—Supervised learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/094—Adversarial learning
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W2050/0062—Adapting control system settings
- B60W2050/0075—Automatic parameter input, automatic initialising or calibrating means
- B60W2050/0083—Setting, resetting, calibration
- B60W2050/0088—Adaptive recalibration
Definitions
- the invention relates to a method and a device for monitoring a KI module which is part of a processing chain of a semi-automatic or
- Kl modules When developing functions for semi-automatic or automatic driving, the use of Kl modules is essential for coping with highly complex situations
- Model intrinsic confidences are subject to training bias and can be very incorrect, which means that adverse examples, namely small changes to the data signal that lead to a change in the model output, lead to misclassifications with a high intrinsic confidence.
- the document US 2016/0335536 A1 relates to a hierarchical neural network and a classifying learning method and a discriminatory method based on the hierarchical neural network. It includes a hierarchical neural
- Network device comprises a weight learning unit for generating loosely coupled parts by establishing links between partial nodes in the hierarchical network based on a checking matrix of an error correcting code and for learning weights between the coupled nodes. Furthermore, the device comprises the hierarchical neural network with an input layer, at least one
- the document US 2016/0071024 A1 relates to a multimodal data analysis device comprising instructions that are embodied in one or more non-volatile machine-readable storage medium, the multimodal data analysis device using a computer system with one or more computer devices to cause a set of time-variable instances of multimodal data are accessed with at least two different modalities, each instance of the multimodal data having a different time component and algorithmically one
- the method comprises two neural networks that perform a zero-sum game.
- One network, the generator creates candidates and the second neural network, the discriminator, evaluates the candidates.
- the generator typically maps a vector of latent variables to the desired result space.
- the aim of the generator is to learn Generate results according to a certain distribution.
- the discriminator is trained to extract the results of the generator from the data from the real,
- the generator's objective function is then to produce results that the discriminator cannot distinguish. As a result, the generated distribution should gradually adjust to the real distribution.
- the invention is based on the object of a method and a device for
- Discriminator is formed, the method having a training phase and a
- the KI module is checked by the monitoring module, in that the trained KI module generates suitable output signals for the driving function from the real input data stream and, in parallel, the real one
- Input data stream is fed to the discriminator of the trained monitoring module, which uses it to determine distances that are used to assess the
- Input data stream and to check the output signals of the Kl module are used.
- the numerical value expressing the distance, which represents a real number, is interpreted as the distance between the input data and the training data set, so that the situation on which the real input data stream of the inference phase is based is based on this
- Input data are "real", while a numerical value of d> 0.5 means that the
- Input data are, so to speak, "fake", that is they did not belong to the training data set.
- the latter indicates an untrained or learned environmental situation, for example a corner case.
- the definition of the value interval is not limited to the above-mentioned interval, but other intervals and assignments of the distance d to the respective interval are possible; it only needs to be recognizable from the distance d whether the currently assessed situation is typical for the training data sets learned or not, ie there is a case that deviates from the learned training data sets.
- the generator preferably uses a background data source to generate the false training data which are fed to the discriminator.
- the background data source can be generated, for example, by a suitable random generator.
- a first device for monitoring the input data stream of a KI module, the device being set up and designed to carry out the method explained above
- a Kl module which is part of a processing chain of a
- a monitoring module that uses real input signals to generate a differential signal that is used to monitor the input data stream of the KI module.
- Forms vehicle comprises three monitoring modules, which are each formed by a generative adversarial network comprising a generator and a discriminator, the method having a training phase and an inference phase and in the training phase
- the first monitoring module for monitoring the input data stream of the Kl module by real training data and false, from the generator of the first
- Monitoring module generated false generated training data is trained and generates first difference signals
- the KI module is trained using real training data and ground truth data, and output signals are generated,
- the second monitoring module for monitoring the output data stream of the KI module is trained by output data from the KI module and spurious, spuriously generated output data generated by the generator of the second monitoring module, and a second difference signal is generated,
- the third monitoring module for monitoring the output data stream of the Kl module through ground truth data and false data from the generator of the third
- Monitoring module generated false ground truth data is trained, and a third difference signal is generated,
- both the input data stream into the KI module and the output data stream from the KI module are monitored by
- Monitoring module is supplied, which determines distances from it, which to
- Monitoring module is supplied, which determines distances from it, which to
- Monitoring module is supplied, which determines distances from it, which to
- the first distance is therefore used to evaluate the typicality of the real input data stream, for example an environment sensor system, relative to the training data of the training phase, the second distance to evaluate the typicality of the output of the CI module relative to the output of the training and the third distance to evaluate the typicality of the output of the Kl module relative to the ground truth seen from the training.
- the three distances what was said for the distance in the first method applies. In other words, the distances are a measure of whether the data stream in question is similar to or deviates from the data streams in training. If the deviation is outside of a specification, it can be concluded that a situation has not been trained and that is to be responded to accordingly.
- the generators preferably generate spurious training data by means of a respective background data source, which are fed to the discriminators of the respective monitoring modules.
- training losses are determined from the three distances with regard to the respective training data, which are used to train the generators and discriminators of the respective monitoring module.
- a second device for monitoring the input data stream and the output data stream of a KI module, which is part of a
- a Kl module which is part of a processing chain of a
- Output signals of the Kl module generate respective distance signals, which for
- Monitoring of the input data stream and the output data stream of the Kl module can be used.
- the central object of the invention is a generative-discriminative situation assessment, which provides a control mechanism for KI modules along the
- the control mechanism is implemented by a monitoring unit for monitoring input and output data of a control module implemented by a Kl module for partially automatic or automatic driving.
- This situation assessment i.e. the control mechanism, measures the data stream flowing into and out of a KI module and measures a distance of the reference data distribution with which the KI module was originally developed and trained.
- the situation assessment uses a preferably triple generative discriminative approach, which is developed in the development phase of the CI model and is generally described below.
- Module output is compared with a basic truth called the Ground Truth.
- a loss value, the loess is calculated from the difference between the last two values and the module is adjusted accordingly to this loess.
- the control unit for monitoring the KI module preferably consists of three independent monitoring modules, the so-called generative-discriminative ones
- Distance measurement modules for measuring the distance to the training input, the distance to the
- Each of the three individual modules consists of a generator and a discriminator.
- the task of the generator is to create data that is as realistic as possible, namely input, output and ground truth.
- the task of the discriminator on the other hand, is to distinguish between real data and generated data. Its issue is therefore learning a measure to the generator.
- the discriminator of a distance measuring module is a Kl module, the training of which is run through during the training of the actual Kl module of the driving function.
- the data used and created in the training namely input, output or ground truth, are used as training data for the respective discriminator.
- Further training data for the discriminator are provided by the generator.
- the generator in turn uses a background data source, called latent space, and generates spurious data from it Training data. It can also be a KI module itself, such as a GAN approach from the area of machine learning, but also a simulation or a
- inference only the discriminators are used by the control unit according to the invention. You then evaluate and monitor the distance between the incoming and outgoing data stream in the actual KI module of the driving function and the reference data set during runtime.
- one of the two distance measurement modules can be used to monitor the
- Output current can be dispensed with.
- the distance measuring module which is responsible for the input data stream of the Kl module, can also be operated alone. In other words, in the simplest embodiment of the control unit, it only comprises the distance measuring module for the input data stream, which, however, leads to a reduced performance.
- the holistic approach of the control device according to the invention with at least two distance measuring modules allows the monitoring of relationships between incoming and outgoing data streams. Since the training of the individual modules can be carried out in parallel with the training of the actual KI module of the driving function and without significant additional technical effort, this represents a considerable savings potential compared to currently known solutions.
- Figure 2 shows the inference phase, i.e. the operating phase, a KI application with the trained discriminator of Fig. 1,
- FIG. 3 shows the training phase of a generative adversarial network including the KI application, with input signals and output signals of the KI application being monitored,
- FIG. 4 shows the inference phase of the KI application with the trained discriminators of FIGS. 3, and
- FIG. 5 shows an application of a generative adversarial network in the processing chain of a driving function in a schematic representation.
- the module M_IN includes a generator G_IN, which generates spurious training data from a background data source L_IN, the so-called “latent space for input data", the generator G_IN having the task of generating training data that is as realistic as possible.
- spurious training data generated by the generator G_IN are fed to a discriminator D_IN together with the real training data TD.
- the task of the discriminator D_IN is to differentiate between the false training data of the generator G_IN and the real training data. For this purpose, a distance DistJN is determined between the real and the false training data, which represents a measure of the distance between the training data TD and the generated data. This distance DistJN then becomes a quantity
- Loss of training TL_IN determines which is used to train the module MJN, so that due to the loss of training TL_IN, the generator GJN generates data that the
- FIG. 2 shows the use of the discriminator D_IN of the module M_IN trained in FIG. 1 in a K1 module K1, for example a K1 module of a vehicle.
- a real input IN is fed to a Kl module Kl, which generates an output OUT.
- the input IN is usually formed by sensor signals from one or more environment sensors that are processed in the Kl module. From these input signals IN, the Kl module Kl consequently generates output signals OUT, which are further processed in the control during automatic driving.
- the input signals IN can be the signals from a camera (not shown) and / or a radar sensor.
- the Kl module Kl is a module for object recognition
- the Kl module Kl should recognize and determine the objects in the vicinity of the vehicle from the signals IN received, so that the objects in the output OUT of the Kl module the surroundings of the vehicle, their spatial arrangement and the type of objects, for example vehicle, pedestrian or two-wheeler, are output, the type of objects representing a probability statement.
- This output OUT can then be fed to a scene recognition and scene prediction (not shown) so that ultimately an automatic driving function (not shown) can be controlled.
- the Kl module Kl can also be a module for lane detection, which from the signals IN of the environment sensors as the lanes of the road
- Output OUT determines on which the vehicle is located, so that when these results are merged with the results of object recognition, it can determine which object is on which lane.
- the list of the use of Kl modules for automatic driving is only to be regarded as an example and not as complete.
- the input signals IN are not only fed to the K1 module K1, but also to the discriminator D_IN of the trained distance measurement module M_IN in parallel.
- the trained discriminator determines a distance DistJN from the input signals IN, which indicates the distance between the input signals IN and known trained situations, so that the determined distance signals IN can be used to determine whether the surrounding situation corresponds to a known situation. In this way, deviations of the input signals from known situations can be determined via the output DistJN and the Kl module can be monitored
- Fig. 3 shows the training phase of a control unit for complete generative discriminative situation assessment in a schematic representation, which one
- Control mechanism for Kl modules along the processing chain of an automatic driving function the control mechanism being a program or a
- Control unit for monitoring input and output data of a Kl control module for semi-automatic or automatic driving is.
- the illustrated control unit ST with Kl module Kl includes, in addition to the Kl module Kl, three generative-discriminative distance measurement modules, namely the module M_IN for determining a distance to a training input, the module M_OUT for determining a distance to a training output and the module M_GT for determining a distance to a training ground truth, the modules being explained in detail below.
- the distance measurement module M_IN includes a generator G_IN, which generates the most realistic possible false input training data, with the generation of the false
- Training data of the generator G_IN is a background data source L_IN, the
- the module M_IN also includes a discriminator G_IN, which compares the false training data generated by the generator G_IN with real
- a distance DistJN As the output of the distance measurement module M_IN, the distance DistJN representing the distance between the spurious training data and the real training data, i.e. a measure of the expected association of the current spurious date with the amount of spurious data generated.
- a function called training loss TL_IN is determined, which is used to train the distance measuring module MJN with the generator GJN and the
- Kl module Kl of a partially or fully automatic driving function such as a parking assistant or the like
- input data such as the environment sensor system
- the training data TD are also fed to the Kl module Kl, which uses it to generate an output OUT that is responsible, for example, for controlling a driving function.
- This output OUT of the Kl module Kl is the discriminator D_OUT of a second Generative-discriminative distance measurement module M_OUT supplied.
- the second module M_OUT includes a generator G_OUT which, with the aid of a further background data source L_OUT, generates spurious training data that are fed to the discriminator D_OUT.
- the discriminator D_OUT generates a distance Dist_OUT from the real output data OUT of the Kl module Kl and the spurious training data generated by the generator G_OUT, the distance Dist_OUT being the distance between the spurious training data and the real one
- Output OUT of the Kl module Kl represents, that is, a measure of the expected association of the current spurious date with the amount of spurious data generated. From the distance Dist_OUT and the associated data, a function called training loss TLJDUT is determined with respect to the output OUT of the Kl module Kl, which is used to train the distance measuring module M_OUT with the generator G_OUT and the discriminator D_OUT.
- the output OUT is linked to ground truth data GT, so that a loss function TL that can be used to train the Kl module results, where TL stands for "Training Loess”.
- the mentioned ground truth data GT are fed to the discriminator D_GT of a third generative-discriminative distance measurement module M_GT.
- the third module M_GT includes a generator G_GT which, with the aid of a third background data source L_GT, generates spurious training data that are fed to the discriminator D_GT. From the real ground truth data GT and the false data generated by the generator G_GT
- the discriminator D_GT generates a distance Dist_GT for training data, the distance Dist_GT representing the distance between the spurious training data and real ground truth data GT, that is to say a measure of the expected association of the current spurious date with the amount of spurious data generated. From the distance Dist_GT and the associated data, a loss function TL_GT with respect to the ground truth data GT is determined, which is used to train the distance measuring module M_GT with the generator G_GT and the
- Discriminator D_GT is used.
- the inference phase that is to say the application phase, of the control unit ST with K1 module K1, with only the discriminators D_IN, D_OUT and D_GT of the three modules M_IN, M_OUT and M_GT being used in the inference phase.
- the input signals for example an environment sensor system, are fed as real input IN to both the Kl module Kl for processing and the discriminator D_IN of the module M_IN, which is responsible for assessing the input signals.
- a first distance DistJN with respect to the input signals IN is determined by the discriminator D_IN from the input signal IN.
- Dist_OUT With respect to the output signal OUT of the Kl module Kl and Dist_GT with respect to the ground truth GT shown in FIG. 3.
- the distances DistJN, Dist_OUT and Dist_GT generated by the three distance measurement modules M_IN, MOUT and M_GT monitor the data streams IN and OUT flowing into and out of the Kl module Kl and therefore provide information about the behavior of the Kl module, especially when the data stream is input IN is an untrained situation so that the output OUT of the Kl module also does not correspond to a trained situation, which is noticeable in the distances Dist_OUT and Dist_GT of the two discriminators D_OUT and D_GT.
- corner cases i.e. borderline cases
- the object recognition implemented by the K1 module (K1) receives input data IN from an environment sensor system US, which can include cameras, radar, lidar, ultrasonic sensors or the like.
- the Kl module Kl serving as object recognition recognizes from the input data IN it is supplied with objects, for example other vehicles, pedestrians, traffic signs, trees, curbs, etc., in the vicinity of the vehicle and outputs these objects with corresponding properties, such as relative speed, position relative to
- the input data stream IN into the Kl module Kl and the output data stream OUT from the Kl module Kl is monitored by the monitoring unit ÜW, the
- Monitoring unit ÜW is formed by the three discriminators D_IN, D_OUT and D_GT of the distance measuring modules M_IN, M_OUT and M_GT described in FIGS. 3 and 4.
- the discriminators D_IN, D_OUT, D_GT generate distances DistJN, Dist_OUT and Dist_GT which show the typicality of the input data stream and the output data stream of the Kl module Kl compared to the corresponding training data.
- both discriminators D_OUT and D_GT are not necessary for monitoring the output data stream OUT of the KI module; with sufficiently good training, one of the mentioned discriminators can be dispensed with.
- a respective threshold SJN, S_OUT and S_GT can now be defined with the following criteria:
- DistJN ⁇ SJN input data stream IN "real” in the sense of known, i.e. trained, DistJN>
- SJN input data stream IN "unknown” in the sense of not trained
- M_IN module for determining the distance between input data
- M_OUT module for determining the distance from output data
- M_GT module for determining the distance from ground truth data
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Automation & Control Theory (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Probability & Statistics with Applications (AREA)
- Traffic Control Systems (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| DE102019206720.4A DE102019206720B4 (de) | 2019-05-09 | 2019-05-09 | Überwachung eines KI-Moduls einer Fahrfunktion eines Fahrzeugs |
| PCT/EP2020/060593 WO2020224925A1 (de) | 2019-05-09 | 2020-04-15 | Überwachung eines ki-moduls einer fahrfunktion eines fahrzeugs |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP3966743A1 true EP3966743A1 (de) | 2022-03-16 |
Family
ID=70391096
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP20720772.1A Pending EP3966743A1 (de) | 2019-05-09 | 2020-04-15 | Überwachung eines ki-moduls einer fahrfunktion eines fahrzeugs |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US20220324470A1 (de) |
| EP (1) | EP3966743A1 (de) |
| CN (1) | CN113811894B (de) |
| DE (1) | DE102019206720B4 (de) |
| WO (1) | WO2020224925A1 (de) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102021205274A1 (de) | 2021-05-21 | 2022-11-24 | Robert Bosch Gesellschaft mit beschränkter Haftung | Sichere Steuerung/Überwachung eines computergesteuerten Systems |
| DE102021208047A1 (de) | 2021-07-27 | 2023-02-02 | Zf Friedrichshafen Ag | Verfahren und Computerprogramm zum Überwachen von Ausgaben eines Generators eines generativen kontradiktorischen Netzwerks |
| WO2024086771A1 (en) * | 2022-10-21 | 2024-04-25 | Ohio State Innovation Foundation | System and method for prediction of artificial intelligence model generalizability |
| DE102024201218A1 (de) * | 2024-02-09 | 2025-08-14 | Robert Bosch Gesellschaft mit beschränkter Haftung | Bewerten von Ergebnissen eines computerbasierten Maschinenlernsystems |
| DE102024208360A1 (de) * | 2024-09-03 | 2026-03-05 | Aumovio Autonomous Mobility Germany Gmbh | Eingabeüberwachung für ein Maschinenlernmodell |
Family Cites Families (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090034809A1 (en) * | 2007-08-01 | 2009-02-05 | Fujifilm Corporation | Abnormal tissue pattern detection apparatus, method and program |
| US10796223B2 (en) | 2014-02-10 | 2020-10-06 | Mitsubishi Electric Corporation | Hierarchical neural network apparatus, classifier learning method and discriminating method |
| US9875445B2 (en) | 2014-02-25 | 2018-01-23 | Sri International | Dynamic hybrid models for multimodal analysis |
| EP3995782B1 (de) * | 2016-01-05 | 2026-03-18 | Mobileye Vision Technologies Ltd. | Systeme und verfahren zur schätzung zukünftiger pfade |
| DE102016207276A1 (de) * | 2016-04-28 | 2017-11-02 | Bayerische Motoren Werke Aktiengesellschaft | Verfahren zur Freigabe einer Fahrfunktion in einem Fahrzeug |
| DE102016009655A1 (de) * | 2016-08-09 | 2017-04-06 | Daimler Ag | Verfahren zum Betrieb eines Fahrzeugs |
| US9989964B2 (en) * | 2016-11-03 | 2018-06-05 | Mitsubishi Electric Research Laboratories, Inc. | System and method for controlling vehicle using neural network |
| US20180336439A1 (en) | 2017-05-18 | 2018-11-22 | Intel Corporation | Novelty detection using discriminator of generative adversarial network |
| US11003995B2 (en) * | 2017-05-19 | 2021-05-11 | Huawei Technologies Co., Ltd. | Semi-supervised regression with generative adversarial networks |
| DE102017213119A1 (de) * | 2017-07-31 | 2019-01-31 | Robert Bosch Gmbh | Verfahren und Vorrichtung zum Ermitteln von Anomalien in einem Kommunikationsnetzwerk |
| US10753997B2 (en) * | 2017-08-10 | 2020-08-25 | Siemens Healthcare Gmbh | Image standardization using generative adversarial networks |
| DE102017215552A1 (de) * | 2017-09-05 | 2019-03-07 | Robert Bosch Gmbh | Plausibilisierung der Objekterkennung für Fahrassistenzsysteme |
| DE102017217443B4 (de) * | 2017-09-29 | 2023-03-02 | Volkswagen Ag | Verfahren und System zur Bereitstellung von Trainingsdaten zum maschinellen Lernen für ein Steuerungsmodell einer automatischen Fahrzeugsteuerung |
| US11120337B2 (en) * | 2017-10-20 | 2021-09-14 | Huawei Technologies Co., Ltd. | Self-training method and system for semi-supervised learning with generative adversarial networks |
| DE102017219441A1 (de) * | 2017-10-30 | 2019-05-02 | Robert Bosch Gmbh | Verfahren zum Trainieren eines zentralen Künstlichen-Intelligenz-Moduls |
| US11543830B2 (en) * | 2017-12-06 | 2023-01-03 | Petuum, Inc. | Unsupervised real-to-virtual domain unification for end-to-end highway driving |
| CN108520155B (zh) * | 2018-04-11 | 2020-04-28 | 大连理工大学 | 基于神经网络的车辆行为模拟方法 |
| DE102018112929A1 (de) * | 2018-05-30 | 2018-07-26 | FEV Europe GmbH | Verfahren zur Validierung eines Fahrerassistenzsystems mithilfe von weiteren generierten Testeingangsdatensätzen |
| GB201809604D0 (en) * | 2018-06-12 | 2018-07-25 | Tom Tom Global Content B V | Generative adversarial networks for image segmentation |
| KR102565279B1 (ko) * | 2018-08-23 | 2023-08-09 | 삼성전자주식회사 | 객체 검출 방법, 객체 검출을 위한 학습 방법 및 그 장치들 |
| EP3654246B1 (de) * | 2018-11-14 | 2024-01-03 | Bayerische Motoren Werke Aktiengesellschaft | Verfahren, fahrzeug, system und speichermedium zur anzeige eines anormalen fahrzeugszenarios unter verwendung eines encodernetzwerks und eines wiedergabepuffers |
| DE102019205519A1 (de) * | 2019-04-16 | 2020-10-22 | Robert Bosch Gmbh | Verfahren zum Ermitteln von Fahrverläufen |
-
2019
- 2019-05-09 DE DE102019206720.4A patent/DE102019206720B4/de active Active
-
2020
- 2020-04-15 US US17/613,018 patent/US20220324470A1/en active Pending
- 2020-04-15 CN CN202080034714.8A patent/CN113811894B/zh active Active
- 2020-04-15 EP EP20720772.1A patent/EP3966743A1/de active Pending
- 2020-04-15 WO PCT/EP2020/060593 patent/WO2020224925A1/de not_active Ceased
Also Published As
| Publication number | Publication date |
|---|---|
| DE102019206720A1 (de) | 2020-11-12 |
| CN113811894B (zh) | 2024-06-28 |
| DE102019206720B4 (de) | 2021-08-26 |
| CN113811894A (zh) | 2021-12-17 |
| US20220324470A1 (en) | 2022-10-13 |
| WO2020224925A1 (de) | 2020-11-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| DE102019206720B4 (de) | Überwachung eines KI-Moduls einer Fahrfunktion eines Fahrzeugs | |
| DE102020205315B4 (de) | Verfahren zur Klassifikation kritischer Fahrsituationen, Auswahl von der kritischen Fahrsituation ähnlichen Daten und Nachtraining des automatischen Systems | |
| DE102017217733A1 (de) | Prüfen eines neuronalen Netzes | |
| DE102018128289A1 (de) | Verfahren und vorrichtung für eine autonome systemleistung und zur einstufung | |
| DE102019208735B4 (de) | Verfahren zum Betreiben eines Fahrassistenzsystems eines Fahrzeugs und Fahrerassistenzsystem für ein Fahrzeug | |
| DE112017008149T5 (de) | VORRICHTUNG FÜR EINE VON EINEM FAHRZEUG AUSGEHENDE KOMMUNIKATION, VERFAHREN ZUM KOMMUNIZIEREN VON EINEM FAHRZEUG NACH AUßEN INFORMATIONSVERARBEITUNGSVORRICHTUNG UND PROGRAMM FÜR EINE VON EINEM FAHRZEUG AUSGEHENDE KOMMUNIKATION | |
| DE102010013943A1 (de) | Verfahren und Vorrichtung für eine Funktionsprüfung einer Objekt-Erkennungseinrichtung eines Kraftwagens | |
| WO2022028935A1 (de) | Computerimplementiertes verfahren zum bereitstellen eines test-verlaufs zu testender verkehrsszenarien | |
| EP3959660B1 (de) | Training trainierbarer module mit lern-daten, deren labels verrauscht sind | |
| DE102017006155A1 (de) | Verfahren zum Betreiben eines Sensorsystems eines Fahrzeugs | |
| WO2020061603A1 (de) | Verfahren und vorrichtung zur analyse eines sensordatenstroms sowie verfahren zum führen eines fahrzeugs | |
| DE102020203707A1 (de) | Plausibilisierung der Ausgabe neuronaler Klassifikatornetzwerke | |
| DE102020203829A1 (de) | Verfahren zum Überwachen eines Fahrzeugsystems zum Erkennen einer Umgebung eines Fahrzeugs | |
| DE102019209463A1 (de) | Verfahren zur Bestimmung eines Vertrauenswertes eines Objektes einer Klasse | |
| EP3985565B1 (de) | Verfahren und vorrichtung zum überprüfen eines beim teilautomatisierten oder vollautomatisierten steuern eines fahrzeugs verwendeten ki-basierten informationsverarbeitungssystems | |
| DE102021208349B3 (de) | Verfahren und Sensorsystem zum Zusammenführen von Sensordaten sowie Fahrzeug mit einem Sensorsystem zum Zusammenführen von Sensordaten | |
| DE102018005864A1 (de) | Verfahren zum Testen eines Totwinkelassistenzsystems für ein Fahrzeug | |
| EP4125005B1 (de) | Kontrastives representation learning für messdaten | |
| WO2023280531A1 (de) | Computerimplementiertes verfahren, computerprogramm und vorrichtung zum erzeugen einer daten-basierten modellkopie in einem sensor | |
| EP4000011B1 (de) | Komponentenbasierte verarbeitung von eingangsgrössen | |
| WO2021122337A1 (de) | Verfahren und vorrichtung zum erkennen einer entfremdung einer sensordatendomäne von einer referenzdatendomäne | |
| DE102022213244A1 (de) | Fahrzeugsystem zur Erhöhung der Sicherheit eines Fahrzeugs, Verfahren und Verwendung | |
| DE102022200735A1 (de) | Verfahren und Steuergerät zum Trainieren eines Objektdetektors | |
| DE102022107820A1 (de) | Verfahren und Prozessorschaltung zum Überprüfen einer Plausibilität eines Detektionsergebnisses einer Objekterkennung in einem künstlichen neuronalen Netzwerk sowie Kraftfahrzeug | |
| DE102019128223B4 (de) | Verfahren, Vorrichtungen und Computerprogramme |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
| 17P | Request for examination filed |
Effective date: 20211209 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAV | Request for validation of the european patent (deleted) | ||
| DAX | Request for extension of the european patent (deleted) | ||
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: EXAMINATION IS IN PROGRESS |
|
| 17Q | First examination report despatched |
Effective date: 20250424 |