WO2022057979A1 - Verfahren zur bereitstellung einer maschinell gelernten steuerfunktion zur fahrzeugsteuerung anhand bereitgestellter fahrzeugsensordaten - Google Patents
Verfahren zur bereitstellung einer maschinell gelernten steuerfunktion zur fahrzeugsteuerung anhand bereitgestellter fahrzeugsensordaten Download PDFInfo
- Publication number
- WO2022057979A1 WO2022057979A1 PCT/DE2021/100760 DE2021100760W WO2022057979A1 WO 2022057979 A1 WO2022057979 A1 WO 2022057979A1 DE 2021100760 W DE2021100760 W DE 2021100760W WO 2022057979 A1 WO2022057979 A1 WO 2022057979A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- control function
- data set
- driving scenarios
- vehicle
- control
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000012549 training Methods 0.000 claims abstract description 46
- 238000013528 artificial neural network Methods 0.000 claims abstract description 44
- 238000012360 testing method Methods 0.000 claims abstract description 24
- 230000001965 increasing effect Effects 0.000 claims abstract description 12
- 238000011156 evaluation Methods 0.000 claims description 25
- 230000009467 reduction Effects 0.000 claims description 14
- 238000012795 verification Methods 0.000 claims description 11
- 210000002569 neuron Anatomy 0.000 claims description 10
- 230000006870 function Effects 0.000 description 66
- 230000008859 change Effects 0.000 description 19
- 238000004590 computer program Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 231100001261 hazardous Toxicity 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
-
- 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
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/02—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to ambient conditions
-
- 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
-
- 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
- B60W2050/0001—Details of the control system
- B60W2050/0019—Control system elements or transfer functions
- B60W2050/0028—Mathematical models, e.g. for simulation
-
- 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 for providing a machine-learned control function for vehicle control using provided vehicle sensor data.
- Vehicles today have control units that perform control functions specified by a system designer.
- the vehicle has a large number of sensors which provide vehicle sensor data for the control function, on the basis of which the control function generates control commands.
- the driver of the vehicle often has no direct influence on the control command generated by the control function.
- control functions are cruise control, distance control or a lane change warning system, which use vehicle sensor data from radar, lidar or a camera.
- control commands that can be output are defined by the developer of the computer program in such a way that the control commands are explicitly assigned to predetermined scenarios.
- This way of providing a control function for vehicle control is not suitable for complex
- reality provides complex driving scenarios that would have to be considered by a developer of the computer program in order to provide a control function that does not issue a hazard-triggering control command.
- this can only be carried out to a very limited extent and for a few driving scenarios, which is why the control function provided in this way is prone to errors under real conditions, in particular to unknown driving situations.
- Another method for providing a control function includes providing a control function based on an artificial neural network.
- the artificial neural network is trained on a training data set, which depicts as many categorically different driving scenarios as possible, in order to independently derive rules according to which the defined control commands are executed. This enables complex control functions to be provided, since no specific rules need to be implemented in the computer program product.
- methods for providing machine-learned control functions for vehicle control are error-prone, since the artificial neural network involved generates an output for scenarios that are not or only insufficiently represented by the training data set, which is associated with low confidence or high statistical uncertainty.
- the invention includes a method for providing a machine-learned control function for vehicle control using provided vehicle sensor data, comprising the method steps:
- the control function uses the at least one control command to control a function of a vehicle. All those functions are referred to as control functions that have an influence on the control (of the vehicle), whether through the provision of information or specific instructions for action.
- a specific action instruction and/or a signal for triggering a control function, which is executed automatically, is referred to as a control command.
- the artificial neural network may be implemented on a vehicle control unit or may be implemented on a server external to the vehicle and interact with the vehicle control unit.
- the evaluation catalog can be designed in such a way to categorize driving scenarios into known or unknown, non-hazard-triggering or hazard-triggering and safe or unsafe scenarios and then to evaluate them.
- Unknown driving scenarios are driving scenarios that are not represented by the training data set. Unknown driving scenarios are driving scenarios that were not considered by the developer, included in the evaluation catalog and/or in the training data set.
- Unsafe driving scenarios are scenarios for which the artificial neural network generates an output that is associated with low confidence or high statistical uncertainty. Unsafe driving scenarios can lead to hazard-triggering control commands from the control function.
- the training data record can depict a simulated journey of the vehicle.
- the complexity of the control function can be reduced if the execution of the control function, for example on a control unit of the vehicle, is either too slow or the memory requirement of the control function itself is too large.
- the complexity reduction of the control function can lead to a reduction or avoidance of over-fitting by the artificial neural network to improve the control function.
- the complexity of the control function can be increased if the control function cannot be adequately adapted (trained) to the driving scenarios provided in the training data set, which can manifest itself, for example, in incorrect behavior of the control function.
- the method comprises the method step of iteratively repeating method steps D) and/or E) and/or F) with the extended training data set and/or test data set and/or the reduced-complexity control function and/or the increased-complexity control function.
- the iterative repetition can be carried out, for example, until all driving scenarios are assessed as not causing a hazard based on the assessment catalogue.
- the evaluation of the driving scenarios using the evaluation catalog includes the determination of unknown driving scenarios.
- the evaluation of the driving scenarios using the evaluation catalog includes the determination of unsafe driving scenarios.
- the training data record and/or the test data record is expanded using the driving scenarios rated as unknown. This enables the successive and targeted expansion of the training data set and training of the control function for driving scenarios relevant to increasing safety.
- the training data record and/or the test data record is expanded using the driving scenarios assessed as unsafe. This enables the successive and targeted expansion and training of the control function in order to avoid control commands that could trigger a hazard.
- the control commands are based on outputs from the artificial neural network, with the output from the artificial neural network being designated as safe if it is met with sufficient confidence or statistical certainty.
- the method includes the method step of analytical verification of the control function, taking into account a defined value range of parameters.
- the value range of parameters can relate to the vehicle sensor data and/or data derived from vehicle sensor data.
- Methods for analytical verification are, for example, known methods such as "Reluplex” or "ReluVal”.
- the method includes the method step of probabilistic verification of the control function.
- Probabilistic verification means the data-driven evaluation and optimization of the test case coverage.
- Test case coverage means quantifying the rules learned by the artificial neural network through inductive and deductive analysis.
- the reduction in complexity of the control function includes the simplification of the artificial neural network.
- a simplification of the artificial neural network can be done in such a way as to decrease a computing and/or memory resource required for the artificial neural network.
- the simplification includes the reduction of the neurons and/or the layers of the artificial neural network.
- the complexity reduction of the control function includes the reduction of vehicle sensor data to be provided.
- the reduction can represent the lowering of the frequency with which vehicle sensor data are provided.
- the computing and storage capacity in a control unit of a vehicle is designed using the artificial neural network.
- the artificial neural network is designed (in terms of complexity) based on the computing and storage capacity of the control unit of a vehicle.
- control function is provided in a vehicle and tested during operation.
- vehicle can be a real or virtual/simulated vehicle.
- control function can be tested during operation in that the control function simulates (generates) the control commands to be output and these are not actually executed by the vehicle.
- control function is provided in a vehicle and the training data record and/or test data record is expanded using driving scenarios provided during (or through) a journey.
- vehicle can be a real or virtual/simulated vehicle.
- the driving scenarios provided during a journey for the expansion of the training data set can be evaluated using the evaluation catalogue.
- the invention includes a vehicle with a control unit that executes a control function that is provided according to the method.
- a method for providing a machine-learned control function for vehicle control based on provided vehicle sensor data is described, the control function being a lane change warning system based on an artificial neural network for predicting a lane change maneuver of surrounding vehicles (LCMP; engl .: lane change maneuver predictor).
- the LCMP is provided with vehicle sensor data from radar and lidar sensors and cameras.
- the process comprises the process steps:
- LCMP lane change maneuver of surrounding vehicles
- Unsafe driving scenarios are scenarios for which the LCMP generates an output that is associated with low confidence or high statistical uncertainty.
- potentially dangerous scenarios can be used to check safety with reference to criteria from the evaluation catalogue, such as a minimum distance between vehicles in the area.
- the training data set can include simulations of driving situations, for example.
- test data set comprising driving scenarios and evaluating the driving scenarios using the evaluation catalogue
- the test data set can, for example, include simulations of driving situations, in particular those that are not represented by the training data set.
- the evaluation of the driving scenarios using the evaluation catalog can include the classification into safe or unsafe and known or unknown driving scenarios.
- complexity can be reduced by reducing the number of layers and/or by reducing the number of neurons per layer of the artificial neural network.
- the complexity can be increased by increasing the number of layers and/or by increasing the number of neurons per layer of the artificial neural network.
- the original artificial neural network can be replaced by the artificial neural network with reduced/increased complexity. For example, this can complexity-reducedZ-increased artificial neural network can be retrained on the provided training data set without using the original artificial neural network for its training, or the original artificial neural network can be used in the process.
- Method steps D) and/or E) and/or F) are repeated iteratively, with the training data set and/or test data set extended in F) being used after each repetition and/or the control function with reduced or increased complexity in F) being used. This ensures that the training data set is successively optimized for training the control function, the number of unsafe driving scenarios is reduced and the number of safe driving scenarios is increased in order to improve the control function.
- the reduction in complexity of the control function includes the simplification of the artificial neural network by reducing the number of neurons and the layers of the artificial neural network. Furthermore, the complexity reduction of the control function includes the reduction of vehicle sensor data to be provided. Using the example of the lane change warning system, data from radar and lidar sensors and cameras are then provided at a reduced frequency.
- Unknown driving scenarios are those that are not or only insufficiently represented by the training data set. Unknown driving scenarios can be specifically generated.
- An example of a method for the targeted generation of unknown driving scenarios is "DeepXplore".
- the input data for the artificial neural network are modified using gradient methods. This happens in such a way that an incorrect classification is forced by the artificial neural network.
- new driving scenarios previously unknown to the artificial neural network are systematically generated. This process occurs under the additional constraint that the neuron coverage increases continuously.
- the neuron coverage indicates how many of the neurons were active after running through a test data set, measured relative to the total of all neurons in the artificial neural network. A neuron counts as active if its activation value exceeds a previously defined threshold.
- Evaluating the driving scenarios using the evaluation catalog includes determining unsafe driving scenarios.
- unsafe driving scenarios are those for which the LCMP generates an output associated with low confidence or high statistical uncertainty.
- Unsafe driving scenarios can be specifically generated.
- An example of a method for the targeted generation of driving scenarios that trigger danger is the gradient method “Fast Gradient Sign”, with the help of which the input data from the driving scenarios are manipulated in such a way that the neural network is prompted to make an incorrect prediction, which systematically uncovers this driving situation that triggers the risk.
- the resulting driving scenarios represent new driving scenarios for which the neural network generates an uncertain output.
- LCMP lane change maneuver of surrounding vehicles
- the training data set and the test data set are expanded using the driving scenarios rated as unknown or unsafe, whereupon method steps D), E) and F) are repeated iteratively.
- the extended training data set and the extended test data set are used for each repetition and/or the control function reduced in complexity in F) is used.
- the method also includes the method step of analytical verification of the control function, taking into account a defined value range of parameters.
- the value range of parameters relates to the vehicle sensor data and data derived from vehicle sensor data.
- the analytical verification of the control function taking into account a defined
- the value range of parameters is based on parameterizable driving scenarios. For the example of the lane change warning system based on an artificial neural network for predicting a lane change maneuver of surrounding vehicles (LCMP), a parameterizable driving scenario is the following:
- the vehicle to be controlled is driving on a three-lane highway in the middle lane and another vehicle is moving at a longitudinal speed between 100 km/h and 120 km/h and at a lateral speed between 2 km/h and 4 km/h over a time interval of 1 s from the right lane towards the center lane.
- the specified speeds and times define the defined value range of the parameters.
- the analytical verification can be carried out using known methods such as "Reluplex” or "ReluVal".
- An advantage of this method step is the statement as to whether the artificial neural network of the control function in driving scenarios, which are characterized by values of the considered value range of parameters, leads to an output that does not result in a hazardous situation.
- the calculation and storage capacity of the vehicle's control unit is designed using the artificial neural network.
- the central processing unit and the main memory of the control unit are suitably designed.
- the trained control function is made available in a vehicle and tested during operation.
- the driving scenarios provided during a journey can expand the training data set and/or test data set, with the previous method steps being able to be repeated.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Automation & Control Theory (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Human Computer Interaction (AREA)
- Combined Controls Of Internal Combustion Engines (AREA)
- Traffic Control Systems (AREA)
- Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
Abstract
Description
Claims
Priority Applications (8)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR1020237012740A KR20230088719A (ko) | 2020-09-16 | 2021-09-15 | 제공된 차량 센서 데이터를 기반으로 하는 차량 제어를 위한 기계 학습 제어 기능을 제공하는 방법 |
CN202180062963.2A CN116157309A (zh) | 2020-09-16 | 2021-09-15 | 根据所提供的车辆传感器数据提供用于车辆控制的机器学习的控制功能的方法 |
EP21786759.7A EP4214642A1 (de) | 2020-09-16 | 2021-09-15 | Verfahren zur bereitstellung einer maschinell gelernten steuerfunktion zur fahrzeugsteuerung anhand bereitgestellter fahrzeugsensordaten |
AU2021343608A AU2021343608A1 (en) | 2020-09-16 | 2021-09-15 | Method for providing a machine-learned control function for vehicle control on the basis of available vehicle sensor data |
US18/245,696 US20230359892A1 (en) | 2020-09-16 | 2021-09-15 | Method for providing of a machine-learned control function for vehicle control on the basis of provided vehicle sensor data |
CA3190157A CA3190157A1 (en) | 2020-09-16 | 2021-09-15 | Method for providing a machine-learned control function for vehicle control on the basis of provided vehicle sensor data |
JP2023540984A JP2023542434A (ja) | 2020-09-16 | 2021-09-15 | 提供された車両センサデータに基づいて車両制御のための機械学習された制御機能を提供する方法 |
IL301179A IL301179A (en) | 2020-09-16 | 2021-09-15 | A method for providing a machine-learned control function to control a vehicle based on available vehicle sensor data |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE102020124174.7A DE102020124174A1 (de) | 2020-09-16 | 2020-09-16 | Verfahren zur Bereitstellung einer maschinell gelernten Steuerfunktion zur Fahrzeugsteuerung anhand bereitgestellter Fahrzeugsensordaten |
DE102020124174.7 | 2020-09-16 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022057979A1 true WO2022057979A1 (de) | 2022-03-24 |
Family
ID=78080101
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/DE2021/100760 WO2022057979A1 (de) | 2020-09-16 | 2021-09-15 | Verfahren zur bereitstellung einer maschinell gelernten steuerfunktion zur fahrzeugsteuerung anhand bereitgestellter fahrzeugsensordaten |
Country Status (10)
Country | Link |
---|---|
US (1) | US20230359892A1 (de) |
EP (1) | EP4214642A1 (de) |
JP (1) | JP2023542434A (de) |
KR (1) | KR20230088719A (de) |
CN (1) | CN116157309A (de) |
AU (1) | AU2021343608A1 (de) |
CA (1) | CA3190157A1 (de) |
DE (1) | DE102020124174A1 (de) |
IL (1) | IL301179A (de) |
WO (1) | WO2022057979A1 (de) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102017006434A1 (de) | 2017-07-07 | 2019-01-10 | Wabco Gmbh | Verfahren zum vorausschauenden Bewerten einer aktuellen Fahrsituation sowie Bewertungsmodul |
US20190244103A1 (en) * | 2018-02-07 | 2019-08-08 | Royal Bank Of Canada | Robust pruned neural networks via adversarial training |
US20190310654A1 (en) * | 2018-04-09 | 2019-10-10 | SafeAI, Inc. | Analysis of scenarios for controlling vehicle operations |
DE102018116036A1 (de) | 2018-07-03 | 2020-01-09 | Connaught Electronics Ltd. | Training eines tiefen konvolutionellen neuronalen Netzwerks für individuelle Routen |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20170006434A (ko) | 2015-07-08 | 2017-01-18 | 주식회사 에이스침대 | 스프링조립체를 내장한 소파쿠션장치 |
KR20180116036A (ko) | 2017-04-14 | 2018-10-24 | 엘지전자 주식회사 | 이동 단말기 |
-
2020
- 2020-09-16 DE DE102020124174.7A patent/DE102020124174A1/de active Pending
-
2021
- 2021-09-15 JP JP2023540984A patent/JP2023542434A/ja active Pending
- 2021-09-15 IL IL301179A patent/IL301179A/en unknown
- 2021-09-15 KR KR1020237012740A patent/KR20230088719A/ko unknown
- 2021-09-15 EP EP21786759.7A patent/EP4214642A1/de active Pending
- 2021-09-15 US US18/245,696 patent/US20230359892A1/en active Pending
- 2021-09-15 AU AU2021343608A patent/AU2021343608A1/en active Pending
- 2021-09-15 WO PCT/DE2021/100760 patent/WO2022057979A1/de active Application Filing
- 2021-09-15 CN CN202180062963.2A patent/CN116157309A/zh active Pending
- 2021-09-15 CA CA3190157A patent/CA3190157A1/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102017006434A1 (de) | 2017-07-07 | 2019-01-10 | Wabco Gmbh | Verfahren zum vorausschauenden Bewerten einer aktuellen Fahrsituation sowie Bewertungsmodul |
US20190244103A1 (en) * | 2018-02-07 | 2019-08-08 | Royal Bank Of Canada | Robust pruned neural networks via adversarial training |
US20190310654A1 (en) * | 2018-04-09 | 2019-10-10 | SafeAI, Inc. | Analysis of scenarios for controlling vehicle operations |
DE102018116036A1 (de) | 2018-07-03 | 2020-01-09 | Connaught Electronics Ltd. | Training eines tiefen konvolutionellen neuronalen Netzwerks für individuelle Routen |
Non-Patent Citations (1)
Title |
---|
ARDI TAMPUU ET AL: "A Survey of End-to-End Driving: Architectures and Training Methods", ARXIV.ORG, CORNELL UNIVERSITY LIBRARY, 201 OLIN LIBRARY CORNELL UNIVERSITY ITHACA, NY 14853, 13 March 2020 (2020-03-13), XP081621133 * |
Also Published As
Publication number | Publication date |
---|---|
CA3190157A1 (en) | 2022-03-24 |
IL301179A (en) | 2023-05-01 |
DE102020124174A1 (de) | 2022-03-17 |
CN116157309A (zh) | 2023-05-23 |
US20230359892A1 (en) | 2023-11-09 |
AU2021343608A1 (en) | 2023-03-23 |
KR20230088719A (ko) | 2023-06-20 |
JP2023542434A (ja) | 2023-10-06 |
EP4214642A1 (de) | 2023-07-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
DE102018128289B4 (de) | Verfahren und vorrichtung für eine autonome systemleistung und zur einstufung | |
EP3899682B1 (de) | Überwachung von auf neuronalen netzwerken basierten fahrfunktionen | |
DE102017200180A1 (de) | Verfahren und Testeinheit zur Bewegungsprognose von Verkehrsteilnehmern bei einer passiv betriebenen Fahrzeugfunktion | |
DE102019124018A1 (de) | Verfahren zum Optimieren von Tests von Regelsystemen für automatisierte Fahrdynamiksysteme | |
AT523834B1 (de) | Verfahren und System zum Testen eines Fahrerassistenzsystems | |
WO2021058223A1 (de) | Verfahren zur effizienten, simulativen applikation automatisierter fahrfunktionen | |
EP4193135A1 (de) | Computerimplementiertes verfahren zum bereitstellen eines test-verlaufs zu testender verkehrsszenarien | |
DE102010013943A1 (de) | Verfahren und Vorrichtung für eine Funktionsprüfung einer Objekt-Erkennungseinrichtung eines Kraftwagens | |
DE102020120141A1 (de) | Verfahren zum Optimieren von Tests von Regelsystemen für automatisierte Fahrdynamiksysteme mittels probabilistisch prädizierter Systemantworten | |
DE102017214531A1 (de) | Verfahren und Vorrichtung zum Betreiben eines Kraftfahrzeugs in einem automatisierten Fahrbetrieb sowie Kraftfahrzeug | |
DE102020209680B3 (de) | Signalverarbeitungspfad, Vorrichtung zur Umfelderkennung und Verfahren zur Validierung eines automatisiert betreibbaren Fahrsystems | |
EP4212980A1 (de) | Fahrassistenzeinrichtung und verfahren zum durchführen einer wenigstens teilautomatischen fahrzeugfunktion in abhängigkeit von einer zu bewertenden fahrstrecke | |
DE102021202813A1 (de) | Verfahren, Vorrichtung und Computerprogramm für eine Unsicherheitsbewertung einer Bildklassifikation | |
WO2021094222A1 (de) | Verfahren zur abschätzung einer abdeckung des raums von verkehrsszenarien | |
EP3959660A1 (de) | Training trainierbarer module mit lern-daten, deren labels verrauscht sind | |
WO2022057979A1 (de) | Verfahren zur bereitstellung einer maschinell gelernten steuerfunktion zur fahrzeugsteuerung anhand bereitgestellter fahrzeugsensordaten | |
DE102019213797A1 (de) | Verfahren zur Bewertung einer Sequenz von Repräsentationen zumindest eines Szenarios | |
EP4047316A1 (de) | Verfahren zum bewerten einer bekanntheit einer fahrstrecke und elektronische recheneinrichtung | |
WO2021089499A1 (de) | Verfahren und system zum prüfen einer automatisierten fahrfunktion durch reinforcement-learning | |
DE102020212921A1 (de) | Verfahren, Computerprogramm und Vorrichtung zum Bewerten einer Verwendbarkeit von Simulationsdaten | |
DE102019130484A1 (de) | Verfahren und Vorrichtung zum Anlernen eines Ensembles von neuronalen Netzen | |
DE102010025977A1 (de) | Verfahren zur Erkennung von Fahrmanövern von Kraftfahrzeugen | |
WO2023275401A1 (de) | Simulation von verkehrsteilnehmern mit emotionen | |
DE102022200158A1 (de) | Verfahren zur Qualifizierung einer Steuerung mittels eines Closed-Loop Simulationsverfahrens | |
WO2024170027A1 (de) | Simulationsbasierte parametervoreinstellung für dynamische fahrfunktionen |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21786759 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 3190157 Country of ref document: CA |
|
ENP | Entry into the national phase |
Ref document number: 2023540984 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202327017504 Country of ref document: IN |
|
ENP | Entry into the national phase |
Ref document number: 2021343608 Country of ref document: AU Date of ref document: 20210915 Kind code of ref document: A |
|
ENP | Entry into the national phase |
Ref document number: 2021786759 Country of ref document: EP Effective date: 20230417 |