AUMOVIO Germany Patente

Embodiments of the present disclosure relate to methods for safe motion planning. A computer program, a computer-readable data carrier and an apparatus are also disclosed. The method for safe motion planning comprises: obtaining, by a hypernetwork, an environment representation comprising one or more signed distance fields, SDFs, obtained for respective one or more time steps of a motion planning time horizon; generating, by the hypernetwork, parameters based on the environment representation; obtaining, by a main network, a residual of a Hamilton Jacobi, HJ, value function based on the generated parameters; applying a Rectified Linear Unit, ReLU, to the residual to obtain a nonnegative residual; determining an HJ value function as a subtraction of the nonnegative residual from an SDF of the one or more SDF; and using the determined HJ value function for safe motion planning.

A system and method for virtualizing an electronic control system (ECS) is disclosed. The method comprises abstracting signal properties in ECS components by categorizing signals into analog and digital groups to enable distinct processing paths; encapsulating peripheral models with standardized interfaces that match the respective ECS component interfaces; obtaining signal operations including logical operations for digital signals and mathematical operations for analog signals to simulate signal processing between components; executing said signal operations to evaluate signal behavior; and using the evaluation of the signal processing and/or the executed signal operations as input to a machine learning model configured to predict cascading malfunctions across multiple ECS components. The method further comprises adapting ECS components that are predicted to malfunction using software-based and/or hardware-based corrective actions. The system may include modules corresponding to each method step and may support semantic interpretation of signal data, dynamic validation, modular logging with interchangeable backends, and integration into continuous integration (CI) or hardware-in-the-loop (HiL) testing pipelines. The invention enables proactive fault detection, scalable ECS simulation, and enhanced system reliability in safety-critical domains.

Embodiments of the present disclosure relate to methods for virtualizing an electronic control system. A computer program, a computer-readable data carrier and an apparatus are also disclosed. The method comprising: abstracting, technical details of signals in ECS components into groups, including analog signals group and digital signals group; obtaining peripheral models associated with respective ECS components, wherein each peripheral model comprises interfaces that correspond to respective interfaces of the respective ECS component, each interface of the peripheral model being associated with a signal, which is an analog signal or a digital signal, matching an interface signal of the respective ECS component; obtaining signal operations that represent signal processing of the respective ECS components, said signal operations including logical operations for the digital signals, and mathematical operations for the analog signals; and executing the signal operations to evaluate the signal processing between the ECS components.

Embodiments of the present disclosure relate to methods for autonomous navigation and for enhancing the performance of a model predictive control, BMPC. A computer program, a computer-readable data carrier and an apparatus are also disclosed. The BMPC being configured to generate control commands to an autonomous system over successive control intervals. The method for autonomous navigation using the BMPC comprises: obtaining, for each control interval, candidate navigation corridors for the autonomous system; obtaining, using backward reachability analysis, respective backward reachable navigation corridors for at least two of the candidate navigation corridors; obtaining, for each control interval, a highest-progress corridor and a lowest-progress corridor from the backward reachable navigation corridors, wherein the highest-progress corridor is associated with the backward reachable navigation corridor where the autonomous system makes most longitudinal progress, and the lowest-progress corridor is associated with the backward reachable navigation corridor where the autonomous system makes least longitudinal progress; and

The application discloses a sliding screen assembly, including: a winding unit; a flexible screen (2), wound on the winding unit, the flexible screen being provided with a free end, and the free end being provided with a traction portion (22); and a flexible screen support frame (3) capable of being deployed and retracted; where the flexible screen support frame (3) is capable of abutting against the traction portion (22), and the flexible screen support frame (3) is deployed to pull the traction portion (22) to move in a forward direction, so that the winding unit unwinds the flexible screen (2), the flexible screen (2) is deployed, and the flexible screen support frame (3) supports the flexible screen (2); the flexible screen support frame (3) is retracted to drive the traction portion (22) to move in a reverse direction, so that the winding unit rewinds the flexible screen (2), and the flexible screen (2) is retracted. The application further discloses an automotive seat, an automotive sunroof, and an automobile.

The present application discloses a sliding screen assembly, including: a winding unit (10); a flexible screen (20) wound on the winding unit, the flexible screen being provided with a free end (30); a support plate (40) hinged to the free end, the support plate being configured to support the flexible screen unwound by the winding unit; a fixed base (50); and a sliding link mechanism (60) having one end slidably connected to the fixed base and the other end hinged to the support plate; where the sliding link mechanism slides in forward and reverse drive the support plate to move in forward and reverse directions, thereby driving the winding unit to unwind and rewind the flexible screen, so that the flexible screen is deployed and retracted. The present application further discloses an automotive seat, an automotive sunroof, and an automobile.

The present application discloses a sliding screen assembly, including a winding unit (1); a flexible screen (2) wound on the winding unit, where the flexible screen is provided with a free end (21); a support plate (3) hinged to the free end, where the support plate is used to support the flexible screen unwound by the winding unit; a fixed base (4); and a linkage (5) having an end hinged to the fixed base and the other end hinged to the support plate, where the linkage is rotatable in a forward direction and a reverse direction around a hinge point between the linkage and the fixed base (6a), to drive the support plate to rotate in the forward direction and the reverse direction, thereby driving the winding unit to unwind and wind the flexible screen, enabling the flexible screen to unfold and fold. The present application further discloses an automotive seat, an automotive sunroof, and an automobile.

Embodiments of the present disclosure relate to a method for generating training data for an application-specific task. A training method and a dataset used for training a machine learning model are also disclosed. A machine learning model, a computer program, a computer-readable data carrier and an apparatus are also presented. The method for generating the training data for the application-specific task comprises: fine-tuning a pre-trained machine learning model using a first dataset; obtaining, using said model, original embeddings for input data obtained from a second application-specific dataset; applying, to each of the obtained original embeddings, a corresponding augmentation function, thereby obtaining corresponding augmented embeddings; and adding the augmented embeddings and the original embeddings to an embeddings dataset.

Embodiments of the present disclosure relate to an apparatus, a computer program, a neural network, a training method, a method for intrusion detection, and a method for watermarking data. Embodiments of the watermarking method comprise obtaining data, determining one or more data features of the data, generating a watermark for the data using a predefined function and the data features as input for the function, and embedding the watermark in the data to obtain watermarked data.

The present disclosure provides an automotive open system architecture and a communication method for an automotive open system architecture. In a specific implementation solution, the automotive open system architecture includes an application layer and a basic software layer. A data processing module of the basic software layer is communicatively connected to a data communication module and a software component of the application layer respectively, and the data communication module is communicatively connected to a corresponding external device based on a pin of a controller. The data processing module is configured to manage correspondence information between a plurality of pins of the controller and a plurality of external devices. The data processing module is further configured to control, according to a first message containing target external device information and generated by the software component and the correspondence information, the data communication module to communicate with a target external device based on a target pin of the controller, so as to realize read and/or write operations of the software component on the target external device. According to the solution of the present disclosure, the software component does not need to know a communicative connection relationship between the pin of the controller and the external device, and can still implement read and write operations on the target external device.

Die Erfindung betrifft ein Gerät mit einer Optikeinheit (3), welche zumindest eine innenliegende Wandung (35) aufweist, die mit einer Oberflächenstruktur (351) versehen ist. Die Oberflächenstruktur (351) weist Kavitäten (352) auf. Ein solches Gerät findet vorteilhaft Verwendung in einem Head-Up-Display mit Störlichtunterdrückung.

Embodiments of the present disclosure relate to methods for detection and removal of adversarial attacks. A computer program, a computer-readable data carrier and an apparatus are also disclosed. The method for detection of adversarial attacks comprises: obtaining a sample embedding from a layer of a machine learning model; obtaining one or more predetermined adversarial perturbation embeddings corresponding to the layer of the machine learning model; obtaining an attribution score between the sample embedding and each of the predetermined adversarial perturbation embeddings; determining that at least one similarity scores of the obtained similarity scores exceed a predetermined threshold; classifying the predetermined adversarial perturbation embeddings, corresponding to the determined at least one similarity scores, as relevant perturbation embeddings; and providing an indication that an adversarial attack is detected.

Die vorliegende Erfindung betrifft eine Anzeigevorrichtung (1) mit Störlichtunterdrückung. Die Anzeigevorrichtung (1) weist ein Anzeigepanel (2) und ein auf einer Oberseite (20) des Anzeigepanels (2) angeordnetes transparentes Element (3) auf. Das Material des transparenten Elements (3) ist so gewählt, dass beim Übergang vom Anzeigepanel (2) zum transparenten Element (3) kein Brechungsindexsprung auftritt. Das transparente Element (3) weist eine gekrümmte Oberseite (30) auf, die ausgestaltet ist, einfallendes Störlicht (SL) in Richtung einer Lichtfalle (LF) abzulenken.

Embodiments of the present disclosure relate to methods for autonomous navigation. A computer program, a computer-readable data carrier and an apparatus are also disclosed. The method for autonomous navigation comprising: obtaining, from a multimodal predictor, one or more candidate trajectories for traffic participants; obtaining, using reachability analysis, respective candidate navigation corridors for each candidate trajectory; obtaining, for each candidate trajectory, a cumulated area of each of the respective candidate navigation corridors; classifying, for each candidate trajectory, a candidate navigation corridor having a greatest cumulated area as relevant corridor, thereby obtaining a class of relevant corridors; obtaining an overlap score between each pair of the relevant corridors in the class of relevant corridors; determining that at least one obtained overlap score, corresponding to at least one pair of the relevant corridors, exceeds a predetermined threshold; replacing, in the class of relevant corridors, each pair of the at least one pair of the relevant corridors with an intersection of said each pair, thereby obtaining an updated class of relevant corridors; and using, by a branch model predictive control, BMPC, the updated class of relevant corridors for generating a predicted trajectory and control commands to an autonomous system.

A temperature maintenance apparatus (1) for a vehicle (2) display, the temperature maintenance apparatus (5) comprising a first connection pipe (10) connected to a display device (5), a second connection pipe (20) having one end portion (21) connected to the first connection pipe (10), and the other end portion (23) connected to a vent hole (6) forming an air flow to the outside of the vehicle (2) when the vehicle (2) is traveling, and a third connection pipe (30) having one end portion (32) connected to at least one among the first connection pipe (10) and the second connection pipe (20), and the other end portion (33) connected to an air conditioning device (7).

The present disclosure provides a method, a system, an apparatus, an electronic device, and a storage medium for displaying in-vehicle information. A specific implementation solution includes: determining, according to a state of a vehicle, first information that needs to be displayed on a mixed reality device worn by a user in the vehicle; determining, according to the user's head posture and gaze, a target viewing zone that the user is gazing at in a field of view of the corresponding mixed reality device; determining, according to the target viewing zone, second information that needs to be displayed on the mixed reality device; determining a display method of the first information according to the target viewing zone; and realizing, according to the first information, the display method of the first information, and the second information, display of in-vehicle information of the vehicle on the mixed reality device. According to the solution of the present disclosure, it can be ensured that the in-vehicle information is always displayed in front of the user's eyes, thereby avoiding driving hazards caused by defocusing of the user's gaze, and improving driving concentration and reaction time of the user.

A computer-implemented method of object-augmented action recognition and an object-augmented action recognition system may be provided, the method comprises processing one or more initial patch tokens to extract one or more human tokens and one or more object tokens; applying the extracted one or more human tokens and the extracted one or more object tokens and the one or more initial patch tokens to a first cross attention mechanism to obtain a set of processed patch tokens containing object-guided information; enumerating to construct a set of Human-Object (HO) tokens, the constructed HO tokens each comprising information associated with visual relationship and information associated with spatial relationship between a human token and an object token; applying the constructed HO tokens and the processed patch tokens containing object-guided information to a second cross attention mechanism to obtain a set of resultant processed HO tokens; applying a feature aggregation process to the resultant processed HO tokens to synthesize the features of M groups of the resultant processed HO tokens into a single feature representation; obtaining a prediction result based on the set of processed patch tokens containing object-guided information; obtaining an action recognition prediction result based on a result of the feature aggregation process and the prediction result based on the set of processed patch tokens containing object-guided information; and outputting for use the action recognition prediction result.

Embodiments of the present disclosure relate to an apparatus, a data carrier, a computer program, an object detector, and a method for pruning a neural network for an object detector. The method comprises identifying one or more critical neurons relevant for a first sub-category of an object class. Further, the method comprises adjusting the critical neurons in order to adapt such that a deviation between representations of the first sub-category and a second sub-category of the object class decreases in the latent space. Also, the method includes removing one or more of the critical neurons which are relevant for the first sub-category and less relevant for the second sub-category.

Embodiments of the present disclosure relate to a method for correcting ground truth labels used in training a machine learning model for a computer vision task. A training method and ground truth labels used for training a machine learning model are also disclosed. A machine learning model, a computer program, a computer-readable data carrier and an apparatus are also presented. The method for correcting ground truth labels comprises obtaining sample images and ground truth labels for the sample images, training the machine learning model using the sample images and the ground truth labels, modifying the sample images, obtaining, by the machine learning model, predicted labels for the modified sample images, and correcting the ground truth labels based on a comparison of the ground truth labels with the predicted labels.

Embodiments of the present disclosure relate to a method for generating training data to be used in training a machine learning model for a computer vision task. A training method and data used for training a machine learning model are also disclosed. A machine learning model, a computer program, a computer-readable data carrier and an apparatus are also presented. The method for generating training data to be used in training a machine learning model comprises obtaining information on a frequency of occurrence of object classes in the training dataset, cutting out, based on the frequency of occurrence, objects from sample images obtained from the training dataset, and generating training data comprising a collage of at least two of the cut-out objects.

Embodiments of the present disclosure relate to an apparatus, an embedded software application, a data carrier, a computer program, and a method for fuzz testing embedded software applications. The method comprises identifying, from a plurality of keywords, network user input (NUI) keywords related to network user inputs. Also, the method comprises providing the identified keywords as seed for fuzz testing an embedded software application and testing the software based on the seed.

A power distribution system 100 comprises a power source 102, multiple consumer units 124,126,128,136 electrically connected to the power source 102, a power failure detection unit 108 electrically connected to the power source 102, the power failure detection unit 108 configured to detect a power failure of the power source 102, a communication means 130 configured to provide methods and processes for exchanging information between the power failure detection unit 108 and the multiple consumer units 124,126,128,136, and at least one capacitive energy source 114 electrically connected to at least one consumer unit 124,126,128,136 of the multiple consumer units 124,126,128,136. Upon detection of the power failure of the power source 102, the at least one capacitive energy source 114 providing power to the at least one consumer unit 124,126,128,136, the power failure detection unit 108 sending an indicator message to the at least one consumer unit 124,126,128,136 via the communication means 130, and the at least one consumer unit 124,126,128,136, upon receiving the indicator message, performing a preventive power-out measure.

Embodiments of the present disclosure relate to a vehicle, an apparatus, a data carrier, a computer program, a neural network, and methods for a human-machine-interface. One method comprises obtaining an appropriate attentiveness of an occupant/passenger of the vehicle for a future traffic situation. Further, the method comprises obtaining information on a predicted future attentiveness of an occupant/passenger of the vehicle for a future traffic situation and determining a cue to be provided to the occupant/passenger to obtain the appropriate attentiveness.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

This disclosure describes an autostereoscopic system comprising: an optical arrangement; and a controller operable to calculate and actuate the optical arrangement relative to a projected gaze position, wherein the projected gaze position is determined based upon a velocity of body movement calculated between at least one coordinate extracted between two frames of video data received by the controller, and at least one category of time factor in relation to a performance of the system. A method for actuating an optical arrangement of an autostereoscopic display screen system as disclosed herein is also provided.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

A compensator arrangement for a vehicle is provided comprising a smart phone terminal configured to be connected to a mobile phone, a compensator configured to improve a link quality and coverage for the mobile phone, a cable having a first wire that carries a high frequency signal. The smart phone terminal is configured to generate a control signal and to transmit the control signal on the first wire to the compensator.

Zusammenfassung wird geladen …