Abstract: A system for training a classification model to be robust against perturbations of multiple perturbation types. A perturbation type defines a set of allowed perturbations. The classification model is trained by, in an outer iteration, selecting a set of training instances of a training dataset; selecting, among perturbations allowed by the multiple perturbation types, one or more perturbations for perturbing the selected training instances to maximize a loss function; and updating the set of parameters of the classification model to decrease the loss for the perturbed instances. A perturbation is determined by, in an inner iteration, determining updated perturbations allowed by respective perturbation types of the multiple perturbation types and selecting an updated perturbation that most increases the loss of the classification model.

Abstract: The invention relates to a remote switch (1) comprising a first wireless interface (2) for sending out switch commands. The remote switch (1) has a second wireless interface (3) arranged separate from the first wireless interface (2). The second wireless interface (3) is configured for the wireless transmission of information (19) between the remote switch (1) and an external device (12). The invention also relates to a method for communication between a remote switch (1) and an external device (12). The method involves creating a communication link between a wireless interface (3) of the remote switch (1) and a wireless interface (14) of the external device (12), as well as transmitting information (19) between the remote switch (1) and the external device (12) by means of the established communication link. This allows for simple configuration or maintenance of the remote switch (1) by means of the external device (12) according to the exchanged information (19).

Abstract: A remote sensor including a first wireless interface for sending out data. The remote sensor has a second wireless interface arranged separate from the first wireless interface. The second wireless interface is configured for the wireless transmission of information between the remote sensor and an external device. The invention also relates to a method for communication between a remote sensor and an external device. The method involves creating a communication link between a wireless interface of the remote sensor and a wireless interface of the external device, as well as transmitting information between the remote sensor and the external device by the established communication link. This allows for simple configuration or maintenance of the remote sensor by the external device according to the exchanged information.

Abstract: A light inspection system positions an autonomous vehicle (AV) in a field of view of a camera such that the camera captures an image of a light of the AV. The light inspection system instructs a camera to capture an image of the light while the light is switched on. The light inspection system receives the captured image, determines a luminance of the light based on the image, and determines to service the light in response to the luminance of the light being below a threshold luminance.

Abstract: An artificial neural network (ANN), including processing layers which are each configured to process input quantities in accordance with trainable parameters of the ANN to form output quantities. At least one normalizer is inserted into at least one processing layer and/or between at least two processing layers. The normalizer includes a transformation element configured to transform input quantities directed into the normalizer into one or more input vectors, using a predefined transformation. The normalizer also includes a normalizing element configured to normalize the input vector(s) using a normalization function, to form one or more output vectors. The normalization function has at least two different regimes and changes between the regimes as a function of a norm of the input vector at a point and/or in a range, whose position is a function of a predefined parameter. The normalizer also includes an inverse transformation element.

Abstract: A method for training an artificial neural network (ANN), that includes a multiplicity of processing units. Parameters that characterize the behavior of the ANN are optimized with the goal that the ANN maps learning input variable values as well as possible onto associated learning output variable values as determined by a cost function. The output of at least one processing unit is multiplied by a random value x and subsequently supplied as input to at least one further processing unit. The random value x is drawn from a random variable with a probability density function containing an exponential function in |x?q| that decreases as |x?q| increases, where q is a freely selectable position parameter and |x?q| is contained in the argument of the exponential function in powers |x?q|k where k?1. A method for operating an ANN is also described.

Abstract: Described herein is a server computing system that controls an autonomous vehicle to perform an operation to at least one of measure or isolate an effect of a variable on an actual power consumption by the autonomous vehicle. Data indicative of the actual power consumption, which is generated based on the operation, and data indicative of a projected power consumption, which is accumulated based on prior execution of the operation by a same or different autonomous vehicle, is received by the server computing system to determine whether an energy efficiency of the autonomous vehicle is degraded. The operation may be performed to identify a degraded vehicle system or component of the autonomous vehicle or to identify an autonomous vehicle in a fleet of autonomous vehicles for which further analysis is desirable. An output is generated by the server computing system that is indicative of the energy efficiency prognostics.

Abstract: In a method for training a first neural network a superposed classification is back-propagated through a second neural network. An output value of the second neural network is utilized to determine whether the input of the first neural network is adversarial.

Abstract: A clamp for fastening an object in the form of a tube or hose, having an elongate clamp strap at each of the ends of which there is provided a radially outwardly projecting fastening leg, and having a rubber profile which is provided on the radial inner face of the clamp strap and which at least partially engages around each of the edges of the clamp strap by way of a lip. At least one first holding geometry is provided on the radial inner face of the clamp strap, and at least one second holding geometry is provided on the radial outer face of the rubber profile.

Abstract: The present disclosure provides a method comprising receiving a rideshare request from a user, the rideshare request including a destination; transporting the passenger to the destination using an autonomous vehicle (AV) having a plurality of windows comprising electrically switchable smart glass; during the transporting, monitoring a metric related to arrival of the AV at the destination; and untinting the windows when the monitored metric has a prescribed relationship to a threshold value.

Abstract: Described herein is a server computing system that controls an autonomous vehicle to perform an operation to at least one of measure or isolate an effect of a variable on an actual power consumption by the autonomous vehicle. Data indicative of the actual power consumption, which is generated based on the operation, and data indicative of a projected power consumption, which is accumulated based on prior execution of the operation by a same or different autonomous vehicle, is received by the server computing system to determine whether an energy efficiency of the autonomous vehicle is degraded. The operation may be performed to identify a degraded vehicle system or component of the autonomous vehicle or to identify an autonomous vehicle in a fleet of autonomous vehicles for which further analysis is desirable. An output is generated by the server computing system that is indicative of the energy efficiency prognostics.

Abstract: An apparatus for detecting the activity of persons or the state of infrastructures or objects influenced by persons by means of acceleration measurement. The device has an acceleration sensor which is arranged to react to a preset threshold value of a measured acceleration and, when the threshold value is exceeded, to trigger at least one of the actions of data storage, modification of a counter or transmission of a data telegram by radio. The apparatus further comprises an energy converter for converting a primary energy into electrical energy for operating the apparatus or the acceleration sensor. The energy converter is arranged to obtain the primary energy independently of an energy resulting from a measured acceleration.

Abstract: A clamp for fastening an object in the form of a tube or hose, having an elongate clamp strap at each of the ends of which there is provided a radially outwardly projecting fastening leg, and having a rubber profile which is provided on the radial inner face of the clamp strap and which at least partially engages around each of the edges of the clamp strap by way of a lip. At least one first holding geometry is provided on the radial inner face of the clamp strap, and at least one second holding geometry is provided on the radial outer face of the rubber profile.

Abstract: A method for influencing driving dynamics of a vehicle, in which the driving dynamics are influenced as a function of parameters allocated to a selected driving dynamics mode when the driving dynamics mode is activated, and the driving dynamics are influenced as a function of road state information representing a road state in the region of the vehicle when an automatic mode is activated.

Abstract: A method for providing an activation signal for activating an actuator. The activation signal is ascertained as a function of an output signal of a neural network. The neural network includes a scaling layer. The scaling layer maps an input signal present at the input of the scaling layer onto an output signal present at the output of the scaling layer in such a way that this mapping corresponds to a projection of the input signal onto a predefinable value range, parameters being predefinable, which characterize the mapping.

Abstract: Described herein is a system and method for detecting malodor within a cabin of an autonomous vehicle, wherein initiation of a remediation system and/or dispatch of the autonomous vehicle to a service hub for mitigating the malodor is based upon sensor signals that monitor the cabin of the autonomous vehicle. The remediation system can include devices such as an HVAC system or a window, which are operated to reduce the concentration of airborne molecules in the cabin of the autonomous vehicle when the concentration of airborne molecules exceeds an air quality threshold. A dispatch protocol may be initiated to dispatch the autonomous vehicle to a service hub when the malodor is associated with certain predefined incidents or when remediation fails to reduce the concentration of airborne molecules below the air quality threshold.

Abstract: A computer-implemented method for training a neural network, which, in particular, is configured to classify physical measuring variables, a fitting of parameters of the neural network occurring as a function of an output signal of the neural network, when the input signal is supplied, and as a function of an associated desired output signal, the fitting of the parameters occurs as a function of an ascertained gradient. The components of the ascertained gradient are scaled as a function of to which layer of the neural network the parameters corresponding to these components belong.

Abstract: The present disclosure provides a method comprising, subsequent to arrival of a vehicle at a charging station, establishing communications between a vehicle side autonomous charging system associated with the vehicle and a charging station side autonomous charging system associated with the charging station; receiving by the charging station side autonomous charging system at least one signal from the vehicle side autonomous charging system, wherein the at least one signal is indicative of a location of a charging port of the vehicle; and using the received at least one signal to guide a robotic extension arm of the charging station to a location proximate the vehicle charging port.

Abstract: A computer-implemented method for training a neural network, which, in particular, is configured to classify physical measuring variables. The neural network is trained with the aid of a training data set. Pairs including an input signal and an associated desired output signal are drawn from the training data set for training. An adaptation of parameters of the neural network occurs as a function of an output signal of the neural network, when the input signal is supplied, and as a function of the desired output signal. The drawing of pairs always takes place from the entire training data set.

Abstract: Predictive maintenance and diagnostics for an electronic module of an autonomous vehicle using modular condition monitoring is described herein. A computing system receives a signal from a data logger which monitors a condition of the electronic module of the autonomous vehicle, wherein the signal is indicative of damage accumulation information thereof. The computing system identifies a type of the electronic module and a damage accumulation threshold for the type of the electronic module to generate a predicted maintenance schedule for the electronic module of the autonomous vehicle. The damage accumulation information can be stored in a data store to define the damage accumulation threshold for the type of the electronic module.