Robert Boschgmbh

Peter Lakatos

Gabor Kovacs

Peter Pozsegovics

Laszlo Szoke

A method for training a style encoder of a neural network. Sensory input variables, which represent a movement of a system and surroundings of the system, are compressed to an abstract driving situation representation in at least one portion of a latent space of the neural network, using a trained situation encoder of the neural network. The sensory input variables are compressed to a driving style representation in at least one portion of the latent space, using the untrained style encoder. The driving style representation and the driving situation representation are decompressed from the latent space to output variables, using a style decoder of the neural network. A structure of the style encoder is changed to train the style encoder until the output variables of the style decoder represent the movement.

Method and device for training a style encoder of a neural network and method for generating a driving style representation representing a driving style of a driver

A method for detecting a curbstone in an environment of a vehicle. The method includes recognizing at least one line segment that belongs to the curbstone with the aid of image data that are read in by an interface to a camera device of the vehicle. The line segment is projected onto a ground plane of the environment in order to generate a projected line segment. A subset of a plurality of three-dimensionally triangulated points in the environment of the vehicle is assigned to the line segment as a function of a position of the points relative to a position of the camera device, relative to a starting point of the projected line segment, and relative to an end point of the projected line segment. A flank plane of the curbstone is ascertained with the aid of the assigned points.

Growing community of inventors

Budapest, Hungary

Peter Lakatos