Robert Boschgmbh

Kilian Rambach

Gor Hakobyan

Jan Hendrik Metzen

Robin Hutmacher

Michael Schoor

Benedikt Loesch

Goetz Kuehnle

Markus Karl Vogel

Chaithanya Kumar Mummadi

Michael Pfeiffer

Jasmin Ebert

Kanil Patel

Evgeny Levinkov

Lisa-Kristina Morgan

Adriana-Eliza Cozma

A method for reconstructing elevation information from measured data that were recorded with the aid of at least one radar device and include a two-dimensional spatial distribution of at least one physical measured variable. The measured data are fed as input variables to at least one generator module that is designed as a neural network. At least one output variable is retrieved from the generator module that represents a measure of the elevation angles from which radar radiation was reflected to the radar device from at least one object. A method for training a generator module, and a method including a complete active chain up to activating a vehicle, are also described.

Reconstruction of elevation information from radar data

A computer-implemented method for adapting a pretrained machine learning system, which has been trained on a first training data set, to a second dataset, wherein the second dataset has different characteristics than the first data set. An input transformation module for partly undoing the distribution shift between the first and second training data set is provided.

Device and method to adapt a pretrained machine learning system to target data that has different distribution than the training data without the necessity of human annotations on target data

A method for classifying objects based on measured data recorded by at least one radar sensor. In the method, a frequency spectrum of time-dependent measured data of the radar sensor is provided; from this frequency spectrum, locations from which reflected radar radiation has reached the radar sensor are ascertained; at least one group of such locations belonging to one and the same object is ascertained; for each location in this group, a portion of the frequency spectrum that corresponds to the radar radiation reflected from this location is ascertained; all these portions for the object are aggregated and are fed to a classifier; the object is assigned by the classifier to one or multiple classes of a predefined classification.

Hybrid evaluation of radar data for classifying objects

A method for training a neural network. The method includes providing learning input images and associated learning output data; providing auxiliary input images; generating modifications of these auxiliary input images by introducing at least one predefined change into them; supplying the modifications to the neural network; ascertaining predictions for the predefined change, using output data onto which the neural network maps the modifications; assessing deviations of the predictions from the predefined change, using an auxiliary cost function; optimizing parameters of the neural network to improve the assessment by the auxiliary cost function during further processing of auxiliary input images; supplying the learning input images to the neural network; assessing deviations of the output data, thus obtained, from the learning output data, using a main cost function; optimizing parameters of the neural network to improve the assessment by the main cost function during further processing of learning input images.

Data-based updating of the training of classifier networks

A method for training a generator. The generator is supplied with at least one actual signal that includes real or simulated physical measured data from at least one observation of the first area. The actual signal is translated by the generator into a transformed signal that represents the associated synthetic measured data in a second area. Using a cost function, an assessment is made concerning to what extent the transformed signal is consistent with one or multiple setpoint signals, at least one setpoint signal being formed from real or simulated measured data of the second physical observation modality for the situation represented by the actual signal. Trainable parameters that characterize the behavior of the generator are optimized with the objective of obtaining transformed signals that are better assessed by the cost function. A method for operating the generator, and that encompasses the complete process chain are also provided.

Translation of training data between observation modalities

A method for training a trainable module for evaluating radar signals. The method includes feeding actual radar signals and/or actual representations derived therefrom of a scene observed using the actual radar signals to the trainable module and conversion thereof by this trainable module to processed radar signals and/or to processed representations of the respective scene, and using a cost function to assess to what extent the processed radar signals are suited for reconstructing a movement of objects or to what extent the processed representations contain artifacts of moving objects in the scene. Parameters, which characterize the performance characteristics of a trainable module, are optimized with regard to the cost function. A method is also provided for evaluating moving objects from radar signals.

Processing of radar signals including suppression of motion artifacts

A method is described for locating and/or classifying at least one object, a radar sensor that is used including at least one transmitter and at least one receiver for radar waves. The method includes: the signal recorded by the receiver is converted into a two- or multidimensional frequency representation; at least a portion of the two- or multidimensional frequency representation is supplied as an input to an artificial neural network, ANN that includes a sequence of layers with neurons, at least one layer of the ANN being additionally supplied with a piece of dimensioning information which characterizes the size and/or absolute position of objects detected in the portion of the two- or multidimensional frequency representation; the locating and/or the classification of the object is taken from the ANN as an output.

Locating and/or classifying objects based on radar data, with improved reliability at different distances

For providing time multiplexing sequences for transmit antennas of a linear or two-dimensional MIMO radar unit that includes antennas situated close to one another (collocated MIMO radar), a method that enables a precise angle estimation includes implementing an algorithm with which a transmit sequence of transmitters and their transmit times are determined so that object movements have essentially no influence on the angle estimation. In this way, as a function of previously known quantities, optimal time multiplexing sequences can be determined in each case.

Method for determining a time multiplexing sequence for a MIMO radar

In a method for operating a MIMO radar, an influence of an object motion on an angle estimate is substantially eliminated, and a time multiplex schema having a transmission sequence and transmission instants of transmitters of the MIMO radar is identified by optimizing the following mathematical relationship: in which: dis optimized positions of the transmitters in the sequence in which they transmit; dis positions of the transmitters in the sequence in which they transmit; t is transmission instants; Varis sample variance; and Covis sample covariance.

Growing community of inventors

Stuttgart, Germany

Kilian Rambach