Method for fully automatic guidance of a vehicle system and motor vehicle

Abstract

The invention relates to a vehicle system (5) for operating a motor vehicle (1, 2, 3) designed for fully automatic guidance of the motor vehicle (1, 2, 3) in various types of driving situations method, wherein - using a computing structure comprising a plurality of evaluation units, as driving situation data for fully automatic guidance of the motor vehicle (1, 2, 3) from the description of the motor vehicle (1, 2, 3) The control data to be used are obtained from the environmental data of the environment and the own data describing the state of the motor vehicle (1, 2, 3) and are used to guide the motor vehicle (1, 2, 3), - each evaluation unit obtains from at least one other The output data is obtained from the output data and / or driving situation data of an evaluation unit, and -at least a part of the evaluation unit is constructed as a neural network based at least in part on software, wherein, -during runtime, by At least a part of the evaluation unit of the neural network is dynamically formed by configuration objects that can be configured through configuration parameter sets, and the current driving situation type is obtained from a plurality of predetermined driving situation types under the use of at least a part of the driving situation data. , wherein at least one evaluation function is assigned to each driving situation type, - an evaluation function that calls the configuration parameter set associated with the current driving situation type from the database (7), and - an evaluation function that does not yet exist, performs the evaluation function A unit is formed by configuring a configuration object with a set of configuration parameters that have been called.

Description

Technical field[0001]The present invention relates to a method for operating a vehicle system of a motor vehicle configured to guide the motor vehicle fully automatically in various types of driving situations, wherein:[0002]-Using a calculation structure including a plurality of evaluation units to obtain control data to be used for fully automatic guidance of the motor vehicle from environmental data describing the environment of the motor vehicle and self data describing the state of the motor vehicle as driving condition data, And used to guide motor vehicles,[0003]-Each evaluation unit obtains output data from the output data and / or driving condition data of at least another evaluation unit,[0004]-At least a part of the evaluation unit is constructed as a neural network at least partly based on software. Furthermore, the invention relates to a motor vehicle.Background technique[0005]The topic of current research and development is to guide the motor vehicle fully automatically ...

AI Technical Summary

Problems solved by technology

[0009] However, with regard to vehicle systems designed for fully automatic guidance of the motor vehicle, in particular in various types of driving situations, the problem arises that neural networks of this type can only be trained for certain traffic situations, for example for Intelligent and accident-free passing through intersections, recognition of the intentions of other traffic participants, automatic parking of certain aspects of the process, situational analysis of overtaking on highways, etc.

Although within these special evaluation functions it is possible to implement by means of artificial neural networks, however due to the complexity of reality it is not possible to implement neural networks or groups of neural networks which can be used without restrictions for all conceivable traffic situations

However, due to the computing power available in modern motor vehicles, it is also not feasible to implement a very large number of individual evaluation functions by means of neural algorithms

Images