US 12,447,976 B2
Method for producing a model for automated prediction of interactions of a user with a user interface of a motor vehicle
David Bethge, Stuttgart-Feuerbach (DE); Jannik Wolf, Filderstadt (DE); Marco Wiedner, Karlsruhe (DE); and Mohamed Kari, Essen (DE)
Assigned to Dr. Ing. h. c. F. Porsche AG, Stuttgart (DE)
Filed by Dr. Ing. h.c. F. Porsche Aktiengesellschaft, Stuttgart (DE)
Filed on Oct. 25, 2022, as Appl. No. 17/972,979.
Claims priority of application No. 10 2021 129 085.6 (DE), filed on Nov. 9, 2021.
Prior Publication US 2023/0146013 A1, May 11, 2023
Int. Cl. B60W 50/00 (2006.01); B60W 40/08 (2012.01); B60W 40/105 (2012.01); B60W 50/14 (2020.01); G06N 5/01 (2023.01); H04L 12/40 (2006.01)
CPC B60W 50/0097 (2013.01) [B60W 40/08 (2013.01); B60W 40/105 (2013.01); B60W 50/14 (2013.01); G06N 5/01 (2023.01); H04L 12/40 (2013.01); B60W 2050/0029 (2013.01); B60W 2050/0083 (2013.01); B60W 2050/146 (2013.01); B60W 2556/10 (2020.02); H04L 2012/40215 (2013.01); H04L 2012/40273 (2013.01)] 11 Claims
OG exemplary drawing
 
1. A method (10) for producing a model (15) for automated prediction of interactions of a user with a user interface of a motor vehicle, the method comprising:
providing vehicle operating logs (11, 12, 13) where each vehicle operating log (11, 12, 13) includes a record of a time sequence of user interactions with the user interface, wherein providing includes filtering or preprocessing raw data read from a data network of the motor vehicle and converting the raw data into structured data sets in the form of a tabular database;
assigning context information (21, 22) to each of the user interactions recorded in the vehicle operating logs (11, 12, 13), the context information (21, 22) including a functional category (21) of the user interaction and a driving state (22) of the motor vehicle at the time of the user interaction, wherein the functional category (21) is converted by a one-hot encoding into a bit encoding that identifies the functional category;
generating training data (14) based on the vehicle operating logs (11, 12, 13) and the associated context information (21, 22)
training a context-sensitive interaction model (15) by machine learning based on the training data (14) to make a prediction about a future user interaction based on a time sequence of past user interactions, wherein training includes executing a training routine over the entire data set to produce a classifier in the form of a decision tree having a maximum tree depth of 8;
adjusting a display field on the user interface based on the prediction about the user interaction to direct attention of the user to a control element or a display element of the user interface; and
triggering the user interface to perform the predicted user interaction without or before interaction of the user with the user interface for functional categories and driving states where diversion of the driver to the user interface should be reduced.