Method and Driving Assistance Device for Occupancy-Grid-Based Predictive Support for an Automatic Vehicle Control, and Correspondingly Configured Motor Vehicle

A method for supporting automatic control of a motor vehicle, comprising:

• capturing sensor data from an environmental sensor system of the motor vehicle;
• capturing vehicle data indicating a current movement state of the motor vehicle;
• capturing position data indicating a current position of the motor vehicle;
• subdividing an environment of the motor vehicle into a multiplicity of cells according to an occupancy grid;
• supplying the sensor data, the vehicle data and the position data as an input to a predefined machine learning model which is configured and trained to predict, on the basis thereof, a respective future movement state of the motor vehicle in the respective cell in a cell-by-cell manner and to output corresponding predictions as an output;
• determine a deviation, at the respective current time, on a basis of the output from the model, wherein the deviation is determined between at least one of:
• at least the prediction last output by the model for the cell in which the motor vehicle is currently located and the current movement state of the motor vehicle;
• or
• at least one prediction last output by the model for at least one cell and an earlier prediction by the model for the same cell or a cell covering the same spatial area;
• and
• automatically determining a measure for automatically controlling the motor vehicle at least on the basis of the determined deviation.