VIRTUAL TRAINING METHOD FOR A NEURAL NETWORK FOR ACTUATING A TECHNICAL DEVICE

1 : A computer-implemented method for training a neural network for actuating a technical device, comprising:

establishing a first data link between the neural network and a first virtual simulation of the technical device via a first data interface of the first simulation for reading out status data from the first simulation and transferring control data to the first simulation;

setting a first training goal for an actuation of the first simulation;

training the neural network based on the first simulation being actuated by the neural network, and checking the training progress of the neural network against the first training goal;

breaking the first data link based on the first training goal having been achieved, and then establishing a second data link between the neural network and a second virtual simulation of the technical device, wherein the second simulation is configured to be more realistic than the first simulation and requires more mathematical operations than the first simulation for a respective simulation cycle owing to its higher degree of realism, by and wherein the second data link is established via a second data interface of the second simulation for reading out status data from the second simulation and transferring control data to the second simulation; and

training the neural network based on the second simulation being actuated by the neural network.

2 : The method according to claim 1 , wherein the first data interface and the second data interface are configured identically.

3 : The method according to claim 1 , further comprising:

generating simulated sensor data which contain information on a virtual environment of the technical device;

transferring the simulated sensor data to the neural network; and

training the neural network to evaluate the simulated sensor data and to take account of the simulated sensor data when actuating the first simulation and/or the second simulation.

4 : The method according to claim 1 , wherein the technical device is a robot belonging to at least one of the following categories:

a vehicle;

a robot arm;

a robot for positioning or attaching material and/or objects;

a robot for cleaning surfaces;

a robot for examining spaces or surfaces;

a robot for applying a chemical; or

a robot for carrying out a medical intervention.

5 : The method according to claim 4 , wherein the first training goal belongs to at least one of the following categories of training goals:

a stretch traveled on a virtual training course or a virtual test route;

a number of movement patterns carried out without any collisions or within a predefined movement range;

a time period within which the neural network actuates the first simulation properly and without any undesired events occurring; or

a threshold value being reached for a reward function.

6 : The method according to claim 1 , wherein the second simulation:

takes account of more mechanical and/or electrical components of the technical device than the first simulation;

takes account of more degrees of mechanical freedom than the first simulation;

simulates physical phenomena and/or laws in a more realistic manner and/or to a higher degree of accuracy than the first simulation;

takes account of a larger number of physical forces and/or interactions than the first simulation; and/or

has a smaller simulation step size than the first simulation.

7 : The method according to claim 1 , further comprising:

setting a second training goal for an actuation of the second simulation;

during the actuation of the second simulation, checking the training progress of the neural network against the second training goal;

establishing a physical data link between the neural network and the technical device based on the second training goal having been achieved; and

the neural network actuating the technical device.

8 : A system, comprising:

at least one processor; and

at least one memory having instructions stored thereon;

wherein the at least one processor is configured to execute the instructions stored on the at least one memory to provide a virtual training environment for training a neural network for actuating a technical device, the virtual training environment comprising:

a first simulation of the technical device having a first data interface for reading out status data from the first simulation and transferring control data to the first simulation;

a programming interface for establishing a first data link between the neural network and the first simulation;

a training functionality configured to set a training goal for the first simulation to be actuated by the neural network and to check the training progress of the neural network against the training goal; and

a second simulation of the technical device, wherein the second simulation is configured to be more realistic than the first simulation and requires more mathematical operations than the first simulation for a respective simulation cycle owing to its higher degree of realism, and wherein the second simulation comprises a second data interface for reading out status data from the second simulation and transferring control data to the second simulation;

wherein the training functionality is configured to:

break the first data link based on the training goal having been achieved;

replace the first data link with a second data link between the neural network and the second simulation; and

train the neural network based on the second simulation being actuated by the neural network.

9 : The system according to claim 8 , wherein the first data interface and the second data interface are configured identically.

10 : The system according to claim 8 , wherein the virtual training environment comprises a virtual environment of the technical device and a sensor simulation; and

wherein the virtual training environment is configured to:

generate simulated sensor data via the sensor simulation, wherein the simulated sensor data contain information on the virtual environment; and

transmit the simulated sensor data to the neural network via the programming interface.

11 : The system according to claim 8 , wherein the technical device is a robot belonging to at least one of the following categories:

a vehicle;

a robot arm;

a robot for positioning or attaching material and/or objects;

a robot for cleaning surfaces;

a robot for examining spaces or surfaces;

a robot for applying a chemical; or

a robot for carrying out a medical intervention.

12 : The system according to claim 11 , wherein the training goal belongs to at least one of the following categories of training goals:

a stretch traveled on a virtual training course or a virtual test route;

a number of movement patterns carried out without any collisions or within a predefined movement range;

a time period within which the neural network actuates the first simulation properly and without any undesired events occurring; or

a threshold value being reached for a reward function.

13 : The system according to claim 8 , wherein the second simulation:

takes account of more mechanical and/or electrical components of the technical device than the first simulation;

takes account of more degrees of mechanical freedom than the first simulation;

simulates physical phenomena and/or laws in a more realistic manner and/or to a higher degree of accuracy than the first simulation;

takes account of a larger number of physical forces and/or interactions than the first simulation; and/or

has a smaller simulation step size than the first simulation.