The present disclosure relates to a self-learning control device. The self-learning control device is configured to initialize a data set including data identifying one or more previously recorded input signals for the control device; Defining one or more optimization goals, each in the form of a reward function; Generate one or more parameterizations for each reward function; Training a neural network for each reward function based on the data set and the respective one or more parameterizations. The present disclosure further relates to a vehicle including the self-learning control device.
展开▼