This paper presents a method for self-learning control systems using neural networks to solve highly nonlinear control problems. The approach involves two neural networks: an emulator that learns to identify the dynamics of a system, and a controller that learns to control the system based on the emulator's output. The controller is trained to minimize the error between the actual system and the desired output, using a back-propagation algorithm. The method is demonstrated using the "truck backer-upper" example, where a neural network controller is used to steer a trailer truck while backing up to a loading dock from any initial position. The controller is able to guide the truck to the dock with high accuracy, even when the initial position is arbitrary. The technique is applicable to a wide variety of nonlinear control problems. The paper also discusses the training process of the neural networks, the use of back-propagation for training, and the use of a more sophisticated objective function that includes control energy minimization. The results show that the controller can learn to control the truck with high accuracy, even when the initial position is arbitrary. The paper concludes with a discussion of future research directions, including the determination of the complexity of the emulator and controller, the convergence and learning rate of the networks, and the robustness of the control scheme.This paper presents a method for self-learning control systems using neural networks to solve highly nonlinear control problems. The approach involves two neural networks: an emulator that learns to identify the dynamics of a system, and a controller that learns to control the system based on the emulator's output. The controller is trained to minimize the error between the actual system and the desired output, using a back-propagation algorithm. The method is demonstrated using the "truck backer-upper" example, where a neural network controller is used to steer a trailer truck while backing up to a loading dock from any initial position. The controller is able to guide the truck to the dock with high accuracy, even when the initial position is arbitrary. The technique is applicable to a wide variety of nonlinear control problems. The paper also discusses the training process of the neural networks, the use of back-propagation for training, and the use of a more sophisticated objective function that includes control energy minimization. The results show that the controller can learn to control the truck with high accuracy, even when the initial position is arbitrary. The paper concludes with a discussion of future research directions, including the determination of the complexity of the emulator and controller, the convergence and learning rate of the networks, and the robustness of the control scheme.