Implementation of a Machine Learning Model Using Neural Networks: a Case Study of Vehicle Motion Modeling Task

The article explores reinforcement learning through the development of a model for automatic control of a vehicle on a track within the context of a computer game. It describes the Q-learning method and its application in the DQN algorithm for agent training. The challenge of selecting the right volume of training data is discussed. The gaming model created to assess the impact of the number of machine vision sensors and neural network structure on the success of learning is outlined. Such a project can be implemented by students for training purposes to enhance their engagement in the learning process.