An AI-based tennis game by application of virtual reality components

Learning a complex task such as tennis game is a challenging problem for both AI and humans. Even after acquiring the necessary skills, a strategy is needed to win the game. This paper represents an environment for a tennis match between a human player and an AI controlled opponent. The game is conducted in an immersive and interactive virtual reality based atmosphere which is provided by the HTC Vive device. We create an integrated system using different techniques for taking best actions of the AI agent. The agent should learn how to hit the ball, how to estimate the best defensive position to stand in, and how to find the weakness of the opponent to beat him. In our simulation, the AI agent is trained with a learning automata combined with RBF networks to learn when and how to use optimized strategies.