Harmonic opponent modeling and behavior structure for 3D soccer simulation agent

As the beginning of the 3D soccer simulation competitions the simulated humanoid robots were used to play soccer in simulated games, this event shifted the aim of the soccer simulation from the design of strategic behaviors into the low level skills like walking and kicking in soccer player itself at first, but by the improvement of these skills new challenge of how to play soccer has been engendered, so having a framework for agents that supports a hierarchy of behaviors is inevitable, in this paper we proposed a behavior structure for agents which consists of low level skills such as kicking and walking and high level skills like ball-handled and go to ball which uses a combination of low level skills, and also a simple method of opponent modeling called harmonic opponent modeling which model the walking speed of opponent agent.