Drone Grid Case Study: Creating the Enemies - Beginning Android 3D Game Development - page 270

Using a finite state machine, each of these behaviors would be implemented in a separate state.
The initial state would process the player's current command and transition to the corresponding
state implementing the command. Once each of these commands was completed, the finite state
machine would transition back to the state that processes the player's commands. The process
would be repeated, and the state machine would transition to the state implementing the player's
current command. (See Figure 8-2 .)

Retreat Finished = false

Retreat
State

Command = Retreat

Retreat Finished = true

Patrol Finished = false

Process
Command
State

Command = Patrol

Patrol
State

Default State

Patrol Finished = true

Attack Enemy Finished = true

Command = Attack Enemy

Attack
Enemy
State

Attack Enemy Finished = false

Figure 8-2. The finite state machine

In terms of implementing a finite state machine in code, we have a Finite State Machine class for
the enemy type we want to control. Each of the behaviors for this type of enemy is represented by
a separate class and is loaded into the finite state machine for processing. The UpdateMachine()
function in the Finite State Machine class is called to update the state machine. See Figure 8-3 for
a diagram illustrating a general class level view of how our state machine in Figure 8-2 would be
implemented in code.