Talk:Robogame Strategy

Google Docs version

Click here to access the google docs editable version.

• Current version: 5.

Reference definitions

State variable: a variable that is needed to identify a state of the game (player, robot, environment); e.g., the distance between robot and player, the relative direction, the time passed when a time triggering event occurs, etc.

Constraint: boundaries on variables that limit the state space.

Game actions: the legal actions in the game, possibly discretised.

Game rule: a description of what is possible to do and the possible consequences; e.g., actions that lead to states where points can be given, or constraints (or effects of actions) change.

Strategy: a criterion to select an action over another one.

Strategy support variable: a variable, computed from available data, used to provide the strategic module with information useful to select a strategy (e.g., timing of player's activity, usual kind of actions, etc.)

Game Description

The game consists of a three-round based interaction between the human player and a robot within an arena — possibly with some predefined obstacles, where the robot tries to reach a specific target and the human player attempts to prevent that by pressing a touch sensor attached to the robot.

There exits 3 specific regions defined by how far the robot detects the player in relation to itself, namely: A, B and C, where

• Region A: denotes the critical region such that the human player can perform an attempt to push the robots sensor.
• Region B: Defines the medium distance meaning that the player is search for an opportunity to perform an strike.
• Region C: Is the furthermost one and the human prevalence in it has some side effects.

Game Rules

• Rule 0: at the beginning of a round the players are apart from each other and the robot detects the human player’s position within region C.

• Rule 1: at each round the human player must try to prevent the robot from reaching its target by pressing its touch sensor.
  • Consequence: if (at any time) the human player succeeds on pushing the robot touch sensor then he wins the current round and Rule 3 is applied. On the other hand, if the robot accomplishes Rule 2 he loses the game and Rule 3 is also applied.

• Rule 2: at each round the robot must try to reach a specific spot in the arena.
  • Consequence: If the robot reach its target spot it wins the round and Rule 3 is applied. On the other hand, if he has its sensor pressed (at any time) he loses the round followed by the application of Rule 3.

• Rule 3: at the end of each round except the last one, the players restart a new round being apart from each other and the robot must be detecting the player’s position within region C (Rule 0).
  • Consequence: the beginning of a new round.

• Rule 4: at any time, if the human is detected inside region A and stays there for more than 2 seconds then the robot earns the right for a 5-secs free run.
  • Consequences: disables the touch sensor (which returns to function after the free run); The human loses the round if he makes any contact with the robot or blocks its trajectory for more than 2 seconds.

• Rule 5: After entering the region A the player has to wait for 5 seconds without touching the robot and only after that period of time he can perform a new attempt to push the robot’s touch sensor.
  • Consequence: The human loses the round if he makes any contact with the robot or blocks its trajectory for more than 2 seconds.

• Rule 6: at any time, if the human player stays in region C for more than 10 secs then the robot earns a 7 seconds free run.
  • Consequence: same as Rule 4.

• Rule 7: The human player cannot be closer than a certain range T from the robot’s target.
  • Consequence: If this rule is broken the robot earns a 3 seconds free run with its touch sensor disabled (the sensor returns to function after the free run).

Winning Conditions

The game ends after three rounds and the winner is the player who has won most rounds.

State Variables

• Robot distance to the human player;
• Player time constraint;
• Robot touch sensor status;
• Direction the human player is current facing at w.r.t. the robot;
• Robot distance to obstacles;
• Rounds left;
• Number of rounds won by the robot
• Distance to the target

Strategy support variable

• Estimated time for the player to reach the robots touch sensor given his current position;
• Frequency of which the player cross in Region A.
• Estimated player’s reaction time;

Future Extensions

• Exploit planning for obstacles;
• Exploit learning strategies (with the aim of keeping the player's interest high);
• Add other robot players to have team strategies;
• Allow more than one human player;