Artificial intelligence in gaming.
7 likes5,397 views
The document discusses various techniques used for artificial intelligence in gaming. It describes how state machines and planning systems are used to simulate human behavior for non-player characters. State machines define a character's states and transitions between states, but have limitations. Planning systems allow characters to work backwards from objectives to determine paths and behaviors. Additional techniques include navigation meshes to guide character movement and online learning from player data. The goal is to improve gaming experiences by making characters seem intelligent through these simulated human behavior methods.
Artificial intelligence in gaming.
- 1. ARTIFICIAL INTELLIGENCE IN GAMINGPresented by: Vivek M BhasiUnder the guidance of Mr PramodPavithran
- 2. School of Engineering, CUSAT 2
- 3. ITS HARD WORK!!! •8192 processors•2.8 TB of memory in the blue gene l supercomputer •22 million neurons •11 billion synapses School of Engineering, CUSAT 3
- 4. SIMULATION OF HUMAN BEHAVIOUR… IMPOSSIBLE??? •Simulation of actual human behaviour would be extremely difficult if not impossible•Quite often the games give a perception of reality•ie: Games appear to have characters who are much more “intelligent” than they actually are… School of Engineering, CUSAT 4
- 5. SO HOW DO THEY DO IT THEN??? School of Engineering, CUSAT 5
- 6. STATE MACHINES•Sometimes non-player characters make decisions using state machines•Each non-player character would have a certain number of states which it will go to when certain criteria are met. •An example can be cited from old FPSs •As the intelligence of the characters increases the decision trees associated with their actions become more complex•The core behaviour of the NPCs are defined only by a few states. The remaining states are for rare circumstatnces. School of Engineering, CUSAT 6
- 7. DISADVANTAGES OF STATE MACHINES•State machines are inherently predictable to a certain degree•There is a limit up to which the number of states can be increased•Defining the transitions between the states is a difficult task School of Engineering, CUSAT 7
- 8. THE PLANNING SYSTEM•Allows various characters on-screen to act uniquely •Aplanning system makes a gaming character work out what it has to do in order to fulfil an objective. •In order to achieve a goal, the computer looks at the situation in which it wants to be, and then work backwards from it to calculate the best way of achieving the desired result. School of Engineering, CUSAT 8
- 9. WORKING BACKWARDS…..AN EXAMPLE•In an FPS the objective of the AI bots is to make sure the target is dead. •But since the objective hasn’t been accomplished yet, the preconditions are analysedin a backtracking fashion•This process continues till the basic precondition is reached . •Then the AI starts satisfying each condition starting from the base case till the final objective is met. School of Engineering, CUSAT 9
- 10. ALL TERRAIN•Now that we have characters that seem intelligent , how do we make them know where they are? •The AI characters can’t see the game world as it would take a huge amount of processing power•The characters need some sort of guide to give them an idea of where they can go School of Engineering, CUSAT 10
- 11. •This often takes the form of a navigation grid that shows obstacles and the shape of the terrain, but developers often also use a navigation mesh (usually called a navmesh) that’s local to the current 3D scene. •A navmeshis basically a low-resolution 3D mesh that sits on top of a 3D scene and shows an AI character the areas to which it can move, as well as featuring tags and hints to the AI that have been placed by the designer •“The designer will place this invisible mesh any place he wants the AI character to potentially move to". School of Engineering, CUSAT 11
- 12. NAVMESHIN ACTION•Animators create a variety of animations that the game designers can use according to the terrain•These animations are used by the designers to make the characters traverse certain areas. •Navmeshesprovide a method for NPCs to find paths and also offers extra spatial information School of Engineering, CUSAT 12
- 13. NAVIGATING THROUGH VOLUME•Navigating through the entirety of the space available to the NPCs makes the gaming experience more realistic•To address this issue something called“segments” are used. •Segments are a series of points and links, which may or may not be on or attached to the navmesh. •The AI is free to create paths using these segments just as it would on a normal navmesh School of Engineering, CUSAT 13
- 14. •The planning system is used to combine different segments, thus creating paths in space where the characters can move in. •The probable destinations of the moving characters are calculated in order to avoid collisions•Highly complicated programming involved School of Engineering, CUSAT 14
- 15. DESTRUCTIBLE ENVIRONMENTS•Adds to the realistic experience of a gamer•Gamers can cause damage to the environment and thereby affect the gameplay School of Engineering, CUSAT 15
- 16. HOW IT WORKS•The NPCs keep checking their environments every 0.5 to 1 sec and respond to the changes using the planning system•Changes can be identified by using Navmeshes•ie: Whenever rigid bodies are removed from the environment , their corresponding outlines are identified in the navmeshand the characters are made to respond accordingly. School of Engineering, CUSAT 16
- 17. AI IN GAMES-WHERE NEXT? •The general mindsetof gamers are changing… •AI acceleration is being welcomed more than ever•Advancements ?......After the state machines and the planning system , the next step is Online learning ie: Online chatbotssuch as Jabberwackylearn as a result of constant online input.(strategies , playing styles etc) School of Engineering, CUSAT 17
- 18. CONCLUSION•Simulating actual human behaviour is very difficult and so games use various forms of trickery to give an impression of intelligent behaviour•The ultimate aim of AI is to improve the gaming experience and give a realistic feel to the game•Various methods including state machines, planning systems ,online learning etcare used to run and improve the AI. School of Engineering, CUSAT 18
- 19. THANK YOU
- 20. QUESTIONS??? School of Engineering, CUSAT 20