Chapter 5.3 Artificial Intelligence: Agents, Architecture, and TechniquesArtificial IntelligenceGame Artificial Intelligence: What is considered Game AI?Possible Game AI DefinitionGoals of an AI Game ProgrammerSpecialization of Game AI DeveloperGame AgentsSense-Think-Act Cycle: SensingSensing: Enforcing LimitationsSensing: Human Vision Model for AgentsSensing: Vision ModelSensing: Human Hearing ModelSensing: Modeling HearingSensing: Modeling Hearing EfficientlySensing: CommunicationSensing: Reaction TimesSense-Think-Act Cycle: ThinkingThinking: Expert KnowledgeThinking: SearchThinking: Machine LearningThinking: Flip-Flopping DecisionsSense-Think-Act Cycle: ActingActing: Showing IntelligenceExtra Step in Cycle: Learning and RememberingLearningRememberingRemembering within the WorldMaking Agents StupidAgent CheatingFinite-State Machine (FSM)Finite-State Machine: In Game DevelopmentPowerPoint PresentationFinite-State Machine: UML DiagramFinite-State Machine: ApproachesFinite-State Machine: Hardcoded FSMFinite-State Machine: Problems with switch FSMFinite-State Machine: Scripted with alternative languageFinite-State Machine: Scripting AdvantagesFinite-State Machine: Scripting DisadvantagesFinite-State Machine: Hybrid ApproachFinite-State Machine: ExtensionsCommon Game AI TechniquesCommon AI Techniques: A* PathfindingCommon AI Techniques: Command HierarchyCommon AI Techniques: Dead ReckoningCommon AI Techniques: Emergent BehaviorCommon AI Techniques: FlockingCommon AI Techniques: FormationsCommon AI Techniques: Influence MappingCommon AI Techniques: Level-of-Detail AICommon AI Techniques: Manager Task AssignmentCommon AI Techniques: Obstacle AvoidanceCommon AI Techniques: ScriptingCommon AI Techniques: Scripting Pros and ConsCommon AI Techniques: State MachineCommon AI Techniques: Stack-Based State MachineCommon AI Techniques: Subsumption ArchitectureCommon AI Techniques: Terrain AnalysisCommon AI Techniques: Trigger SystemPromising AI TechniquesPromising AI Techniques: Bayesian NetworksPromising AI Techniques: Blackboard ArchitecturePromising AI Techniques: Decision Tree LearningPromising AI Techniques: Filtered RandomnessPromising AI Techniques: Fuzzy LogicPromising AI Techniques: Genetic AlgorithmsPromising AI Techniques: N-Gram Statistical PredictionPromising AI Techniques: Neural NetworksPromising AI Techniques: PerceptronsPromising AI Techniques: Perceptrons (2)Promising AI Techniques: PlanningPromising AI Techniques: Player ModelingPromising AI Techniques: Production SystemsPromising AI Techniques: Reinforcement LearningPromising AI Techniques: Reputation SystemPromising AI Techniques: Smart TerrainPromising AI Techniques: Speech RecognitionPromising AI Techniques: Text-to-SpeechPromising AI Techniques: Weakness Modification LearningChapter 5.3Artificial Intelligence:Agents, Architecture, and TechniquesCS 4455 2Artificial IntelligenceIntelligence embodied in a man-made deviceHuman level AI still unobtainableCS 4455 3Game Artificial Intelligence:What is considered Game AI?Is it any NPC behavior?–A single “if” statement?–Scripted behavior?Pathfinding?Animation selection?Automatically generated environment?Best shot at a definition of game AI?CS 4455 4Possible Game AIDefinitionInclusive view of game AI:“Game AI is anything that contributes to the perceived intelligence of an entity, regardless of what’s under the hood.”CS 4455 5Goals of anAI Game ProgrammerDifferent than academic or defense industry1. AI must be intelligent, yet purposely flawed2. AI must have no unintended weaknesses3. AI must perform within the constraints4. AI must be configurable by game designers or players5. AI must not keep the game from shippingCS 4455 6Specialization ofGame AI DeveloperNo one-size fits all solution to game AI–Results in dramatic specialization Strategy Games–Battlefield analysis–Long term planning and strategyFirst-Person Shooter Games–One-on-one tactical analysis–Intelligent movement at footstep levelReal-Time Strategy games the most demanding, with as many as three full-time AI game programmersCS 4455 7Game AgentsMay act as an–Opponent–Ally–Neutral characterContinually loops through the Sense-Think-Act cycle–Optional learning or remembering stepCS 4455 8Sense-Think-Act Cycle:SensingAgent can have access to perfect information of the game world–May be expensive/difficult to tease out useful infoGame World Information–Complete terrain layout–Location and state of every game object–Location and state of playerBut isn’t this cheating???CS 4455 9Sensing:Enforcing LimitationsHuman limitations?Limitations such as–Not knowing about unexplored areas–Not seeing through walls–Not knowing location or state of playerCan only know about things seen, heard, or told aboutMust create a sensing modelCS 4455 10Sensing:Human Vision Model for AgentsGet a list of all objects or agents; for each:1. Is it within the viewing distance of the agent?•How far can the agent see?•What does the code look like?2. Is it within the viewing angle of the agent?•What is the agent’s viewing angle?•What does the code look like?3. Is it unobscured by the environment?•Most expensive test, so it is purposely last•What does the code look like?CS 4455 11Sensing:Vision ModelIsn’t vision more than just detecting the existence of objects?What about recognizing interesting terrain features?–What would be interesting to an agent?CS 4455 12Sensing:Human Hearing ModelHumans can hear sounds–Can recognize sounds•Knows what emits each sound–Can sense volume•Indicates distance of sound–Can sense pitch•Sounds muffled through walls have more bass–Can sense location•Where sound is coming fromCS 4455 13Sensing:Modeling HearingHow do you model hearing efficiently?–Do you model how sounds reflect off every surface?–How should an agent know about sounds?CS 4455 14Sensing:Modeling Hearing EfficientlyEvent-based approach–When sound is emitted, it alerts interested agentsUse distance and zones to determine how far sound can travelCS 4455 15Sensing:CommunicationAgents might talk amongst themselves!–Guards might alert other guards–Agents witness player location and spread the wordModel sensed knowledge through communication–Event-driven when agents within vicinity of each otherCS 4455 16Sensing:Reaction TimesAgents shouldn’t see, hear, communicate