Behavior Trees

Definition

Behavior trees compose autonomous agent logic from a tree of nodes (sequence, selector, parallel, decorator) whose tick semantics produce running, success, or failure states. They are widely adopted in robotics and game AI because they are inspectable, easily reusable, and naturally support fallback chains. Loyal-wingman autonomy stacks frequently express tactical behaviors—escort, ingress, decoy, RTB—as composable subtrees orchestrated by a higher-level planner.

Reference attributes

Pattern

Hierarchical, tick-driven

Common nodes

Sequence, Selector, Parallel, Decorator

Strength

Inspectability, composability

Common pairings

BT + HTN planner; BT + utility AI

Related terms

• Hierarchical Task Networks (HTN)
  Planning paradigm that decomposes tasks into ordered subtasks via methods.
• Utility AI
  Decision pattern that scores candidate actions by weighted utility curves.
• Finite State Machines (FSM)
  Classical state-and-transition control model for bounded autonomous behavior.