From game AI to production robotics
Behavior trees originated as a way to control non-player characters in video games, where designers needed AI that felt reactive and layered without hand-coding a huge web of state transitions for every character. That same need, modular, reactive decision-making that's easy for a team to build and debug together, turned out to map directly onto robotics, and the pattern migrated over largely unchanged in structure.
In robotics, BehaviorTree.CPP is a widely used open-source C++ library for building and executing behavior trees, and it includes tooling for visualizing and monitoring a tree's ticks in real time, which matters enormously when you're trying to understand why a robot chose one fallback branch over another. ROS2's Nav2 navigation stack, the same Nav2 you've seen if you've taken Autonomous Navigation & SLAM, uses a behavior tree under the hood to orchestrate the higher-level logic of navigating: deciding when to plan a new path, when to recover from being stuck, and when to give up and report failure, while the path planning and obstacle avoidance itself is handled by separate, lower-level components the tree simply calls into.