Skip to content

fabid/BehaviorTree.jl

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

BehaviorTree.jl

Behavior Trees (BTs) are a powerful way to describe the behavior of autonomous agents with applications in robotics and AI.

This implementation is based on Behavior Trees in Robotics and AI: An Introduction. It was developped by Team L3 to qualify in NASA'Space Robotics Challenge Phase 2 (More details on usage in this paper: Human-Robot Teaming Strategy for Fast Teleoperation of a Lunar Resource Exploration Rover).

Installation

Optional System Dependencies

Visualization tools depends on graphviz.
On Ubuntu/Debian:

sudo apt-get install graphviz

Install Julia Package

With Julia ≥ 1.4 (may work on previous 1.x version but not tested) add package

julia> ]
(v1.4) pkg> add BehaviorTree

Basic Usage

Two Primitives are available to build behavior trees: Sequence and Selector. They accept a list of tasks, that can be a Sequence, a Selector, or, in the case of a leaf, a function returning one of :success, :failure or :running.

doSuccess(bb=Dict()) = :success
isTrue(bb=Dict()) = :success
doFailure(bb=Dict())= :failure
doRunning(bb=Dict()) = :running
bt = Selector([
    doFailure,
    Sequence([isTrue, doFailure, doSuccess], "choice"),
    doRunning,
    doSuccess!
], "head")

Execution of the tree happen via the tick function, that accepts a shared blackboard object use to share state between tasks.

blackboard = Dict{Symbol, Any}()
status, results = tick(bt, blackboard)

Ticking usually happens in a loop at at a frequency determined by the needs of the application.

Visualization

As BehaviorTrees use the AbstractTree interface, it is possible to use D3Trees for visualization:

d3tree = D3Tree(bt)
inbrowser(tree, "firefox")

Utilities to generate graphviz output via the .dot format are also provided.

dot_graph=toDot(bt)
filename= "example.dot"
open(filename, "w") do dot_file
    write(dot_file, dot_graph)
end
filename= "example.png"
png_graph = dot2png(dot_graph)
open(filename, "w") do png_file
    write(png_file, png_graph)
end

behavior tree overview

passing the execution results in the toDot function generates a visualization of the current state:

dot_graph=toDot(bt, results)
filename= "status.png"
png_graph = dot2png(dot_graph)
open(filename, "w") do png_file
    write(png_file, png_graph)
end

behavior tree overview

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages