Simple Robot with NN

A minimal grid-world tree where the locomotion leaf is driven by a trained ONNX classifier. Useful starting point for your own neural-symbolic behavior tree.

Location

examples/Simple-Robot-With-NN/

Features exercised

• A 5×5 grid blackboard.

• A single classification neural_network leaf producing one of {left, right, up, down, stay}.

• A selector that guards the NN leaf behind a bounds-check check node.

Invocation

Formal verification (requires nuXmv):

python -m behaverify nuxmv \
    examples/Simple-Robot-With-NN/model.tree ../out/ \
    --generate --ctl --invar --nuxmv_path ../nuXmv

Simulation (pure Python):

python -m behaverify python \
    examples/Simple-Robot-With-NN/model.tree ../out_py/

Then run python ../out_py/python/runner.py to watch the robot navigate.