cs4100-crowd-navigation

Pygame-based sandbox for crowd-aware robot navigation experiments.

Quick setup

python3 -m venv .venv
source .venv/bin/activate
pip install -U pip
pip install -r requirements.txt

Run the simulation:

crowd-sim

You can also run it directly:

python -m crowd_navigation.main

Controls:

• WASD or arrow keys to move the robot
• Close the window to exit

Optional extras for this proposal

Install all extras (analysis + RL + crowd + dev):

pip install -r requirements-dev.txt

Or install only what you need:

pip install -e ".[analysis]"
pip install -e ".[rl]"
pip install -e ".[crowd]"
pip install -e ".[dev]"

What each extra is for:

• analysis: metrics and plots (time-to-goal, collisions, personal-space intrusions)
• rl: reinforcement learning baselines in partially observable settings
• crowd: optional velocity-obstacle style crowd-flow tooling (pyRVO)
• dev: testing and linting

Project structure

src/crowd_navigation/main.py   # Pygame simulation entrypoint
tests/test_smoke.py            # Basic test scaffold
pyproject.toml                 # Package + dependencies