Youtube Video

Daisy

A prototype built to explore simulating dynamic nested parameter spaces for autonomous driving in a model-based manner, at a relatively high level of abstraction.

The name comes from the analogy that each single parameter instantiation grows multiple petals like a daisy , in which other dynamic parameters are found, ad infinitum.

Prerequisites

• Python >= 3.7
• pip
• virtualenv

Setup Instructions

• cd to ./dev
• sh ./setup_venv.sh
• source ./activate_venv.sh
• pip install -r ./requirements.txt

Running Instructions

• When starting a development session, enter the venv with
  • source ./dev/activate_venv.sh
• When ending a development session, exit venv with
  • deactivate

Example Scenario

• ./run_scenario.py lidar
• ./run_param_master.py -scenario lidar -flow.frames 300 -flow.headed -flow.samples 12 -flow.threads 6
  • Note that a tiling window manager is recommended, since this will open many windows, and by default they will hide each other

Scenarios

• Scenarios are found in ./py_scenarios
• The first argument to ./run_scenario is a pattern against which the files in ./py_scenarios are matched

Documentation

• To generate a class diagram, you can use pyreverse
  • pip install pylint
  • cd to proto3
  • pyreverse ./
  • xdot classes.dot