RoboCup 2D Half Field Offense

WARNING: HFO is not actively maintained! We accept pull requests and bug fixes, but please be prepared to do your own debugging in case of a problem. If you'd like to become a maintainer of HFO, get in touch.

Half Field Offense in RoboCup 2D Soccer is a subtask in RoboCup simulated soccer, modeling a situation in which the offense of one team has to get past the defense of the opposition in order to shoot goals. This repository offers the ability to quickly and easily interface your learning agent with the HFO domain. Interfaces are provided for C++ and Python.

NOTE: The previous distances, normalized to a range of -1 to 1, returned in the two state spaces were being divided by a too-small maximum distance; the correction for this should be accounted for in existing code. (An example of such a correction can be seen in ./example/high_level_custom_agent.py.)

Dependencies

Boost-system, filesystem
Qt4 [Required for SoccerWindow2 visualizer]: To not build the visualizer, add cmake flag -DBUILD_SOCCERWINDOW=False

Python Dependencies

Python 2.7 or 3.2+ (tested with 2.7 and 3.5)
numpy

Install

mkdir build && cd build
cmake -DCMAKE_BUILD_TYPE=RelwithDebInfo ..
make -j4
make install

Note: There is no need for 'sudo' with 'make install' - the files will be installed below the starting directory.

Python Install [required for python interface]

From the main HFO directory: pip install [--user] .

Demos

From the main HFO directory:

./example/passing_agents.sh

Start a 1v1 game played by Agent2D:

./bin/HFO --offense-npcs=1 --defense-npcs=1 --no-sync

Start an example agent on the empty goal task. This agent will move forwards slowly. First start the server: ./bin/HFO --offense-agents=1 --no-sync & and then connect the agent: ./example/hfo_example_agent

Or do both in a single command:

(./bin/HFO --offense-agents=1 --no-sync &) && sleep 1 && ./example/hfo_example_agent

Example Agents

Example agents are provided in the example directory. Below are two minimal examples:

C++ Agent

HFOEnvironment hfo;
hfo.connectToServer(...);
for (int episode=0; episode<5; episode++) {
  status_t status = IN_GAME;
  while (status == IN_GAME) {
    const std::vector<float>& feature_vec = hfo.getState();
    hfo.act(DASH, 20.0, 0.0);
    status = hfo.step();
  }
  cout << "Episode " << episode << " ended";
}

Python Agent

hfo = hfo.HFOEnvironment()
hfo.connectServer(...)
for episode in range(5): # replace with xrange(5) for Python 2.X
  status = IN_GAME
  while status == IN_GAME:
    features = hfo.getState()
    hfo.act(DASH, 20.0, 0.0)
    status = hfo.step()
  print 'Episode', episode, 'ended'

Documentation

The state and action spaces provided by the HFO domain are documented in the manual.

LARG/HFO