Mapf is a (currently experimental) Rust library for multi-agent planning, with a focus on cooperative path finding.
This repo is just getting started, so please bear with us for a while as we work towards making useful tools.
This is being developed as part of the Open-RMF project which provides an open source framework for enabling interoperability between heterogeneous fleets of mobile robots, including cooperation across different platforms and vendors.
We tend to always target the latest versions of Rust, so we recommend using the rustup tool: https://www.rust-lang.org/tools/install
$ curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | shIf you already have rustup then you can use this to bring your installation up to date:
$ rustup updateFollow instructions for installing rustup-init.exe here.
We are currently using a few unstable features from the nightly toolchain of Rust
because they allow mapf to have extreme customizability without any loss to
performance. The easiest way to use the nightly version is to make it your default
toolchain with this command:
$ rustup default nightlyFrom the root directory:
$ cargo run --release --example gridThis example will allow you to experiment with multiple agents cooperatively planning from their starts to their goals in a grid environment. Use left/right click to set the start/goal cells of each agent. Use shift+left/right click to add/remove occupancy from cells. The velocities and sizes of each agent can be customized independently, and you can save/load scenarios into yaml files.
Some premade scenarios can be found in the mapf-viz/scenarios folder. Load a
scenario on startup by passing the scenario name as an executable argument, e.g.:
$ cargo run --release --example grid -- mapf-viz/scenarios/sizes.yamlNote that the scenario filename to load at startup must come after a -- with a space on each side.