This library IS NO LONGER BEING MAINTAINED! Please see Pathfinder2. You're free to use this library if you choose, but I'm just going to warn you right now that there are a decent amount of bugs in it. Pathfinder2 does pretty much exactly the same thing, but better.
A feature-packed and thoroughly implemented library for autonomous navigation designed for wheeled mobile robots. Click right here to get to Pathfinder's online documentation, or keep reading to learn more.
- Virtual Euclidian geometry - prevent collisions and autonomously navigate around complex paths.
- Out-of-box support for most drivetrain types, including...
- Meccanum Drive
- Tank Drive
- Swerve Drive
- Abstract and re-usable codebase that can be implemented anywhere. Offically supported environments include:
- FIRST Robotics Competition
- FIRST Tech Challenge
- Easy-to-use and expansive API.
- Tons of customizability and configuration options.
- Well-documented and in active development.
- Officially endorsed by NASA! Not really. Please don't sue me. Please.
Please see the online quickstart, available here.
Pathfinder began development around the end of November, 2020. Pathfinder was/is designed to be a full-featured solution to path planning and following for wheeled mobile robots.
- Provide a simple interface for interacting with paths, following paths, and doing whatever else you want to do with paths. Pathfinder has the intention of providing everything from path finding (object avoidance, etc) to path following. The original intention of this library was to create a solution that would allow an end user to command the robot to go to a point with as little as a single line of code. Other similiar libraries already exist, such as ACME Robotic's road-runner library, but Pathfinder is (as of now) the only library that provides everything you need as soon as you install the library.
- Allow for serialization and deserialization of paths - paths should be able to be stored as JSON files, meaning they can be loaded and saved during execution. This means you can create a path using another GUI tool (or a text editor), drop the JSON file into your robot's "deploy" section (or source files, or resources - whatever you'd like) and load the path from your robot's internal storage system and begin following said path.
- Allow users to track the position of their robot by using odometry. A variety of odometry systems should be provided - notably, tank drive, swerve drive, three wheel, and two wheel. Gyroscopes should be used whenever possible to increase positional accuracy.
- Allow users to follow a given path using a built-in path following system. Path followers should cover the following categories: linear, proportional, and hopefully trajectory.
- Allow users to load paths in absolute and relatives modes. An absolute path will command the robot to go to the exact postions you notated - if the path starts at (10, 10) and your robot is at (20, 20), your robot will drive to (10, 10) and then continue the path as intended. Relative paths, on the other hand, are loaded relative to the robot's current facing. A robot at (10, 10) told to go to (20, 20) wouldn't go to (20, 20) - it would go to (30, 30), as (10, 10) + (20, 20) = (30, 30).
- Provide kinematics for each major type of holonomic drivetrain. These kinematics should be usable both by Pathfinder internally and users externally - if you get a brand-new meccanum drive and want to set it up really quickly, you can easily do so by using these provided kinematics.
- Allow for advanced movement trajectories and splines with error-based following, allowing robots to follow a path at the robot's maximum movement speed.
- Allow any users to use any of their own code in place of provided code. At the very core of Pathfinder is customizability - just about every feature should be expandable, extensible, replacable, customizable, optimizable, etc. If you don't like the provided drivetrain classes, you can make your own. If you don't like the provided odometry classes, you can make your own.
- Provide a graphical user interface that allows users to create and modify paths with ease. These paths should be able to be created by providing a set of waypoints. Trajectories should be able to be generated by using simple and intuitive controls.
- Be really epic. Yes, I do mean really epic - like... REALLY epic. Cool guys don't look at explosions. Cool robotics libraries are named "Pathfinder."
Pathfinder is a rather expansive library designed for a wide variety of pathfinding situations. As a result, it can be pretty challenging to document every single feature and option contained in the library. It's a lot easier to provide guides on how to make use of specific features in video form, as they take a lot less time to create and are generally easier to understand. Please see this link to learn more.
If you notice an issue or bug with Pathfinder, please report it on the issue tracker and send me an email. It's more important that you submit an issue report so the issue is documented in a public fashion. Issues or bugs will be resolved as quickly as possible whenever possible.
Pathfinder is a fairly large project, and, as such, we always welcome contributors. If you notice any bugs with Pathfinder, or would like to add any functionality, feel free to submit a pull request. You can always feel free to add new features that don't detract from Pathfinder's functionality. Examples include:
- New drivetrains (add drivetrains that aren't already supported)
- New following systems (new way to follow paths)
- New path generation system alternatives (new generator classes)
Pathfinder uses Gradle as a build system. Because Pathfinder is a library, the only task
you'll really need to run is jar or shadowJar. shadowJar will produce a jar with all
of the required dependencies bundled inside. This jar is about 300kb and includes all of
the libraries Pathfinder needs to operate. jar produces a jar much closer to 100kb in size.
It doesn't include any of the libraries Pathfinder needs to function - you'll need to add them
manually.
- OdometryCore Used for three wheel positional tracking. This library is very lightweight, so including it even when it isn't used shouldn't harm anyone.
- EDT Optimized data types used internally by Pathfinder. EDT is user-usable, but you can still use more standard data types if you'd prefer.
- intra_utils A variety of utilities used by Pathfinder internally - intra_utils is really just a collection of random code I've found I've needed in many projects.
- PathfindingCore Actual path finding code - given point A and B and obstacles {X, Y, Z}, find a path between A and B that doesn't intersect with obstacles {X, Y, Z}.
Pathfinder strives to be a well-documented project - every method, class, field, interface, enum, etc - should all have documentation. Documentation, in its current state, is available from either of these two places.
If you notice any documentation issues, or are struggling to understand the meaning of any documentation, please contact me (Colin) ([email protected]) and I'll try to get any issues sorted out as quickly as possible.