Skip to content

Ideas page for MRPT Google Summer of Code 2016

Jump to bottom Edit New page
Jose Luis Blanco-Claraco edited this page Feb 16, 2016 · 53 revisions

Important information for Students

Being accepted as a GSoC student is a quite competitive process. Please, if you wish to submit a proposal, consider becoming familiar with the involved technologies first. Simply copying and pasting from this page will not be accepted.

How you will be evaluated if you are an accepted student

Student projects will be paid only if:

  • Midterm (20-27 June 2016): A pull request is requested that...
    • Builds, ideally in travis-ci
    • Has, at least, stubbed out new functionality
    • Code has appropriate Doxygen documentation
    • Has a stubbed out example/tutorial/ROS launch file that builds/runs without errors
    • It observes the recommendations in "How to contribute", which include C++ style suggestions.
  • End of summer (23-29 August 2016):
    • A complete pull request that builds, full Doxygen documentation, unit test if applicable, complete functionality.
    • A video (e.g. on YouTube) demonstrating your code.

Students interested in applying

  • You must already be proficient in C++.
  • Take your time to learn about MRPT. Try downloading it and launching demo applications, for example.
  • Take a look at the projects in the "GSoC ideas page". Discuss those of your interest, or your own ideas, in the MRPT forum/mailing list.
  • Read carefully about Google's student eligibility rules.
  • Sign up in GSoC and post your project proposal to the MRPT Organization.
  • Don't forget including a coding portfolio of past/current projects and your GitHub username in your application.
  • Rules say that all communications between students and mentors should happen in public: please, use the MRPT mailing list or GitHub comments in pull-requests or commits whenever possible.

Ideas list for GSoC 2016

MRPT provides developers with portable and well-tested applications and C++ libraries covering data structures and algorithms employed in common robotics research areas. ROS (Robot Operating System), supported by the OSRF, provides libraries and tools to help software developers create robot applications. It provides hardware abstraction, device drivers, libraries, visualizers, message-passing, package management, and more.

At present (Feb-2016), only a minimal part of MRPT functionality is available to ROS users, hence that many of the project ideas are aimed at developing MRPT-powered ROS packages.

List of potential mentors:

Project #1: Robust SLAM/localization with artificial visual landmarks and stereo vision

  • Brief description: Design a set of "QR code-like" bidimensional markers suitable for recognition from several meters away, write (or integrate existing) C++ code to detect and unequivocally identify them by some unique ID numbers in stereo images. A set of non-aligned points (a bidi array of corners?) must be detected with (sub)pixel accuracy to enable the optimization algorithm to converge to a non-degenerate solution. With all these, you can write code for automatically reconstruct the 3D pose (SE3) of all markers and the exact trajectory of the camera. This is called Bundle Adjustment or SLAM, and consists in solving an optimization problem, most of its pieces already implemented in MRPT. Also, once a map of markers is generated, one can run the optimization for the camera poses only, allowing accurate and robust localization, with applications to autonomous vehicle navigation, augmented reality, etc.

  • Expected results:

  • A new C++ class in mrpt-vision for QR-like markers detection. When finished, its integration into OpenCV will be considered.

  • Generic new C++ classes in mrpt-slam for the SLAM and localization problems.

  • A working application capable of running the new methods on live or offline stereo visual datasets.

  • The method must work for professional stereo cameras (e.g. Bumblebee2). The organization will provide the student with stereo images of the markers designed by him/her.

  • The method may optionally also work with a pair of consumer-grade webcams, in the case the student wants to print his/her own markers for testing at home.

  • Knowledge prerequisites: C++, computer vision, OpenCV, numerical optimization.

  • Difficulty level: Middle

Project #2: Integrate MRPT reactive navigators into ROS move_base

  • Brief description: Use ROS pluginlib to integrate MRPT TP-Space reactive navigation for autonomous vehicles/robots as a local_planner in move_base. Further reading:
  • Expected results: Being able to launch a navigation command in ROS' RViz so it is firstly processed by existing ROS global planners and the path followed locally by MRPT methods. Writing a good tutorial on ROS Wiki about the process would be also a great outcome.
  • Knowledge prerequisites: C++, GNU/Linux, vehicle kinematics.
  • Difficulty level: Middle-high.

Project #3: ROS package for 2D/3D range-bearing EKF SLAM/Localization

Useful for features-based mapping and localization: laser corners This task only includes the definition of a generic SLAM/Localization node, without integration with real sensors.

  • Brief description:
  • Expected results:
  • Knowledge prerequisites: C++, GNU/Linux.
  • Difficulty level: Middle.

Project #4: ROS package(s) for versatile RBPF-based SLAM

MRPT features RBPF-SLAM with different kinds of map models and sensors. Create ROS nodes for some practical cases:

  • LIDARs + occupancy grid maps.

  • Range-only sensors + beacon maps.

  • Brief description:

  • Expected results:

  • Knowledge prerequisites: C++, GNU/Linux.

  • Difficulty level: Middle / Middle-high.

Project #5: Extend existing ROS packages for grabbing and playing back MRPT datasets

  • Brief description: Many datasets are available in MRPT Rawlog format or can be imported to it (e.g. Carmen logs). ROS log format (rosbag) is incompatible, so run-time or off-line conversion tools are required if Rawlog datasets want to be used in a ROS application.
    • mrpt_rawlog ROS wiki page of existing packages. Very limited set of sensors are implemented at present.
  • Expected results: Being able to view all/most sensory data from a complex Rawlog dataset, e.g. Málaga Urban Dataset, in ROS RViz. It may be done via run-time on-the-fly conversion (easier) or by a new command-line tool capable of converting among rawlog and rosbag formats (harder).
  • Knowledge prerequisites: C++, GNU/Linux.
  • Difficulty level: Low / Middle-Low.

Project #6: New GUI app to visualize and edit robot-made maps

  • Brief description:

    • During 2005-2008 there existed a complex GUI application with this purpose called SimpleMapsViewer, but it was Windows-only (written in Borland C++ Builder!) so it was dropped when MRPT became crossplatform.
    • It is strongly recommended to have a look at a working snapshot of the old app (from 2008!) to get an idea of the expected basic functionality: (SimpleMapsViewer-MRPT-0.5.3-win32.zip)[http://ingmec.ual.es/~jlblanco/downloads/SimpleMapsViewer-MRPT-0.5.3-win32.zip] Use: Extract the ZIP, execute SimpleMapsViewer.exe and load the example map file localization_demo.simplemap.
    • At present, the only similar app in MRPT is observations2map, which allows generating metric map but without a GUI and without edit possibilities.

  • Expected results: A new GUI app should be able to load recent versions of MRPT simple maps as generated from ICP-SLAM or RBPF-SLAM. The app must allow inspecting the raw observations (at least, 2D LIDAR scans and monocular and stereo images) and generate one or more metric maps according to user-given parameters, e.g. 2D point clouds, 2D occupancy grid maps, a 3D octomap, etc.

  • Knowledge prerequisites: C++, GUI design (Qt or wxWidgets).

  • Difficulty level: Middle

New project template:

  • Brief description:
  • Expected results:
  • Knowledge prerequisites:
  • Difficulty level:
Add a custom footer
Add a custom sidebar
Clone this wiki locally