Social Navigation with RoboComp

Social Navigation

My project for GSOC’17 has consisted in Social Navigation. The main goal has been to give mobile robots the capacity of behaving in a social acceptable manner.

To achieve this aim, in first place, the personal space of humans has been modelled, which has allowed the robot to respect their personal space. Secondly, the present humans in the environment have been clustered according to their distances in order to avoid the robot interfering among interacting humans. The navigation architecture has been modified to allow the robot to recognize the personal space of the individuals and the different groups of people.

It was necessary to fix different bugs in the code that prevented the progress in the development. The time required for this task is the reason why all the goals set couldn’t be achieved.

Once these problems were solved, parts of the code were improved in order to make it more efficient. However, the proposed algorithm has still room for improvement.

Finally, a new method that allows the robot to recognize if humans are interacting with objects was created.


For a more detailed explanation of the work done during GSOC’17 visit the following links:

  1. Introduction to Social Navigation
  2. Modifications in the Navigation Architecture
  3. My progress

Content in GitHub

The work done during GSOC’17 can be found in my GitHub repository. It can be accessed in the following link.

Also, all the commits made for this project can be found in the following link.

Araceli Vega Magro