View on GitHub

Communication in SHOP4CF

Documentation of communication among SHOP4CF components

Human Aware Mobile Robot Navigation in Large-Scale Dynamic Environments

Short name: HA-MRN

By JVERNE-FZI

FZI Component

Purpose

This component takes sensor information and robot localization, calculates the best trajectory avoiding obstacles on the way, and returns trajectory points for the most efficient path

Data interfaces

Input and output data (but not user interfaces):

  1. INPUT: Real-time connection to lidar
    • Format: ROS Topics
    • Real-time constraints: Real time for trajectory avoidance is requested
    • Expected data volume: at least 30 Hz
    • Other details: the expected data volume depends on the LIDAR
  2. INPUT: Real-time Camera stream
    • Format: ROS Topics
    • Real-time constraints: Real time for trajectory avoidance is requested
    • Expected data volume: at least 30 Hz
    • Other details: the expected data volume depends on the camera to be used and the trained Neural Network for image post-processing
  3. INPUT: Robot positioning
    • Format: ROS Topics
    • Real-time constraints: During the goal achievement update on current position of the robot in a real time is requested to avoid collision and estimate if the goal (delivery) was achieved or not
    • Expected data volumen: at least 30 Hz
  4. INPUT: Robot destination definition
    • Format: Coordinate point in a bitmap (Global map) with respect to the origin
    • Real-time constraints: No
  5. INPUT: Global map
    • Format: Bitmap and YAML File (specifying meta information about the map, like origin, resolution and size)
    • Real-time constraints: No, once the map is stored no updates on the workspace are done
  6. OUTPUT: Global/Local plan, local navigation waypoints
    • Format: Coordinates in the bitmap (Global map) with respect to the origin
    • Real-time constraints: yes. Obstacles might change the pre-established path
    • Expected data volume: at least 30 Hz
  7. OUTPUT: Mobile platform commands
    • Format: ROS topics
    • Real-time constraints: yes. The platform might be requested to stop, change direction or return the path, if an obstacle appears in the path
    • Expected data volume: at least 10 Hz
  8. OUTPUT: ROS MoveBase Stack
    • Format: ROS Topic
    • Real-time constraints: yes. The platform will be constantly informed about the trajectory to follow, and any changes in its path if there is an obstacle.
    • Expected data volume: at least 10 Hz

The interaction with another components to obtain or use the input and output data is still to be decided.