build details

Show: section status errors & todos local changes recent changes last change in-page changes feedback controls

Duckietowns and Duckiebots

Modified 2018-12-19 by tanij

Knowledge of the relation between Duckietown and Duckiebots, or, what can Duckiebots do in different Duckietowns.

Not all functionalities of the Duckiebots require all Duckietown city elements to work. If you are wondering what learning experiences you can explore with what Duckietown, you are in the right place. We define Duckietown configurations of increasing complexity depending on what Duckiebot functionalities they support.

Only traffic signs (no city)

Modified 2018-12-16 by tanij

Without a Duckietown, it will not be possible for Duckiebots to showcase most of their behaviors. Nonetheless, traffic signs have AprilTags, which allow to play with the AprilTag detection and relative pose estimation pipeline, and in turn with the camera calibration and system identification procedures.

City loops

Modified 2018-12-16 by tanij

The fundamental Duckiebot behaviors can be explored with a very simple Duckietown. We call these simple cities: loops.

City loops are closed road patterns, without intersections, that meet the appearance specifications. They can be made only of floor layer elements, with no signals.

City loops enable to play with the:

  • perception pipeline: what happens when data is obtained in form of measurements (i.e., images in the DB17 and DB18 configurations), and how it is used to extract information, leading the Duckiebots to generate a belief of their position and orientation in the lanes.

  • lane control: the process that Duckiebots apply to transform beliefs provided by the perception pipeline in decisions on how to steer in order to stay inside the lane.

  • traffic management: the functionality for which Duckiebots stay at a safety distance from other Duckiebots driving in front of them.

When city loops are augmented with traffic signs, e.g., road names, it is then possible to explore the functionality of:

  • AprilTag detection and relative pose estimation: how Duckiebots detect AprilTags, and determine their position and orientation relative to them. This information can be used to localize in the map.

If you acquired a Duckietown Starter Pack, you have a city loop with traffic signs.

Navigable cities

Modified 2018-12-19 by tanij

Thee more complex Duckiebot behaviors require intersections and other city elements such as traffic lights to work. We define cities that include intersections as navigable cities.

Navigable cities are city loops connected by intersections, meeting the appearance specification. Navigable cities require traffic signs.

Navigable cities enable testing of Duckiebot behaviors. A complete list (for one or more Duckiebots) of behaviors can be found in the demo section of the Duckiebot operation manual.


Modified 2018-12-19 by tanij

Robotariums (Robotic aquariums) are experimental Duckietowns designed for continuous operations. Robotariums require a maintenance area in addition to a Duckietown, providing essential functions like automatic charging. More information on Robotariums can be found in the Robotarium book.

TODO: switch last two links with proper inter-book references

previous task next (5 of 26) index

The following was marked as "todo".

TODO: switch last two links with proper inter-book references

File book/opmanual_duckietown/05-duckietown-preliminaries-definitions.md.

File book/opmanual_duckietown/05-duckietown-preliminaries-definitions.md
in repo duckietown/docs-opmanual_duckietown branch master17 commit 8689dcb1
last modified by tanij on 2018-12-19 17:30:10

Created by function create_notes_from_elements in module mcdp_docs.task_markers.
Because of mathjax bug

No questions found. You can ask a question on the website.