Ant-like Navigation
top-view-last.jpg
size="square"
size="square"
size="square"

Ant-like Navigation

by Francesco

Done with 1 Thymio

Keywords:

Navigation, mapping

Construction projects

Basic concept

The navigation of a robot consists in going from one point to another. For example one can call navigation the action of a robot that goes from a base in the desert to a source of water to be pumped, then taken back to the base. Imagine that the robot can at any time return to the base because the base emits a signal it can receive. How to search for the water and find the path that leads from the base to the water? In robotics, the classical solution consists in creating a map in the memory of the robot, but this requires a sophisticated approach, good sensors or high computational power. In this example we will not make the map in the memory of the robot, instead we will make the map directly on the ground. This is what ants do to create a track between their nest and a food source. Having once found the food, they return to the nest dropping on the ground a chemical trace which they can then follow to find the food.

How to make robots that navigate like ants

We will therefore reproduce the ant behaviour which is to leave a trace on the ground when going back to the base having once found the water. The goal is then to be able to follow the track and find the water starting from the base.

Ideally we should have a robot that can deposit on the ground a substance which will be needed when going back to the base. However we have a robot that can write on the ground with a pen. It cannot lift the pen up and down, it writes constantly. So let's reinforce the pen trace when it goes back to the base. Then, ifn the trace is strong enough, it will follow it to find the water.

How to return to the base? Thymio could do it in at least two ways:

  • With a light on top of the base. This method has been tested and, in the configuration used, was not very precise.
  • By putting the whole environment on a slope and following the slope to find the base. This is what is done in this example. This works quite well.

Here are a few images of the test configuration and the final result.

Configuration on a tilted table:

video-final-setup-700.jpg

Result in video:

Explanation of the video:

  • The robot starts to move randomly (white Thymio)
  • If it finds the target (in the centrer, black circle) ite turns towards the base (Thymio becomes red) and moves toward it, moving in such a way as to leave a stronger mark on the ground.
  • The mark on the ground must be strong enough to be followed. So initially the Thymio continues to seek randomly and marks the return, up to when it starts to detect the trace it left. In this case it tries to follow (Thymio takes the green colour).
  • Once the trace is strong enough, the Thymio can go from the base to the target directly by the shortest path.

Download the code

| Click here

Link to more info

Comment this creation!

Add a New Comment
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License