Homing behaviour experiments


A Simple Homing Strategy

We decided to use a strategy which restricts the a priori knowledge necessary to the robot.
For this, the ants behaviour will be imitated: they mark their trajectory with pheromones.
We don't have pheromones detector, but the robot is equipped with a tokens distributor.
The tokens are painted in red and are used by the robot to mark the corridor at his arrival
at a crossing on the way; a simple process detects the places where to put marks. These
red marks on the ground are the only a priori knowledge the robot will use on its way back.
In addition, we consider that the number of bolts in the last crossing is more than 5. The robot
has then just to follow up the corridor until a crossing.
The first behaviour is to follow up the corridor. When the robot arrives at a crossing, he chooses
the corridor direction which has more red marks on the ground. He turns then in that direction.
If the number of marks is less than 5, he resumes a follow up corridor behaviour. When the
detected marks number is sufficient (more than 5), we consider that the robot has arrived in the
vicinity of the base. He then goes back to the base guided on a visual landmark.

Demo (AVI, 16.6 Mo)


Report (in french):
Nicolas Mansard (2003); Behaviour Hierarchy: Homing Behaviour Application / Hiérarchie de Comportements : Application au Retour à la Base;
DEA IVR ; Institut National Polytechnique de Grenoble / Université Joseph Fourier
PDF (2.7 Mo)