Evolution of Implicit and Explicit Communication in a Group of Mobile Robots
Joachim de Greeff, Stefano Nolfi
This page contains support material of the paper "Evolution of Implicit and Explicit Communication in a Group of Mobile Robots"
Experimental Setup
The environment and robots are either simulated or real. In simulation, the program Evorobot is used and in reality the e-Puck robotic platform (Mondada & Bonani, 2007). The parameters are equal in both cases. The experimental setup involves two wheeled robots situated in an arena containing two target areas which are evolved for being concurrently located in the two target areas and for switching areas as often as possible. In this section the environment, the robot, the neural controller and the evolutionary algorithm is described. For details of all parameters used see the paper.
The Environment
The environment consists of an arena of either 110x110 or 150x150 cm surrounded by walls and containing two target areas with a diameter of 34 cm placed on two randomly selected but non-overlapping positions inside the arena. The floor of the arena and the walls are grey. The two circular portions of the arena corresponding to the two target areas are painted in black and white, respectively.The Robot
The robotic platforms consist of two e-Puck robots provided with the ground sensory board extension. The robots, which have a diameter of 7.5 cm, are equipped with 2 motors which control the 2 corresponding wheels, 8 infrared proximity sensors located around the robot’s body, 3 infrared sensors placed on the frontal side of the robot and oriented toward the ground, a VGA camera with a field of view of 36° pointing in the direction of forward motion and a wireless Bluetooth interface which can be used to send and receive signals to and from other robots.The Neural Controller
The Evolutionary Algorithm
The free parameters of the neural controller - the connection weights, the biases of internal neurons and hand actuators and the time constant of leaky-integrator neurons - have been adapted through an evolutionary robotics approach using the following parameters:
Population size: 100 Number of generations: 1000 Genotype: 158 genes/parameters encoded in 8 bits each, a string of 1264 bits in total Mutation rate: 2% Crossover: not used Selection type: ranking selection Offspring: five (with elitism) Fitness: number of switches between target areas
Videos of evolved solutions, in simulation and in real robots
Two different strategies evolved in environments with different sizes, we refer to these different strategies as "symmetrical strategy" and "assymetrical strategy". Below are videos of the two strategies, both in simulation and in a real environment using e-puck robots. For a detailed description, see the paper.