Evolution of Communication

Exploring the emergence and evolution of communication
in social organisms using collective, evolutionary robotics

Sara Mitri
Steffen Wischmann
Prof. Dario Floreano
Prof. Laurent Keller

The aim of this project is to address questions on the emergence and evolution of communication in groups of social organisms by using evolutionary robotics to build societies of autonomous robots that evolve a communication system to solve a particular survival task collectively. The project began as part of the ECAgents project, and is now funded by the Swiss National Science Foundation.


Communication is an integral part of the social behaviour of all organisms, ranging from bacteria to humans. Despite extensive research to uncover its mysteries, its evolutionary dynamics remain elusive. Since communication does not fossilise, studying its evolution is an immensely challenging task. We propose that by applying experimental evolution with groups of robots, we can overcome some of the difficulties faced by empirical studies, as well as other modelling techniques, such as game-theoretical or mathematical modelling. The evolutionary approach allows us to access all the data about the evolving communication system and the corresponding behaviour in a relatively complex mechanistic model.

In previous and current work we have focused on the following issues:

  • Conditions for the emergence of communication

  • How does a communication system evolve and stabilise in a population? If the cost for signalling is high, why would an individual be interested in sharing information with the rest of the group? Our research has shown that different levels of genetic relatedness and levels of competition between/within groups highly influence the emerging communication system and its stabilisation.
  • The evolution of communication in competing populations

    Communication between conspecifics can significantly enhance cooperative behaviour. However, such signals can also have detrimental effects because they can be exploited by other competing organisms. Here, we investigate the evolution and effectiveness of intragroup communication strategies and their robustness to intergroup competition. 
  • The transition to symbolic communication

  • What are the differences between human and non-human communication? What is meant by symbolic communication? In what way is animal communication less symbolic than language? We aim to explore these questions through a number of experiments representing the transition from non-symbolic to symbolic communication.


In order to tackle these questions, we evolve neural networks for a population of agents required to solve a survival task of foraging for food and avoiding poison. Artificial evolution takes place under a physics-based simulated environment Enki, where both the robots' sensors and actuators and the experimental setup are modelled. An evolutionary robotic framework Teem is then used to evolve the best controllers to survive in this simulated environment, which are then transferred onto real robots. The resulting behaviours are compared to biological models and used to answer various theoretical questions, such as those listed above.

An evolved communication system used to recruit robots toward a food source can be seen here and here.

Communicating S-bot robots foraging


As a tool for demonstrating various communication mechanisms in artificial systems between groups of robots, two types of robots are being used within this project. The first robot, the s-bot was originally designed for the swarm-bots project (Mondada et al. 2003), which aims to study the self-organisation and self-assembly of groups of robots. The design of the robot was slightly modified in order to fit the requirements of a demonstrator for artificial communication.

The s-bots have a diameter of 12 cm and a height of 15cm and possess 2 Lilon batteries, which give it about an hour of autonomy. A 400 MHz custom Xscale CPU board with 64 MB of RAM and 32 MB of flash memory is used for processing, as well as 12 distributed PIC microcontroller for low-level handling. The s-bots run a custom Linux operating system and communicate with the work-station via WiFi. The robot's actuators consist of 2 treels, a turret capable of rotation and a rigid gripper. Its sensory capabilities include infrared sensors (15 around the turret and 4 below the robot), force and speed sensors, humidity and temperature sensors, 8 ambient light sensors, an omnidirectional camera and 4 microphones. The robots can also emit sound using 2 speakers and light from its light-emitting LED colour turret. We are currently in possession of around 10 s-bots to use in our research.

S-bot mobile robot, built for the swarm-bots project

The e-puck is a miniature (7cm diameter) battery-powered wheeled robot developed at EPFL for use in teaching and behavioural research. The robot is controlled by a dsPIC microcontroller from Microchip capable of basic signal processing. Being relatively inexpensive, the e-pucks are designed to be used for projects on collective robotics, where interactions between multiple robots are needed. For this purpose, we are using 30 e-pucks at our lab. One of the e-puck's main strengths is its versatility: Two connectors on its surface allow the user to stack extension turrets on the main robot, which can be custom-tailored to meet a specific task. Several turrets have already been developed including an LED turret for communication. Due to its higher battery autonomy (2-3 hours of intensive use) and lower cost, we hope to shift the focus from the s-bots to the use of the e-pucks in our future scientific experiments.

EPuck mobile robot

Main Findings

The conditions for the emergence of communication were explored in experiments where two parameters were monitored for their influence on the emergence of communication (see Fig. 1, left):
1. Genetic relatedness within a group of robots (homogeneous, r=1, vs. heterogeneous, r=0, colonies)
2. Level of evolutionary selection (individual or group)

Our results show that honest communication evolves in three of the four tested conditions, thereby increasing fitness compared to a baseline experiment, where no blue light was allowed (see Fig. 1, right). The robots evolved two different stable communication strategies, which were not equally efficient: In the first, they signalled when by the food, whereby receivers evolved an attraction to blue light (Fig. 2, left). In the second strategy, less efficient strategy, signallers emitted light by the poison, while receivers were repelled by blue light (Fig. 2, right). This shows that evolved communication systems need not be optimal in order to be stable.

When agents were unrelated and selected individually, we observed that communication reduced the fitness of the groups. Further investigation revealed that this was caused by the spread of deceptive communication. Due to the large amount of competition between individuals, it was in the signallers’ best interest to reduce the fitness of receivers. This was done by signalling far from the food, given that individuals were attracted to blue light. One might expect that this would lead receivers to cease to be attracted to blue light. However, this does not occur. The receiving strategy is very stable, in conjunction with the deceptive signalling strategy.

Our hypothesis is that the stability of deceptive communication is due to a large amount of noise in the selection process. Receivers remain attracted to blue light because there are always some signallers in the population that signal by the food. Thus, there was always enough blue light by the food to make it worth following. These results are currently under investigation.

In conclusion, we show that in order for honest communication to evolve, either genetic relatedness or group-level selection is needed. Furthermore, we show that the use of realistic models can lead to dynamics (such as the stability of deceptive communication) that one would not observe in simplified mathematical or game-theoretical models.
Four conditions tested Performance comparison with and without communication
Fig. 1. Left: Four conditions tested in our experiments. Right: Comparison of mean performance with and without communication in the four cases.

Food signalling strategy  Poison signalling strategy
Fig. 2. Left: Evolved food signalling strategy. Right: Evolved poison signalling strategy.


Press coverage

Previous projects and collaborators

Previous collaborators

Previous student projects

  • Vincent Porchet took part in the design of the Enki simulator, the s-bot implementation and the development of the sound pre-processing for the s-bot.
  • Kevin Frugier and Matthieu Bontemps ported some of our experiments from simulation to the s-bot robot.
  • Alban Laflaquière designed a similar system to transfer evolved controllers from simulation to the e-puck robots.
  • Alban also worked on a project to explore the evolution of quorum-sensing bacteria using a robotic model.

Related links

Cartoon on child language learning
Published: 29.05.06      Last update: 02.03.09