The goal of this project is to develop control strategies and neuromorphic chips for autonomous microflyers capable of navigating in confined or cluttered areas such as houses or small built environments using vision as main source of information.

Flying in such environements implies a number of challenges that are not found in high-altitude, GPS-based, unmanned aerial vehicles (UAVs). These include small size and slow speed for maneuverability, light weight to stay airborne, low-consumption electronics, and smart sensing and control. We believe that neuromorphic vision chips and bio-inspired control strategies are very promising methods to solve this challenge.

The project is articulated along three, tightly integrated, research directions:

Mechatronics of indoor microflyers (Adam Klaptocz, EPFL);
Neuromorphic vision chips (Rico Möckel, INI);
Insect-inspired flight control strategies (Antoine Beyeler, EPFL).
We plan to take inspiration from flying insects both for the design of the vision chips and for the choice of control architectures. Instead, for the design of the microflyers, we intend to develop innovative solutions and improvements over existing micro-helicopter and micro-airplanes.

Our final goal is to better understand the minimal set of mechanisms and strategies required to fly in confined environments by testing theoretical and neuro-physiological models in our microflyers.

A 10-gram microflyer that flies autonomously in a 7x6m test arena
The purpose of this ongoing experiment is to demonstrate autonomous steering of a 10-gram microflyer (the MC2) in a square room with different kind of textures on the walls (the Holodeck). This will first be achieved with conventional linear cameras before migrating towards aVLSI sensors.

The synthesis and reverse engineering of analog networks (see sidebar) are recognized as knowledge-intensive activities, where few systematic techniques exist. Given the importance and pervasiveness of analog networks, there is a founded interest in the development of automatic techniques capable of handling both problems. Evolutionary methods appear as one of the most promising approaches for the fulfillment of this objective.

Analog Genetic Encoding (AGE) is a new way to represent and evolve analog networks. The genetic representation of Analog Genetic Encoding is inspired by the working of biological genetic regulatory networks (GRNs). Like genetic regulatory networks, Analog Genetic Encoding uses an implicit representation of the interaction between the devices that form the network. This results in a genome that is compact and very tolerant of genome reorganizations, thus permitting the application of genetic operators that go beyond the simple operators of mutation and crossover that are typically used in genetic algorithms. In particular, Analog Genetic Encoding permits the application of operators of duplication, deletion, and transpositions of fragments of genome, which are recognized as fundamental for the evolution and complexification of biological organisms. The resulting evolutionary system displays state-of-the-art performance in the evolutionary synthesis and reverse engineering of analog networks.

The aim of this project is to address questions on the emergence and evolution of communication in groups of social animals by using evolutionary robotics to build societies of autonomous robots that evolve a communication system to solve a particular survival task collectively. Our work is part of the ECAgents project, sponsored by the Future and Emerging Technologies program of the European Community (IST-1940), whose aim is to develop embodied agents capable of interacting with the physical world, including other agents and humans through the use of communication.

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

• The pre-requisites for communication
• Conditions for the emergence of communication
• The transition to symbolic communication
• Social learning and the evolution of communication

Method
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.

Robots
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.

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.

Eyebots are autonomous flying robots with powerful sensing and communication abilities for search, monitoring, and pathfinding in built environments. Eyebots operate in swarm formation, as honeybees do, to efficiently explore built environments, locate predefined targets, and guide other robots or humans.

Eyebots are part of the Swarmanoid, a European research project aimed at developing an heterogeneous swarm of wheeled, climbing, and flying robots that can carry out tasks normally assigned to humanoid robots. Within the Swarmanoid, Eyebots serve the role of eyes and guide other robots with simpler sensing abilities.

Eyebots can also be deployed on their own in built environments to locate humans who may need help, suspicious objects, or traces of dangerous chemicals. Their programmability, combined with individual learning and swarm intelligence, makes them rapidly adaptable to several types of situations that may pose a danger for humans.

Eyebots are currently under development at LIS, EPFL. We will post additional information on this site as soon as technical documents will be available for public disclosure.

Gliding flight is powerful -to overcome obstacles and travel from A to B.

It can be applied in miniature robotics as a very versatile and easy to use locomotion method. In this project we aim at developing a palm sized Microglider that possesses the ability to deploy from ground or walls, to then open its wings, recover from almost every position in mid-air and perform subsequent goal directed gliding.

A potential source of inspiration on how to accomplish this task efficiently is nature. In the animal kingdom, many small animals are able to get into the air by jumping, fast running or by dropping down from trees. Once air-borne, they recover and stabilize passively or actively and perform goal directed aerial descent (e.g. gliding frogs, flying geckos, gliding lizards, locusts, crickets, flying squirrels, gliding fish, gliding ants etc.). These animals do not use steady state gliding, but change their velocity and angle of attack dynamically during flight to optimize the trajectory in order to increase the gliding ratio or land on a spot. The same principles may be advantageous for small aerial robots as well.

The critical issues on the path towards the realization of an efficient deploying Microglider at this scale are (i) the trade-off between passive stability, maneuverability and maximal gliding ratio, (ii) the low Reynolds number (<10'000) that leads to increased influence of boundary layer effects and renders the applicability of the conventional and well known large scale aerodynamics impossible and (iii) the control of the unsteady dynamics during recovery and flight.

The work in progress addresses these aspects. Embedded mechanisms for autonomous deployment from ground or walls into the air will be considered at the next stage.

Fatigue is a major source of stress and accidents in today's world, but there are no objective ways of monitoring and preventing the build-up of fatigue.

Sleep and wake periods are major factors, but not the only ones, that contribute to regulate the onset of fatigue. In this project, we start by developing a non-intrusive, wearable device for monitoring sleep and wake phases.

Since body signals related to sleep and wake are different from person to person, our device incorporates learning technologies adapted from our work on autonomous robotics. This allows the device to self-tune to the user.

The output of the sleep/wake device will then be incorporated into a model of fatigue that takes into account also other body signals and can adapt to the style and physiology of the user.

A version of the sleep/wake device will be tested within the framework of Solar Impulse, where the pilot has to be alert during the entire flight, which can take up to five days and nights. Our device can be used to predict the pilots fatigue and to calculate his optimal break times, always taking into account the mission status.

We aim at designing a swarm of Micro Air Vehcles (SMAVs) capable of autonomously establishing emergency wireless networks (SMAVNETs) between multiple ground-users in a disaster area. The SMAVNET is required to be rapidly deployable in any environment following the paradigms of swarm robotics. MAVs are designed to be minimal (low-cost, light-weight, simple electronics) and do not use position sensors (cameras, GPS), which are dependent on the environment, expensive in terms of energy, cost, size and weight or unusable at large ranges. Rather than that, agents rely on local communication with immediate neighbors and proprioceptive sensors which provide heading, speed, altitude and angular velocities.

Such SMAVNETs could replace damaged, inexistent or congested networks and can play an important role in disaster mitigation. The aerial nature of the system is interesting in that it allows for line-of-sight transmissions between MAVs, which is more energy-efficient than communication in cluttered environments at ground level. Furthermore, MAVs can fly over difficult terrain such as flooded areas or debris.

In the scope of the SMAVNET project we are developing a generic MAV platform usable in most aerial robotic applications. The construction of 10 to 20 of these MAVs will serve as a demonstration platform for developed SMAVNET swarm algorithms.