Search

Evolutionary Robotics is a method for automatically generating artificial brains and morphologies of autonomous robots. This approach is useful both for investigating the design space of robotic applications and for testing scientific hypotheses of biological mechanisms and processes. In this chapter we provide an overview of methods and results of Evolutionary Robotics with robots of different shapes, di- mensions, and operation features. We consider both simulated and physical robots with special consideration to the transfer between the two worlds.

Cambridge, MA: MIT press.

Evolutionary robotics is the attempt to develop robots through a self- organized process based on artificial evolution. This approach stresses the importance of the study of systems that: (1) have a body and are situated in a physical environment, and (2) autonomously develop their own skills in close interaction with the environment. In this article we will briefly illustrate the method and the main idea, we will review the most significant contribution in this area and finally, we will discuss some of the contributions that this research area is making to the foundational debate in cognitive science.

In this paper we investigate a novel approach to the evolutionary development of autonomous situated agents based on the assumption that the neural mechanisms underlying ontogenetic learning are themselves developed and shaped by evolutionary process. A genetic algorithm is used to evolve neural structures that can be continuously modified during life according to the mechanisms specified in the genotype. The evolutionary process is carried out on a real mobile robot. The analysis of one of the best evolved individuals shows rapid development of stable behavior mediated by fast-changing synapses which are dynamically stable.

Artificial life IV : proceedings of the Fourth International Workshop on the Synthesis and Simulation of Living Systems, ISBN:978-0262521901

Artificial evolution of computer software (evolutionary neural networks, genetic programming, evolutionary fuzzy systems, etc.) has been shown to generate software that in many cases is more performant than that designed by engineers. Evolved software performs well under the same conditions used during evolutionary training. However, in situations where unpredictable change may affect normal operation, evolved systems often fail. In this paper we describe a new approach for evolving software that remains adaptive and is therefore very robust to unpredictable change after evolution. To illustrate the idea, we present the case of evolutionary robots that quickly and reliably adapt online to several types of new situations, including sensory, environmental, and mechanical change, while still performing their task. The core of the methodology consists of evolving the mechanisms of parameter adaptation instead of the parameters themselves. We shall conclude by showing how this methodology can be applied to a variety of other situations beyond robotics.

We describe an evolutionary approach to active vision systems for dynamic feature selection. After summarizing recent work on evolution of a simulated active retina for complex shape discrimination, we describe in detail experiments that extend this approach to an all-terrain mobile robot equipped with a mobile camera. We show that evolved robots are capable of selecting simple visual features and actively maintaining them on the same retinal position, which largely simplifies the "recognition'' task, in order to generate efficient navigation trajectories with an extremely simple neural control system. Analysis of evolved solutions indicates that robots develop a simple and yet very efficient version of edge detection and visual looming to detect obstacles and move away from them. Two evolved sensory-motor strategies are described, one where the mobile camera is actively used throughout the entire navigation and one where it is used only at the beginning to point towards relevant environmental features. The relationship between these two strategies are discussed in the context of the underlying visuo-motor mechanisms and of the evolutionary conditions.

Co-evolution of competitive species provides an interesting testbed to study the role of adaptive behavior because it provides unpredictable and dynamic environments. In this paper we experimentally investigate some arguments for the co-evolution of different adaptive protean behaviors in competing species of predators and preys. Both species are implemented as simulated mobile robots (Kheperas) with infrared proximity sensors, but the predator has an additional vision module whereas the prey has a maximum speed set to twice that of the predator. Different types of variability during life for neurocontrollers with the same architecture and genetic length are compared. It is shown that simple forms of proteanism affect co-evolutionary dynamics and that preys rather exploit noisy controllers to generate random trajectories, whereas predators benefit from directional-change controllers to improve pursuit behavior.

Indirect genotype to phenotype mappings in the form of developmental systems may allow better scalability to larger phenotypes in evolvable hardware. This report reviews developmental systems used in evolvable hardware and it proposes a new classifications based on hardware characteristics. It then describes a genetic encoding and developmental system called the morphogenetic system which has been designed for multi-cellular circuits. This morphogenetic system is inspired upon gene expression and cell differentiation but it focuses on efficient hardware implementation. An hardware implementation on the dynamically reconfigurable POEtic circuit is described. It uses serial arithmetics and time-multiplexing to minimize ressource use.

This paper is concerned with artificial evolution of neuro-controllers with adaptive synapses for autonomous mobile robots. The method consists of encoding on the genotype a set of local modification rules that synapses obey while the robot freely moves in the environment [2]. The synaptic weights are not encoded on the genotype. In the experiments presented here, a "behavior-based fitness" function gives reproductive advantage to robots that can solve a sequential task. The results show that evolutionary adaptive controllers solve the task much faster and better than evolutionary standard (non-adaptive) controllers, that the method scales up well to large architectures whereas standard controllers do not, and that evolved adaptive controllers are not trivial and cannot be reduced to a fixed-weight network.

This paper describes an evolutionary algorithm based on a statistical representation of populations of individuals. Experiments on robot navigation and on numerical fitness functions are presented in order to measure the performance of the algorithm compared to traditional genetic algorithms. Results show that the method is suitable for onboard online evolution because it requires low amount of memory resources. Furthermore, it allows for incremental evolution in dynamic environments in order to cope with complex tasks that require several evolutionary stages.

Evolutionary robotics is the attempt to develop robots through a self-organized process based on artificial evolution. This approach stresses the importance of the study of systems that have a body and that are situated in a physical environment, and which autonomously develop their own skills in close interaction with the environment. In this review we briefly illustrate the method and the main concept of evolutionary robotics, and examine the most significant contribution in this area. We also discuss some of the contributions that this research area is making to the foundational debate in cognitive science.

After reviewing current approaches in Evolutionary Robotics, we point to directions of research that re likely to bring interesting results in the future. e then address two crucial aspects for future developments of Evolutionary Robotics: choice of fitness functions and scalability to real-world situations. In the first case we suggest framework to describe fitness functions, choose them according to the situation constraints, and compare available experiments in the literature on evolutionary robotics. In the second case, we suggest way to make experimental results applicable to real- world situations by evolving online continuous adaptive controllers. We also give an overview of recent experimental results showing that the suggested approaches pro- duce qualitatively superior abilities, scale up to more complex architectures, smoothly transfer from simulations to real robots and across different robotic platforms, and autonomously adapt in few seconds to several sources of strong variability that were not included during the evolutionary run.

Evolutionary robotics is an interesting novel approach to shape the control system of autonomous robots. This explores issues related to re-adaptation in changed environments of a population of evolved individuals. Experimental studies are reported for genetic evolution of neurocontrollers that have developed the ability to perform homing navigation for battery recharge of miniature mobile robot. It is shown that re-adaptation to important changes in the environment is very rapid and does not disrupt previously acquired knowledge. The results are discussed in relation to the internal representation of the neurocontroller and to the variability within the population.

In this paper I explore the hypothesis that some of today's robots might possess a form of consciousness whose substrate is a mere algorithm. First, consciousness is defined within an evolutionary framework as awareness of one's own state in relation to the external environment. Then, the basic prerequisites for such conscious activity are discussed, namely embodiment, autonomy, and adaptation mechanisms. Artificial evolution, rather than evolutionary optimisation, is presented as a viable methodology to create conscious robots, accompanied by some examples of behaviours of artificially evolved robots. Finally, I argue that what might be problematic with the concept of robot consciousness is not the robot, but the notion of consciousness itself.

Service robotics, as it has been intended so far, views the accomplishment of a service mission mainly as the result of the action of a single robot. Swarm robotics tackles the very same problem from a different stance, \\ie as the result of a team effort of simple units. The project described here shows this particular approach. It defines first one simple unit (s-bot) capable of independently moving about on the ground and of dynamically establishing rigid or semi-rigid connections with other fellow units, and then it shows how a large group of them can, as a whole entity (swarm-bot), carry out a given task. Thanks to the ductility in assembling and forming its connections, a swarm-bot can readily cope with occasional failures of some components and promptly reshape the remaining swarm so as to replace the role of the failing units. Given such a plasticity, their possible applications is rather large ranging from harsh environment exploration to goods harvesting or goods transportation. At the moment, the project is at the stage of having defined a first simulating environment to be used both for the on-going hardware design and for the software control. The present paper describes this particular aspect of the project.

In the simplest scenario of two co-evolving populations in competition with each other, fitness progress is achieved at disadvantage of the other population's fitness. The everchanging fitness landscape caused by the competing species (named the "Red Queen effect") makes the system dynamics more complex, but it also provides a set of advantages with respect to single-population evolution. Here we present results from an experiment with two mobile robots, a predator equipped with vision and a much faster prey with simpler sensors. Without any effort in fitness design, a set of interesting behaviors emerged in relatively short time, such as obstacle avoidance, straight navigation, visual tracking, object discrimination (robot vs. wall), object following, and others. Although such experiments cannot yet be performed in real-time on populations of robots for technical reasons, the approach seems promising.

Co-evolution (i.e. the evolution of two or more competing populations with coupled fitness) has several features that may potentially enhance the power of adaptation of artificial evolution. In particular, as discussed by Dawkins and Krebs [3], competing populations may reciprocally drive one another to increasing levels of complexity by producing an evolutionary “arms race”. In this paper we will investigate the role of co-evolution in the context of evolutionary robotics. In particular, we will try to understand in what conditions co-evolution can lead to “arms races”. Moreover, we will show that in some cases artificial co-evolution has a higher adaptive power than simple evolution. Finally, by analyzing the dynamics of co-evolved populations, we will show that in some circumstances well adapted individuals would be better advised to adopt simple but easily modifiable strategies suited for the current competitor strategies rather than incorporate complex and general strategies that may be effective against a wide range of opposing counter-strategies.

We review recent experiments in evolutionary robotics carried out in dynamic environments and across different robotic platforms. We then introduce a new evolutionary approach where robots are evolved for their ability to adapt online. Several experiments show that this new approach is much faster, more powerful, and scalable that the traditional approach.

This paper addresses the trajectory tracking problem of a 2D caged flying robot in contact with a wall. To simplify the contact problem, the models are constructed on a vertical two-dimensional plane, and our objective is to let the quadrotor hover or move along the wall with arbitrary velocity and attitude. The control law is derived using the Lyapunov stability theory, applying backstepping techniques to achieve exponential stability under mild assumptions. To overcome the unknown friction force between robot and wall, we design estimators for the friction coefficient, which include a projection operator that provides an upper bound for the obtained estimates. Realistic simulation results are provided to validate the proposed methodology.

In this paper an evolutionary method consisting of encoding a set of local adaptation rules that synapses obey while a robot freely moves in the environment is compared to standard evolution of fixed-weight control networks. The results show ha evolutionary adaptive controllers can adapt online without additional evolutionary training to strong environmental changes where instead the performance of evolutionary fixed- weight controllers is significantly degraded. Two cases are described: transfer of evolved controllers from simulated to real robots and across different robotic platforms that vary in size, shape, and sensor response profile. In both cases evolved adaptive controllers autonomously and quickly adjust synaptic weights to successfully accomplish the task in the new conditions.

Evolutionary Robotics is a powerful method to generate efficient controllers with minimal human intervention, but its applicability to real-world problems remains a challenge because the method takes long time and it requires software simulations that do not necessarily transfer smoothly to physical robots. In this paper we describe a method that overcomes these limitations by evolving robots for the ability to adapt on-line in few seconds. Experiments show that this method requires less generations and smaller populations to evolve, that evolved robots adapt in a few seconds to unpredictable change - including transfers from simulations to physical robots - and display non-trivial behaviors. Robots evolved with this method can be dispatched to other planets and to our homes where they will autonomously and quickly adapt to the specific properties of their environments if and when necessary.