In industry 4.0, Machine Learning coupled with sensors monitoring leverages new ways to optimise maintenance strategies.In a predictive maintenance case, failure diagnoses are an excellent way to prevent any breakdowns.Up to now, failure diagnoses are focused on the classification of only one failure among many (multi-label classification), even if multiple failures can occur simultaneously.This study proposes an extension to classify simultaneous failures with the most popular classification methods such as random forests or artificial neural networks.Validated on a public predictive maintenance dataset, our methodology achieved classification with equal or best accuracy compared to multi-label classification.

To act cohesively as a group, robot swarms must be able to make decisions collectively. Collective decision-making refers to a process in which once a group decision is reached, it cannot be attributed to any single individual. Although extensive research has been conducted in swarm robotics using hand-coded design techniques to develop individual mechanisms for collective decision-making, the proposed mechanisms are generally limited in terms of robustness, scalability, and adaptability. In this paper, we employ evolutionary computation techniques to synthesise neural network-based decision-modules underpinning the individual opinion selection in robots. We describe the group dynamics underlying the decision process that leads to consensus in a three-option perceptual discrimination task. We test the robustness, scalability and adaptability of the decision-module in a variety of conditions. We show that the decision-making mechanisms underpinned by the evolved decision-module are more effective in supporting the collective decision-making process than the hand-coded voter and majority models, both in terms of accuracy and with respect to time to convergence to consensus.

In swarm robotics, random walks have proven to be efficient behaviours to explore unknown environments. By adapting the parameters of the random walk to environmental and social contingencies, it is possible to obtain interesting collective behaviours. In this paper, we introduce two novel aggregation behaviours based on different parameterisations of random walks tuned through numerical optimisation. Cue-based aggregation allows the swarm to reach the centre of an arena relying only on local discrete sampling, but does not guarantee the formation of a dense cluster. Neighbour-based aggregation instead allows the swarm to cluster in a single location based on the local detection of neighbours, but ignores the environmental cue. We then investigate a heterogeneous swarm made up of the two robot types. Results show that a trade-off can be found in terms of robot proportions to achieve cue-based aggregation while keeping the majority of the swarm in a single dense cluster.

A swarm of robots can collectively select an option among the available alternatives offered by the environment through a process known as collective decision-making. This process is characterised by the fact that once a decision is made by the group, it can not be attributed to any of its group members. In the large majority of the swarm robotics literature, only a few types of mechanisms have been used to allow the robots of a swarm to make collective decisions. Namely, the mechanisms referred to as the Voter and the Majority model or variations of these two models. Recently, an alternative solution has been proposed which is based on the use of dynamical neural networks as individual decision-making mechanisms. This alternative solution proved effective in a perceptual discrimination task. In this paper, we carry out extensive comparative tests that quantitatively evaluate, on a perceptual discrimination task, the Voter model and the dynamic neural network model on a variety of operating conditions and for swarms that differ in their size. The results of our study clearly indicate that the performances of a swarm employing dynamical neural networks as the decision-making mechanism are more robust, more adaptable to a dynamic environment, and more scalable to a larger swarm size than the performances of a swarm employing the Voter model as the decision-making mechanism. We account for these performance differences with an analysis of the two models.

In collective decision-making, individuals in a swarm reach consensus on a decision using only local interactions without any centralized control. In the context of the best-of-n problem - characterized by n discrete alternatives - it has been shown that consensus to the best option can be reached if individuals disseminate that option more than the other options. Besides being used as a mechanism to modulate positive feedback, long dissemination times could potentially also be used in an adversarial way, whereby adversarial swarms could infiltrate the system and propagate bad decisions using aggressive dissemination strategies. Motivated by the above scenario, in this paper we propose a bio-inspired defence strategy that allows the swarm to be resilient against options that can be disseminated for longer times. This strategy mainly consists in reducing the mobility of the agents that are associated to options disseminated for a shorter amount of time, allowing the swarm to converge to this option. We study the effectiveness of this strategy using two classical decision mechanisms, the voter model and the majority rule, showing that the majority rule is necessary in our setting for this strategy to work. The strategy has also been validated on a real Kilobots experiment.

Abstract The autonomous coordinated flying for groups of unmanned aerial vehicles that maximise network coverage to mobile ground-based units by efficiently utilising the available on-board power is a complex problem. Their coordination involves the fulfilment of multiple objectives that are directly dependent on dynamic, unpredictable and uncontrollable phenomena. In this paper, two systems are presented and compared based on their ability to reposition fixed-wing unmanned aerial vehicles to maintain a useful airborne wireless network topology. Genetic algorithms and non-cooperative games are employed for the generation of optimal flying solutions. The two methods consider realistic kinematics for hydrocarbon-powered medium-altitude, long-endurance aircrafts. Coupled with a communication model that addresses environmental conditions, they optimise flying to maximising the number of supported ground-based units. Results of large-scale scenarios highlight the ability of genetic algorithms to evolve flexible sets of manoeuvres that keep the flying vehicles separated and provide optimal solutions over shorter settling times. In comparison, game theory is found to identify strategies of predefined manoeuvres that maximise coverage but require more time to converge.

Classic opinion dissemination models such as Majority model, and Voter model are particularly important in swarms robotics because they regulate the interactions among the agents while the swarm is engaged in collective decision-making processes requiring the consensus of the large majority or the unanimity of the group's members. In this paper, we compare the effectiveness of three different opinion dissemination models in a specific scenario where consensus is searched on an opinion disseminated within the swarm by a small number of legitimate agents. The task of the legitimate agents is hindered by few adversarial agents which disseminate within the swarm an “invalid” message. This scenario is meant to model a data communication manipulation attack in a swarm of robots. By comparing the dissemination models in different experimental conditions, the results of our study inform us on which model is more effective in supporting the case of legitimate agents, while reducing the disruptive effects of the data communication manipulation attack.

This paper illustrates a method to design dynamic neural network controllers to allow a robot to autonomously navigate roads based on color perception. The neuro-controller moves the robot by setting the speed of the wheels and adjusts the robot visual systems by setting the value of three parameters that determine how much of the red, green and blue components of the RGB camera images contribute in generating the network input vector. Results show that the best evolved controller can successfully drive a real robot in environments with color characteristics never encountered during evolution. Moreover, we show that the dynamic color perception abilities are based on complex patterns of activation of the three color parameters. These patterns are generated by evolved neural mechanisms that successfully adapt the robot perceptual system to the color characteristics of the different visual scenes.

Active perception refers to a theoretical approach to the study of perception grounded on the idea that perceiving is a way of acting, rather than a process whereby the brain constructs an internal representation of the world. The operational principles of active perception can be effectively tested by building robot-based models in which the relationship between perceptual categories and the body-environment interactions can be experimentally manipulated. In this paper, we study the mechanisms of tactile perception in a task in which a neuro-controlled anthropomorphic robotic arm, equipped with coarse-grained tactile sensors, is required to perceptually categorize spherical and ellipsoid objects. We show that best individuals, synthesized by artificial evolution techniques, develop a close to optimal ability to discriminate the shape of the objects as well as an ability to generalize their skill in new circumstances. The results show that the agents solve the categorization task in an effective and robust way by self-selecting the required information through action and by integrating experienced sensory-motor states over time.

Active perception refers to a theoretical approach to the study of perception grounded on the idea that perceiving is a way of acting, rather than a cognitive process whereby the brain constructs an internal representation of the world. The operational principles of active perception can be effectively tested by building robot-based models in which the relationship between perceptual categories and the body-environment interactions can be experimentally manipulated. In this paper, we study the mechanisms of tactile perception in a task in which a neuro-controlled anthropomorphic robotic arm, equipped with coarse-grained tactile sensors, is required to perceptually discriminate between spherical and ellipsoid objects. The results of this work demonstrate that evolved continuous time nonlinear neural controllers can bring forth strategies to allow the arm to effectively solve the discrimination task.

This research work illustrates an approach to the design of controllers for self-assembling robots in which the self-assembly is initiated and regulated by perceptual cues that are brought forth by the physical robots through their dynamical interactions. More specifically, we present a homogeneous control system that can achieve assembly between two modules (two fully autonomous robots) of a mobile self-reconfigurable system without a priori introduced behavioral or morphological heterogeneities. The controllers are dynamic neural networks evolved in simulation that directly control all the actuators of the two robots. The neurocontrollers cause the dynamic specialization of the robots by allocating roles between them based solely on their interaction. We show that the best evolved controller proves to be successful when tested on a real hardware platform, the swarm-bot. The performance achieved is similar to the one achieved by existing modular or behavior-based approaches, also due to the effect of an emergent recovery mechanism that was neither explicitly rewarded by the fitness function, nor observed during the evolutionary simulation. Our results suggest that direct access to the orientations or intentions of the other agents is not a necessary condition for robot coordination: Our robots coordinate without direct or explicit communication, contrary to what is assumed by most research works in collective robotics. This work also contributes to strengthening the evidence that evolutionary robotics is a design methodology that can tackle real-world tasks demanding fine sensory-motor coordination.

This article describes a simulation model in which artificial evolution is used to design homogeneous control structures and adaptive communication protocols for a group of three autonomous simulated robots. The agents are required to cooperate in order to approach a light source while avoiding collisions. The robots are morphologically different: Two of them are equipped with infrared sensors, one with light sensors. Thus, the two morphologically identical robots should take care of obstacle avoidance; the other one should take care of phototaxis. Since all of the agents can emit and perceive sound, the group's coordination of actions is based on acoustic communication. The results of this study are a proof of concept: They show that dynamic artificial neural networks can be successfully synthesized by artificial evolution to design the neural mechanisms required to underpin the behavioral strategies and adaptive communication capabilities demanded by this task. Postevaluation analyses unveil operational aspects of the best evolved behavior. Our results suggest that the building blocks and the evolutionary machinery detailed in the article should be considered in future research work dealing with the design of homogeneous controllers for groups of heterogeneous cooperating and communicating robots.

In biology/psychology, the capability of natural organisms to learn from the observation/interaction with conspecifics is referred to as social learning. Roboticists have recently developed an interest in social learning, since it might represent an effective strategy to enhance the adaptivity of a team of autonomous robots. In this study, we show that a methodological approach based on artifcial neural networks shaped by evolutionary computation techniques can be successfully employed to synthesise the individual and social learning mechanisms for robots required to learn a desired action (i.e. phototaxis or antiphototaxis).

We present a first attempt to accomplish a simple object manipulation task using the self-reconfigurable robotic system swarm-bot. The number of modular entities involved, their global shape or size and their internal structure are not pre-determined, but result from a self-organized process in which the modules autonomously grasp each other and/or an object. The modules are autonomous in perception, control, action, and power. We present quantitative results, obtained with six physical modules, that confirm the utility of self-assembling robots in a concrete task

This article illustrates the methods and results of two sets of experiments in which a group of mobile robots, calleds-bots, are required to physically connect to each other, that is, to self-assemble, to cope with environmental conditions that prevent them from carrying out their task individually. The first set of experiments is a pioneering study on the utility of self-assembling robots to address relatively complex scenarios, such as cooperative object transport. The results of our work suggest that the s-bots possess hardware characteristics which facilitate the design of control mechanisms for autonomous self-assembly. The control architecture we developed proved particularly successful in guiding the robots engaged in the cooperative transport task. However, the results also showed that some features of the robots' controllers had a disruptive effect on their performances. The second set of experiments is an attempt to enhance the adaptiveness of our multi-robot system. In particular, we aim to synthesise an integrated (i.e., not-modular) decision-making mechanism which allows the s-bot to autonomously decide whether or not environmental contingencies require self-assembly. The results show that it is possible to synthesize, by using evolutionary computation techniques, artificial neural networks that integrate both the mechanisms for sensory-motor coordination and for decision making required by the robots in the context of self-assembly.

We survey developments in artificial neural networks, in behavior-based robotics, and in evolutionary algorithms that set the stage for evolutionary robotics (ER) in the 1990s. We examine the motivations for using ER as a scientific tool for studying minimal models of cognition, with the advantage of being capable of generating integrated sensorimotor systems with minimal (or controllable) prejudices. These systems must act as a whole in close coupling with their environments, which is an essential aspect of real cognition that is often either bypassed or modeled poorly in other disciplines. We demonstrate with three example studies: homeostasis under visual inversion, the origins of learning, and the ontogenetic acquisition of entrainment.

In this paper, we aim to design decision-making mechanisms for a simulated Khepera robot equipped with simple sensors, which integrates over time its perceptual experience in order to initiate a simple signalling response.Contrary to other previous similar studies, in this work the decision-making is uniquely controlled by the time-dependent structures of the agent controller, which in turn are tightly linked to the mechanisms for sensorimotor coordination.The results of this work show that a single dynamic neural network, shaped by evolution, makes an autonomous agent capable of 'feeling' time through the flow of sensations determined by its actions.Further analysis of the evolved solutions reveals the nature of the selective pressures that facilitate the evolution of fully discriminating and signalling agents.Moreover, we show that, by simply working on the nature of the fitness function, it is possible to bring forth discrimination mechanisms that generalize to conditions never encountered during evolution.

Research in the field of evolutionary robotics has begun to investigate the evolution of learning controllers for autonomous robots. Research in this area has achieved promising results, but research to date has focussed on the evolution of neural networks incorporating synaptic plasticity. There has been little investigation of possible alternatives, although the importance of exploring such alternatives is recognised. This paper describes a first step towards addressing this issue. Using networks with fixed synaptic weights and 'leaky integrator' neurons, we evolve robot controllers capable of learning and thus exploiting regularities occurring within their environment.