Search

This paper focuses on the collective cognition by robotic swarms. The robotic swarms expected to perform tasks that are beyond the capability of a single robot by collective behavior that emerge from local interactions, similar to biological swarms. However, the robotic swarms have to rely on the collective cognition more than biological swarms when considering the limitation in sensory capabilities and the cost of each robot. In this paper, we develop controllers for a robotic swarm to accomplish a foraging task that requires collective cognition. In this task, robots have to both collectively distinguish two objects, namely food and poison, and cooperatively transport food objects to the nest. We applied an evolutionary robotics approach with the covariance matrix adaptation evolution strategy to develop controllers for robotic swarms. The results of computer simulations show that collective cognition was successfully developed, which allows the robots to transport only food objects. In addition, we also perform experiments to examine the scalability and the flexibility of the developed controllers.

Redundancy in the number of robots is a fundamental feature of robotic swarms to confer robustness, flexibility, and scalability. However, robots tend to interfere with each other in a case, where multiple robots gather in a spatially limited environment. The aim of this paper is to understand how a robotic swarm develops an effective strategy to manage congestion. The controllers of the robots are obtained by an evolutionary robotics approach. The strategy of managing congestion is observed in the process of generating a collective path of robots visiting two landmarks alternately. The robotic swarm exhibits autonomous specialization that the robots traveling inside the path activate the LEDs, while the robots in the outer side deactivate them. We found that the congestion is regulated in an emergent way of autonomous specialization by the result of an artificial evolution.

Swarm robotics is a field in which multiple robots coordinate their collective behavior autonomously to accomplish a given task without any form of centralized control. In swarm robotics, task allocation refers to the behavior resulting in robots being dynamically distributed over different sub-tasks, which is often required for solving complex tasks. It has been well recognized that evolutionary robotics is a promising approach to the development of collective behaviors for robotic swarms. However, the artificial evolution often suffers from two issues—the bootstrapping problem and deception—especially when the underlying task is profoundly complex. In this study, we propose a two-step scheme consisting of task partitioning and autonomous task allocation to overcome these difficulties. We conduct computer simulation experiments where robotic swarms have to accomplish a complex collective foraging problem, and the results show that the proposed approach leads to perform more effectively than a conventional evolutionary robotics approach.

We extend an abstract agent-based swarming model based on the evolution of neural network controllers, to explore further the emergence of swarming. Our model is grounded in the ecological situation, in which agents can access some information from the environment about the resource location, but through a noisy channel. Swarming critically improves the efficiency of group foraging, by allowing agents to reach resource areas much more easily by correcting individual mistakes in group dynamics. As high levels of noise may make the emergence of collective behavior depend on a critical mass of agents, it is crucial to reach sufficient computing power to allow for the evolution of the whole set of dynamics in simulation. Since simulating neural controllers and information exchanges between agents are computationally intensive, to scale up simulations to model critical masses of individuals, the implementation requires careful optimization. We apply techniques from astrophysics known as treecodes to compute the signal propagation, and efficiently parallelize for multi-core architectures. Our results open up future research on signal-based emergent collective behavior as a valid collective strategy for uninformed search over a domain space.

International audience; This paper presents a behavioral ontogeny for artificial agents based on the interactive memorization of sensorimotor invariants. The agents are controlled by continuous timed recurrent neural networks (CTRNNs) which bind their sensors and motors within a dynamic system. The behavioral ontogenesis is based on a phylo- genetic approach: memorization occurs during the agent's lifetime and an evolutionary algorithm discovers CTRNN parameters. This shows that sensorimotor invariants can be durably modified through interaction with a guiding agent. After this phase has finished, agents are able to adopt new sensorimotor invariants relative to the environment with no further guidance. We obtained these kinds of behaviors for CTRNNs with 3-6 units, and this paper examines the functioning of those CTRNNs. For instance, they are able to internally simulate guidance when it is externally absent, in line with theories of simulation in neuroscience and the enactive field of cognitive science.

Lately, development in robotics for utilizing in both industry and home is in much progress. In this research, a group of robots is made to handle relatively complicated tasks. Cooperative action among robots is one of the research areas in robotics that is progressing remarkably well. Reinforcement learning is known as a common approach in robotics for deploying acquisition of action under dynamic environment. However, until recently, reinforcement learning is only applied to one agent problem. In multi-agent environment where plural robots exist, it was difficult to differentiate between learning of achievement of task and learning of performing cooperative action. This paper introduces a method of implementing reinforcement learning to induce cooperation among a group of robots where its task is to transport luggage of various weights to a destination. The general Q-learning method is used as a learning algorithm. Also, the switching of learning mode is proposed for reduction of learning time and learning area. Finally, grid world simulation is carried out to evaluate the proposed methods.

We show how a biologically inspired model of multicellular development combined with a simulated evolutionary process can be used to design the morphologies and controllers of soft-bodied virtual animats. An animat’s morphology is the result of a developmental process that starts from a single cell and goes through many cell divisions, during which cells interact via simple physical rules. Every cell contains the same genome, which encodes a gene regulatory network (GRN) controlling its behavior. After the developmental stage, locomotion emerges from the coordinated activity of the GRNs across the virtual robot body. Since cells act autonomously, the behavior of the animat is generated in a truly decentralized fashion. The movement of the animat is produced by the contraction and expansion of parts of the body, caused by the cells, and is simulated using a physics engine. Our system makes possible the evolution and development of animats that can run, swim, and actively navigate toward a target in a virtual environment.

Embodied evolution (EE) is a methodology in evolutionary robotics in which, without simulations on a host computer, real robots evolve on the basis of their interactions with the actual environment. However, when adopting EE, we had to accept robot behavior with a low fitness, especially in the early generations. This article introduces pre-evaluation into the EE framework for a biped robot in order to restrain the behavior of a robot of which the fitness is estimated to be low, especially falling down onto the ground. We provide a comparative discussion on the conventional simulate-and-transfer method, the original EE method, and the proposed one in terms of calculation time, cost of fitness evaluation, and cost of simulation or modeling based on the evaluation experiments. We believe that the EE framework with pre-evaluation is applicable to a wide variety of optimization tasks in which the cost or the risk of fitness evaluation is not negligible.

In this article we explore some of the issues currently facing researchers in the interface between the twin fields of Artificial Life and Robotics, and the challenges and potential synergy of these two areas in the creation of future robotic life forms. There are three strands of research which we feel will be of key importance in the possible development of future embodied artificial life forms. These are the areas of evolutionary robotics and evolutionary humanoid robotics in particular, probabilistic robotics for deliberation, and robot benchmarking with associated metrics and standards. We briefly explore each of these areas in turn, focusing on our current research in each field and what we see as the potential issues and challenges for the future.

In this paper, we present a new method based on multiobjective evolutionary algorithms to evolve low-complexity neural controllers for agents that have to perform multiple tasks simultaneously. In our method, each task and the structure of the neural controller are considered as separated objective functions. We compare the results of two different encoding schemes: (1) connectionist encoding, and (2) node-based encoding. The results show that multiobjective evolution can be successfully applied to generate low-complexity neural controllers. In addition, node-based encoding outperformed connectionist encoding in terms of agent performance and the robustness of the neural controller.

The field of evolutionary humanoid robotics is a branch of evolutionary robotics specifically dealing with the application of evolutionary principles to humanoid robot design. Previous studies demonstrated the possible future potential of this approach by evolving walking behaviors for simulated humanoid robots with up to 20 degrees of freedom. In this paper we examine further the evolutionary process by looking at the changes in diversity over time. We then investigate the effect of the immobilization of an individual joint or joints in the robot. The latter study may be of potential future use in prosthetic design. We also explore the possibility of the evolution of humanoid robots which can cope with different environmental conditions. These include reduced ground friction (ice) and modified gravitation (moon walking). We present initial results on the implementation of our simulated humanoid robots in hardware using the Bioloid robotic platform, using a model of this robot in order to evolve the desired motion patterns, for subsequent transfer to the real robot. We finish the article with a summary and brief discussion of future work.

This article describes a new approach for control systems for an autonomous mobile robot by using sandwiches of two different types of neural network. One is a neural network with competition and cooperation, and is used for recognizing sensor information where synaptic coupling are fixed. The second is a neural network with adaptive synaptic couplings corresponding to a genotype in a creature, and used for self-learning for the wheel controls. In a computer simulation model, we were successful in obtaining four types of robot with good performance when going along a wall. The model also showed robustness in a real environment.

Deploying autonomous robot teams instead of humans in hazardous search and rescue missions could provide immeasurable benefits. In such operations, rescue workers often face environments where information about the physical conditions is impossible to obtain, which not only hampers the efficiency and effectiveness of the effort, but also places the rescuers in life-threatening situations. These types of risk promote the potential for using robot search teams in place of humans. This article presents the design and implementation of controllers to provide robots with appropriate behavior. The effective utilization of genetic algorithms to evolve controllers for teams of homogeneous autonomous robots for area coverage in search and rescue missions is described, along with a presentation of a robotic simulation program which was designed and developed. The main objective of this study was to contribute to efforts which attempt to implement real-world robotic solutions for search and rescue missions.

The use and benefits of self-adaptive mutation operators are well known within evolutionary computing. In this paper, we begin by examining the use of self-adaptive mutation in learning classifier systems with the aim of improving their performance as controllers for autonomous mobile robots. We implement the operator in a zeroth level classifier system, and examine its performance in two animat environments. It is shown that although no significant increase in performance is seen over results presented in the literature using a fixed rate of mutation, the operator adapts to an appropriate rateregadless of the initial range. The same concept is then applied to the learning rate parameter, but results show that a modification must be made to produce stable/effective controllers. Finally, results from a fully self-adaptive system are presented, with marked benefits being found in a nonstationary environment.

Recently, evolutionary robotics (ER) has been attracting a lot of attention in the field of robotics artificial life and so on. ER approaches are expected to provide feasible methods to design controllers for autonomous mobile robots with less human intervention. However, most of the conventional studies in ER have been aiming at obtaining very simple (trivial) behaviors such as obstacle avoiding, wall following, and target approaching. To make the ER approach more fruitful, we should pay close attention to obtaining nontrivial behaviors. Based on these considerations, in this paper we propose a method for obtaining nontrivial behaviors using a developmental process with a carefully arranged grafting method. To verify the validity, we apply our idea to the construction of neural controllers that can cope with rough terrain.

Non-Cartesian robotics, which began with the introduction of subsumption architecture by Rodney Brooks, now encompasses a wide range of robotics that do not follow traditional cartesian principles in the running of a robot. The new field is sometimes called biorobotics as it draws its guiding principles from biology, physiology, behavioural sciences, genetics and theories of evolution, brain sciences, genetics and theories of evolution, brain sciences, ethology, psychology, and other related non-engineering disciplines. The difference in principles of operation, however, has roots deeper in the philosophical underpinnings of the way we view controlling artifacts and the concept of control itself when it is contrasted against the concept of autonomy. Realization of increasingly higher levels of autonomy is routinely demanded today not only in industry where most robotic applications occur, but also in areas closer to our daily life where a gradual but steady increase in service applications of robotics is observed. This paper introduces the concept of non-Cartesian robotics as an antithesis to conventional (Cartesian) robotics and describes various aspects of this new way of running a robotic system.

Artificial evolution as a design methodology allows the relaxation of many of the constraints that have held back conventional methods. It does not require a complete prior analysis and decomposition of the task to be tackled, as human designers require. However, this freedom comes at some cost; there are a whole new set of issues relating to evolution that must be considered. Standard genetic algorithms may not be appropriate for incremental evolution of robot controllers. Species adaptation genetic algorithms, (SAGA) have been developed to meet these special needs. The main cost of an evolutionary approach is the large number of trials that are required. Simulations-especially those involving vision in complex environments, or modeling detailed semiconductor physics—may not be adequate or practical. Examples of evolved robots will be discussed, including a specialized piece of equipment which allows a robot to be tested using simple vision in real time, and what is believed to be the first successful example of an evolved hardware controller for a robot.

Intelligent systems are required in knowledge engineering, computer science, mechatronics, and robotics. This article discusses machine (system) intelligence from the viewpoints of the learning and adaptation of living things. We then introduce computational intelligence, including neural networks, fuzzy systems, and genetic algorithms, and end with a discussion of recent work on computational intelligence in robotics.