This paper combines the center-crossing condition in artificial neural networks that incorporate synaptic delays in their connections and which act as Central Pattern Generators (CPGs) for biped controllers. Recurrent...
详细信息
ISBN:
(纸本)9781479904549;9781479904532
This paper combines the center-crossing condition in artificial neural networks that incorporate synaptic delays in their connections and which act as Central Pattern Generators (CPGs) for biped controllers. Recurrent synaptic delay based neural networks allow greater time reasoning capabilities in the neural controllers, outperforming the results of continuous time recurrent neural networks, the neural model most used as CPG for biped robot locomotion related behaviors. Simulated evolution is used to automatically obtain neural controllers for walking behaviors, showing the capabilities of the synaptic delay based neural networks for the temporal coordination of the biped joints in difficult surfaces.
Heterogeneity is present in many collective systems found in nature and considered fundamental for effective task execution in several complex, real-world scenarios. evolutionary computation has the potential to autom...
详细信息
ISBN:
(纸本)9781450327381
Heterogeneity is present in many collective systems found in nature and considered fundamental for effective task execution in several complex, real-world scenarios. evolutionary computation has the potential to automate the design of multirobot systems, but to date, it has mostly been applied to the design of homogeneous systems. We have recently demonstrated that novelty search can overcome deception and bootstrapping issues in the evolution of homogeneous robot swarms. In this research, we study how evolutionary techniques based on behavioural diversity (such as novelty search) can contribute to the evolution of heterogeneous multirobot systems. The results obtained so far show that novelty search can overcome open issues in the cooperative coevolution of multiagent systems, and lead to more effective and diverse solutions.
How to drive a learning process towards the emergence of a memory? It is hypothesized here that a reward function which evaluates the fulfillment of a task requiring memory does not necessarily reward the stepping sto...
详细信息
ISBN:
(纸本)9781450311786
How to drive a learning process towards the emergence of a memory? It is hypothesized here that a reward function which evaluates the fulfillment of a task requiring memory does not necessarily reward the stepping stones to this cognitive ability. This question is studied from an evolutionary robotics perspective. Both structure and parameters of a neural network supposed to exhibit a memory are generated through an evolutionary search. Results show that selective pressures driving the evolutionary search are of critical importance. We further hypothesize that one feature of controllers with a memory is their ability to exhibit consistent behaviors over different contexts. To validate this hypothesis, a new fitness objective rewarding behavior consistency in different contexts is introduced and tested on a T-maze ER task - a task involving both navigation and working memory. The efficiency of the fitness objective is studied, as well as its effects on the overall performance and generalization ability of the controller. Results show that it is complementary to a behavioral diversity objective, thus leading to improved results when using both selection pressures.
With the recent advances of modular robots and low-cost manipulators, the evolution of robots, including morphologies and controllers, has become possible to perform in a physical setup without using any simulators. I...
详细信息
ISBN:
(纸本)9783031024627;9783031024610
With the recent advances of modular robots and low-cost manipulators, the evolution of robots, including morphologies and controllers, has become possible to perform in a physical setup without using any simulators. In this scenario, the evolution cannot be parallelized and the wall time becomes a scarce resource that should be used wisely. This paper analyses different algorithms by using the wall time as a stopping criterion for evolution, and it takes into account that wall time depends on the evaluation time plus the time to assemble and disassemble robots before and after an evaluation. The experiments have been performed in simulation, but the evaluation and assembly time have been carefully modelled from previous hardware experiments. Results suggest that (i) genetic algorithms are severely penalized, (ii) genetic algorithms can be improved by performing several evaluations of controllers for each morphology, and that (iii) evolutionary strategies that can chain several evaluations of robots with close morphologies can outperform other evolutionary algorithms. This finding is not surprising, but to the best of our knowledge previous attempts to evolve modular robots in hardware have not employed evolutionary strategies.
The landscape of isolated areas has been changed due to human intervention to support vehicular transport, however, this is a hectic job, therefore, if our vehicles are morphed to mimic nature, the landscape would not...
详细信息
ISBN:
(纸本)9783030345006;9783030344993
The landscape of isolated areas has been changed due to human intervention to support vehicular transport, however, this is a hectic job, therefore, if our vehicles are morphed to mimic nature, the landscape would not need to be changed. Robots and vehicles inspired from nature are very hard to control because of multiple number of actuators. Manual methods (such as programming individual actuators to form a walking pattern) fall short because of the complexity. Therefore, an automated process that employs artificial intelligence (AI) to evolve locomotion gaits for quadrupedal robots is needed. AI has been used before as well;however, most of the AI implementations are only done in simulation without hardware execution. This article attempts to use genetic algorithms to evolve locomotion gaits that are later implemented on robots both via simulations and real implementation. The simulation is run for 200 generations and the best result is put into effect on a hardware robot. Our results show that the gait is successfully transferred;however, the results are not perfect and suffer from the reality gap. These results also help us conclude that gaits designed for a specific environment have a better chance of transferring than gaits that have been designed without taking into account the surface the robot walks on.
The tasks of odor detection, plume tracking and odor source localization constitute an important, yet complex, real world problem. One possible solution for them is based on the use of a group of mobile robots whose c...
详细信息
ISBN:
(纸本)9781509022557
The tasks of odor detection, plume tracking and odor source localization constitute an important, yet complex, real world problem. One possible solution for them is based on the use of a group of mobile robots whose controllers have to be defined. Artificial Neural Networks (ANN) have already been used as controllers, but the task of hand defining their topology and parameters can be very challenging and time consuming. In this paper, we propose an approach to evolve, rather than design, ANN-based controllers. Our approach relies on Genetic Programming (GP), a family of stochastic search procedures loosely inspired by the biological principles of Natural Selection and Genetics. We compare our approach with a classic one, inspired by the chemotaxis behavior of the E. coli bacteria. Our results show that this approach is able to outperform the chemotaxis in the experiments performed.
In this paper we report our experiments with the e-puck robots for developing a communication system using evolutionary robotics. In order to do the latter we follow the evolutionary approach by using Neural Networks ...
详细信息
ISBN:
(纸本)9783642253232;9783642253249
In this paper we report our experiments with the e-puck robots for developing a communication system using evolutionary robotics. In order to do the latter we follow the evolutionary approach by using Neural Networks and Genetic Algorithms. The robots develop a communication scheme for solving tasks like: locating food areas, avoiding obstacles, approaching light sources and locating sound-sources (other robots emitting sounds). Evorobot* and Webots simulators are used as tools for computing the evolutionary process and optimization of the weights of neural controllers. As a consequence, two different kinds of neural controllers emerge. On one hand, one controller is used for robot movement;on the other hand the second controller process sound signals.
This paper presents a Genetic Algorithm (GA) approach to evolving robot behaviours. We use fuzzy logic controllers (FLCs) to design robot behaviours. The antecedents of the FLCs are pre-designed, while their consequen...
详细信息
ISBN:
(纸本)0780378660
This paper presents a Genetic Algorithm (GA) approach to evolving robot behaviours. We use fuzzy logic controllers (FLCs) to design robot behaviours. The antecedents of the FLCs are pre-designed, while their consequences are learned using a GA. The Sony quadruped robots are used to evaluate proposed approaches in the robotic football domain. Two behaviours, ball-chasing and position-reaching, are studied and implemented. An embodied evolution scheme is adopted, by which the robot autonomously evolves its behaviours based on a layered control architecture. The results show that the robot behaviours can be automatically acquired through the GA-based learning of FLCs.
Enabling artificial agents to automatically learn complex, versatile and high-performing behaviors is a long-lasting challenge. This paper presents a step in this direction with hierarchical behavioral repertoires tha...
详细信息
ISBN:
(纸本)9781450356183
Enabling artificial agents to automatically learn complex, versatile and high-performing behaviors is a long-lasting challenge. This paper presents a step in this direction with hierarchical behavioral repertoires that stack several behavioral repertoires to generate sophisticated behaviors. Each repertoire of this architecture uses the lower repertoires to create complex behaviors as sequences of simpler ones, while only the lowest repertoire directly controls the agent's movements. This paper also introduces a novel approach to automatically define behavioral descriptors thanks to an unsupervised neural network that organizes the produced high-level behaviors. The experiments show that the proposed architecture enables a robot to learn how to draw digits in an unsupervised manner after having learned to draw lines and arcs. Compared to traditional behavioral repertoires, the proposed architecture reduces the dimensionality of the optimization problems by orders of magnitude and provides behaviors with a twice better fitness. More importantly, it enables the transfer of knowledge between robots: a hierarchical repertoire evolved for a robotic arm to draw digits can be transferred to a humanoid robot by simply changing the lowest layer of the hierarchy. This enables the humanoid to draw digits although it has never been trained for this task.
This paper summarizes the keynote I gave on the SEAMS 2020 conference. Noting the power of natural evolution that makes living systems extremely adaptive, I describe how artificial evolution can be employed to solve d...
详细信息
ISBN:
(纸本)9781450379625
This paper summarizes the keynote I gave on the SEAMS 2020 conference. Noting the power of natural evolution that makes living systems extremely adaptive, I describe how artificial evolution can be employed to solve design and optimization problems in software. Thereafter, I discuss the Evolution of Things, that is, the possibility of evolving physical artefacts and zoom in on a (r)evolutionary way of creating 'bodies' and 'brains' of robots for engineering and fundamental research.
暂无评论