Evolving the “Feeling” of Time Through Sensory-Motor Coordination: A Robot Based Model

Tuci, Elio; Trianni, Vito; Dorigo, Marco

doi:10.1007/978-3-540-30217-9_101

Cited by 5 publications

(8 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nolfi (2002) and Croon et al (2004) investigate a task in which a robot is evolved to recognise whether it is located in one room rather than another, when navigating a maze. Tuci et al (2004Tuci et al ( ,2005 report an intriguing investigation in which the evolved control system of a simulated Khepera robot is shown to be able to change its behaviour as a result of its experience of the environment. More specifically, the robot is able to discriminate between the equivalents of productive and unproductive foraging situations.…”

Section: Separating the Various Swarm Areasmentioning

confidence: 99%

Swarm robotics and minimalism

Sharkey

2007

Connection Science

View full text Add to dashboard Cite

Swarm Robotics (SR) is closely related to Swarm Intelligence, and both were initially inspired by studies of social insects. Their guiding principles are based on their biological inspiration and take the form of an emphasis on decentralised local control and communication. Earlier studies went a step further in emphasising the use of simple reactive robots that only communicate indirectly through the environment. More recently SR studies have moved beyond these constraints to explore the use of non-reactive robots that communicate directly, and that can learn and represent their environment. There is no clear agreement in the literature about how far such extensions of the original principles could go. Should there be any limitations on the individual abilities of the robots used in SR studies? Should knowledge of the capabilities of social insects lead to constraints on the capabilities of individual robots in SR studies? There is a lack of explicit discussion of such questions and researchers have adopted a variety of constraints for a variety of reasons. A simple taxonomy of swarm robotics is presented here with the aim of addressing and clarifying these questions. The taxonomy distinguishes sub-areas of SR based on the emphases and justifications for minimalism and individual simplicity.

show abstract

Section: Separating the Various Swarm Areasmentioning

confidence: 99%

Swarm robotics and minimalism

Sharkey

2007

Connection Science

View full text Add to dashboard Cite

show abstract

“…On the explicit modelling side, a most simple approach is to retain all sensory information in the last T time steps, as done in classifier systems [3]; a more elaborate and demanding approach proceeds by encoding every possible situation of the robot in the search space [4]. Implicit memory modelling mostly proceeds by representing robotic controllers as recurrent neural nets (NN) [5], [6], [7], allegedly coding the memory of the previous time steps within the neuron states.…”

Section: Introductionmentioning

confidence: 99%

Memory-enhanced Evolutionary Robotics: The Echo State Network Approach

Hartland

Bredèche

Sebag

2009

2009 IEEE Congress on Evolutionary Computation

View full text Add to dashboard Cite

Interested in Evolutionary Robotics, this paper focuses on the acquisition and exploitation of memory skills. The targeted task is a well-studied benchmark problem, the Tolman maze, requiring in principle the robotic controller to feature some (limited) counting abilities. An elaborate experimental setting is used to enforce the controller generality and prevent opportunistic evolution from mimicking deliberative skills through smart reactive heuristics. The paper compares the prominent NEAT approach, achieving the non-parametric optimization of Neural Nets, with the evolutionary optimization of Echo State Networks, pertaining to the recent field of Reservoir Computing. While both search spaces offer a sufficient expressivity and enable the modelling of complex dynamic systems, the latter one is amenable to robust parametric, linear optimization with Covariance Matrix Adaptation-Evolution Strategies.

show abstract

“…It is notable that some "reactive" robots employ internal state to deal with perceptual aliasing and their control systems are not actually purely reactive [3]. Recently some researchers studied recurrent neural networks and plastic mechanism to solve real-world robotic problems [21,8,5]. It has been shown that the dynamic property in the neural networks can handle perceptually aliased problems without difficulty.…”

Section: Introductionmentioning

confidence: 99%

A quantitative analysis of memory requirement and generalization performance for robotic tasks

Kim

2007

Proceedings of the 9th Annual Conference on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

In autonomous agent systems, memory is an important element to handle agent behaviors appropriately. We present the analysis of memory requirements for robotic tasks including wall following and corridor following. The robotic tasks are simulated with sensor modeling and motor actions in noisy environments. In this paper, control structures are based on finite state machines for memory-based controllers, and we use the evolutionary multiobjective optimization approach with two objectives, behavior performance and memory size. For each task, a quantitative approach to estimate internal states with a different number of sensors is applied and the best controllers are evaluated in several test environments to examine their generalization characteristics and efficiency. Finite state machines with a hierarchy of memory are also compared with feedforward neural networks for the behavior performance.

show abstract

Evolving the “Feeling” of Time Through Sensory-Motor Coordination: A Robot Based Model

Cited by 5 publications

References 5 publications

Swarm robotics and minimalism

Swarm robotics and minimalism

Memory-enhanced Evolutionary Robotics: The Echo State Network Approach

A quantitative analysis of memory requirement and generalization performance for robotic tasks

Contact Info

Product

Resources

About