Real-World Evolution of Robot Morphologies: A Proof of Concept

Jelisavcic, Milan; Carlo, Matteo De; Hupkes, Elte; Eustratiadis, Panagiotis; Orlowski, Jakub; Haasdijk, Evert; Auerbach, Joshua E.; Eiben, A. E.

doi:10.1162/artl_a_00231

Cited by 59 publications

(63 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Triangle of Life concept (discussed below) has been demonstrated with a simplified setup by Jelisavcic et al (2017), where physical robots were 3D-printed and handassembled, also using prefabricated organ, such as servo motors and Raspberry Pis. The ARE project will take this further through the automated Robot Fabricator, allowing larger numbers of individuals and a complete evolutionary system to be fabricated.…”

Section: Related Workmentioning

confidence: 99%

The ARE Robot Fabricator: How to (Re)produce Robots that Can Evolve in the Real World

Hale¹,

Buchanan²,

Winfield³

et al. 2019

The 2019 Conference on Artificial Life

Self Cite

View full text Add to dashboard Cite

The long term vision of the Autonomous Robot Evolution (ARE) project is to create an ecosystem of both virtual and physical robots with evolving brains and bodies. One of the major challenges for such a vision is the need to construct many unique individuals without prior knowledge of what designs evolution will produce. To this end, an autonomous robot fabrication system for evolutionary robotics, the Robot Fabricator, is introduced in this paper. Evolutionary algorithms can create robot designs without direct human interaction; the Robot Fabricator will extend this to create physical copies of these designs (phenotypes) without direct human interaction. The Robot Fabricator will receive genomes and produce populations of physical individuals that can then be evaluated, allowing this to form part of the evolutionary loop, so robotic evolution is not confined to simulation and the reality gap is minimised. In order to allow the production of robot bodies with the widest variety of shapes and functional parts, individuals will be produced through 3D printing, with prefabricated actuators and sensors autonomously attached in the positions determined by evolution. This paper presents details of the proposed physical system, including a proof-ofconcept demonstrator, and discusses the importance of considering the physical manufacture for evolutionary robotics.

show abstract

Section: Related Workmentioning

confidence: 99%

The ARE Robot Fabricator: How to (Re)produce Robots that Can Evolve in the Real World

Hale¹,

Buchanan²,

Winfield³

et al. 2019

The 2019 Conference on Artificial Life

Self Cite

View full text Add to dashboard Cite

show abstract

“…This means that the evolutionary operator's mutation and crossover are applied to the genotypes (to the code) and not to the phenotypes (to the robotic organisms). The first stage in the ToL is the creation of a new robotic organism in a so-called Production Center [19]. This stage starts with a new piece of genetic code that is created by mutating or recombining existing pieces of code (of the robot parents) and ends with the delivery of a new robot.…”

Section: System Architecture: the Triangle Of Lifementioning

confidence: 99%

Revolve: A Versatile Simulator for Online Robot Evolution

Hupkes

Jelisavcic

Eiben

2018

Applications of Evolutionary Computation

Self Cite

View full text Add to dashboard Cite

Developing robotic systems that can evolve in real-time and real-space is a long term objective with technological as well as algorithmic milestones on the road. Technological prerequisites include advanced 3D-printing, automated assembly, and robust sensors and actuators. The necessary evolutionary mechanisms need not wait for these, they can be developed and investigated in simulations. In this paper, we present a system to simulate online evolution of constructible robots, where (1) the population members (robots) concurrently exist and evolve their morphologies and controllers, (2) all robots can be physically constructed. Experiments with this simulator provide us with insights into differences of using online and offline evolutionary setups.

show abstract

“…We design the modular robots using a subset of the components: brick components, a core component, and active hinges [29], which is based on Robogen [4]. The brick components are cubic with four slots on their sides available for other modules to attach to.…”

Section: Robot Morphologiesmentioning

confidence: 99%

“…"go 2 Related work Evolutionary robotics aims to design adaptive robots automatically that can evolve to accomplish a specific task while adapting to environmental changes [2]. A number of studies have demonstrated the feasibility of evolutionary methods for generating robotic control and/or morphology [29].…”

Section: Introductionmentioning

confidence: 99%

Directed Locomotion for Modular Robots with Evolvable Morphologies

Lan

Jelisavcic

Roijers

et al. 2018

Parallel Problem Solving From Nature – PPSN XV

Self Cite

View full text Add to dashboard Cite

We generalize the well-studied problem of gait learning in modular robots in two dimensions. Firstly, we address locomotion in a given target direction that goes beyond learning a typical undirected gait. Secondly, rather than studying one fixed robot morphology we consider a test suite of different modular robots. This study is based on our interest in evolutionary robot systems where both morphologies and controllers evolve. In such a system, newborn robots have to learn to control their own body that is a random combination of the bodies of the parents. We apply and compare two learning algorithms, Bayesian optimization and HyperNEAT. The results of the experiments in simulation show that both methods successfully learn good controllers, but Bayesian optimization is more effective and efficient. We validate the best learned controllers by constructing three robots from the test suite in the real world and observe their fitness and actual trajectories. The obtained results indicate a reality gap that depends on the controllers and the shape of the robots, but overall the trajectories are adequate and follow the target directions successfully.North" or "go forward". To validate our results, we test the best controllers on real robots as well. The corresponding general problem statement is as follows:How to learn controllers efficiently for directed locomotion in a variety of modular robots with different morphologies?The paper addresses this problem by three specific research goals:1. Describe possible candidates, i.e., algorithms to learn controllers for directed locomotion in modular robots. 2. Evaluate these algorithms on a test suite of modular robots with different shapes and sizes by comparing their efficacy and efficiency in simulation. 3. Demonstrate the learned directed locomotion behavior on real robots and reflect on the reality gap in this application.

show abstract

Real-World Evolution of Robot Morphologies: A Proof of Concept

Cited by 59 publications

References 39 publications

The ARE Robot Fabricator: How to (Re)produce Robots that Can Evolve in the Real World

The ARE Robot Fabricator: How to (Re)produce Robots that Can Evolve in the Real World

Revolve: A Versatile Simulator for Online Robot Evolution

Directed Locomotion for Modular Robots with Evolvable Morphologies

Contact Info

Product

Resources

About