Autonomous Robot Evolution: Cradle To Grave

Lead Research Organisation: Edinburgh Napier University
Department Name: Computing

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
 
Description The project has demonstrated that it is possible to design a complete robot (body and brain) using a fully autonomous system based on the principles of natural evolution. We have developed a framework in which hybridises artificial evolution in software and hardware to create blueprints for robots that are optimised for solving a particular task. The evolved designs are autonomously manufactured using a robotic arm, hence the entire system can operate without a human in the loop. The robot consist of a 3d printed plastic chassis to which a variety of sensors and actuators can be attached. To the best of our knowledge this is the first time that evolution of a complete robot with closed loop control in a rich morphological space has been demonstrated. The evolved robots are capable of completing simple tasks such as maze navigation and targeted locomotion. The framework could be adapted by other researchers.

A number of challenges were overcome: from the hardware perspective, a significant amount of effort was put into the engineering design of components to ensure the robots could be reliably assembled by the automated arm. From the software perspective, a number of advances in evolutionary algorithms were made in order to address the issue of evolving designs that were manufacturable, and to allow evolved 'brains' to quickly adapt to new complex bodies, which required mixing learning with evolution.

A number of open questions arise from the results, for example (1)i how the learning phase can be made more efficient so that robots learn more quickly (2) how to minimise the reality-gap and (3) how to enable evolution to mantain sufficient diversity within an evolving population to enable very complex forms/brains to emerge
Exploitation Route The results could be taken forward by sectors which require the use of roots that are capable of adapting autonomously to previously unseen environment and to unexpected scenarios. This has future applications for example in space (robots for exploring planets etc) and in environmental applications (e..g monitoring or exploration in harsh environments such as glaciers/forests that are hard to reach by humans.
Sectors Aerospace, Defence and Marine,Environment,Manufacturing, including Industrial Biotechology

URL https://www.york.ac.uk/robot-lab/are/
 
Description There have been a number of "popular press" articles written about the project, e.g. in the Guardian, the Telegraph, The Scotsman, The New Scientist Demonstrations have been given at York Festival of Ideas in 2019. Prof. Hart gave a TED talk about the work in the project at TEDWomen 2021 As a result of this media attention, the work has been brought to the attention of potential end users (e.g Prof. Hart was invited to give a talk to the European Space Agency)
First Year Of Impact 2021
Sector Digital/Communication/Information Technologies (including Software),Electronics
Impact Types Societal

 
Description Article in Daily Telegraph about the project 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Media (as a channel to the public)
Results and Impact Article in which Prof Hart and Prof Winfield were interviewed re the research undertaken in the project
Year(s) Of Engagement Activity 2021
URL https://www.telegraph.co.uk/technology/2021/03/14/british-engineers-creating-self-replicating-robots...
 
Description Article in Technology Magazine 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact article in German edition of magazine "Technology Review"
Year(s) Of Engagement Activity 2019
URL https://www.heise.de/tr/artikel/Darwins-Maschinen-4404076.html
 
Description Article in national newspaper (the Scotsman) 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Media (as a channel to the public)
Results and Impact Article in Scottish national newspaper the Scotsman discussing the project and its implications
Year(s) Of Engagement Activity 2022
URL https://www.scotsman.com/news/opinion/columnists/robots-are-being-redesigned-by-computer-programs-th...
 
Description Article in the Conversation: We're teaching robots to evolve autonomously - so they can adapt to life alone on distant planets 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact The article was published in the Conversation. It had over 7000 reads in one week and over 16000 reads in the first 3 weeks. It was shared 617 times from the Conversation site.
It was republished in Italian and in Hungarian
Year(s) Of Engagement Activity 2021
URL https://theconversation.com/were-teaching-robots-to-evolve-autonomously-so-they-can-adapt-to-life-al...
 
Description Department open day 
Form Of Engagement Activity Participation in an open day or visit at my research institution
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Undergraduate students
Results and Impact talk and demonstration of project results at an UG open day
Year(s) Of Engagement Activity 2022
 
Description Feature article in New Scientist 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Media (as a channel to the public)
Results and Impact Feature article in New Scientist ( Vol. 253 Issue 3375, 25th Feb 2022)
Year(s) Of Engagement Activity 2022
URL https://www.newscientist.com/article/mg25333751-700-meet-the-robots-that-can-reproduce-learn-and-evo...
 
Description Guardian article describing research in ARE project 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Media (as a channel to the public)
Results and Impact Guardian article (authored by Emma Hart) : created much discussion online and resulted in several other newspapers running articles that reported on this
Year(s) Of Engagement Activity 2021
URL https://www.theguardian.com/commentisfree/2021/jun/21/robots-reproduce-evolution-nature-technology
 
Description Invited Seminar: University of La Laguna Tenerife 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Postgraduate students
Results and Impact An invited talk at the Latest Advances in Computer Science series of seminars at University La Laguna
Year(s) Of Engagement Activity 2020
 
Description Keynote talk at International conference 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Other audiences
Results and Impact Keynote talk at major international conference
Year(s) Of Engagement Activity 2022
URL https://wcci2022.org/invited-speakers/
 
Description Keynote talk at major international conference 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact Keynote talk at IEEE Congress on Evolutionary Computation in Wellington New Zealand centred on the project work describing it to an international audience. The title was "Toward Autonomous Evolution of Robotic Ecosystems"
Year(s) Of Engagement Activity 2019
URL https://ieeetv.ieee.org/conference-highlights/towards-the-autonomous-evolution-of-robotic-ecosystems...