[alife] PhD Studentship in Complex Systems/Bio-Inspired Computing/Swarm Robotics

[Rene Doursat]

PHD STUDENTSHIP IN COMPLEX SYSTEMS/BIO-INSPIRED COMPUTING/SWARM ROBOTICS

Manchester Metropolitan University, UK, Spring 2017

"Devobot: From biological morphogenesis to morphogenetic swarm robotics"

https://goo.gl/aLkxF0

SUMMARY

Swarm robotics promises to offer adaptivity and robustness through decentralisation, but its challenge remains to (meta-)design the motion control of individual units so they can spontaneously form single functional entities. Using 40 Psi Swarm bots, this "Morphogenetic Engineering" project will combine programmable self-assembly with high redundancy, following inspiration from biological development.

AIMS AND OBJECTIVES

Complex systems are characterised by the self-organisation of small, repeated elements into large-scale patterns. They are abundant in our environment and provide a powerful source of inspiration toward a new generation of technological systems, as they often exhibit "organic" properties (decentralisation, autonomous development, evolution) still largely absent from classical engineering.

Swarm robotics, a new technological family of complex systems propelled by recent advances in smaller, cheaper and faster hardware, promises several benefits: multiple robots can accomplish different functions collectively and dynamically; they can collect data from larger areas; decentralisation is robust against local failures; etc. Its major challenge, however, remains to (meta-)design the motion control and self-assembly of individual units so that they can collectively operate as a single entity when physically connected or loosely aggregated.

The Devobot project targets the best of two worlds by combining programmable self-assembly with high redundancy and robustness, following the principles of Morphogenetic Engineering (Doursat et al. 2012, 2013) [1,2]. In multicellular organisms, the number and positions of cells are imprecise, while tissues and organs are reliably formed and arranged. Similarly, multirobot development (by recruitment instead of division) should afford to be irregular at the level of individual agents while preserving an order at the level of the swarm.

MAIN CHALLENGES

* Design a realistic simulation of morphogenetic swarms of many small mobile robots that can flock and self-assemble into specific shapes, based on robotics simulator platforms such as MORSE or Gazebo (including two-wheeled motion, infrared proximity sensors, RF transceivers, and various physical constraints).

* Cross the reality gap through a real-word hardware implementation of these realistic simulations by programming a group of 40 Psi Swarm robots [3] provided by the York Robotics Lab (Prof. Jon Timmis, Director).

TO APPLY

Informal enquiries can be made to Prof. Rene Doursat, Informatics Research Centre (IRC), School of Computing, Mathematics & Digital Technology (SCMDT), Manchester Metropolitan University, r.doursat at mmu.ac.uk

Please apply here: https://goo.gl/aLkxF0

** CLOSING DATE: 31 January 2017 (9am GMT) **

[1] http://doursat.free.fr/mebook.html
[2] http://doursat.free.fr/mewnaco.html
[3] http://www.york.ac.uk/robot-lab/psiswarm/

======
Rene Doursat, PhD, Habil.
Professor of Complex Systems,
Deputy Head, Informatics Research Centre,
School of Computing, Mathematics & Digital Technology,
John Dalton Building E158, Chester Street,
Manchester Metropolitan University,
Manchester, M1 5GD, UK

Web: http://doursat.free.fr -
email: R.Doursat at mmu.ac.uk -
work: +44 (0)161 247 3589


"Before acting on this email or opening any attachments you should read the Manchester Metropolitan University email disclaimer available on its website http://www.mmu.ac.uk/emaildisclaimer "