Welcome to MARS

The Martian Autonomous Robot Swarm | Smartgeometry 2016 Gothenburg

Workshop Goal

We will develop a distributed 3D printing system to manipulate material on site resulting in emergent forms.

Cluster participants will experiment with different scenarios, goals and behaviours in simulation and on actual robot.

Autonomous Robotics

Follow how we are developing our MARS robots on the Blog


Research Question

This workshop aims to determine the advantages and drawbacks of utilizing a robot swarm to autonomously complete a construction task.

Architecture and construction is most often a complex orchestration of dozens of trades under the guidance of a skilled conductor and it would seem that such a multiplicity cannot be managed any other way. But looking at nature, it has been established that many of its inner workings do not depend on global control and a top-down organization. school of herring thousands strong can evade the ocean’s most highly evolved predator using coordinated movements.

Bottom-up systems not only have great advantages for fish, birds and insects, but also for construction teams. Unimagined solutions are enabled because each agent works autonomously reacting to local conditions. But all these local interactions emerge into an efficient and effective global result. Also, with the intelligence of the system distributed to the individual agents, the potential for catastrophic failure is mitigated by redundancy. These systems are more flexible and more resilient to change or error than a top-down approach, which fails once the central command stops working.

This idea is particularly useful for architecture in alien environments, where human presence and control is too difficult or dangerous. A redundant autonomous system is preferable because the long communication delay is impractical and it is difficult to repair machines remotely. A system comprising of multiple autonomous parts can adapt to change and failure of a component and will not result in a mission fail if one or a few components malfunction.

Cluster Goals

We will teach a self-organizing swarm of small construction robots which follow local rules to build an emergent structure.

This goal is twofold: It is a research of the aesthetics and forms which can be generated by a stochastic bottom-up approach without a forced top-down design. Secondly it is a research on arising challenges when setting-up, deploying and utilizing a bottom-up construction robotic system.

The robots will navigate the site and remove and place local material in three dimensions, whilst sensing their local environment and avoiding obstacles and each other. The robots will act independently, but their programming and rules will be evaluated and optimised during the task.

Background & Research methods

The research explores a prototype for a 3D printing system to locate and sinter Martian soil, as previously envisioned by Foster + Partners.

Behaviour has significance for understanding biological intelligence and therefore it will be the main focus of the workshop. It is understood as the interactions between an organism and its environment where the behaviour of the organism affects its own perceptions, and thus its future actions and perceptions. Applying this concept to robots results in a field of autonomous behavioural machines capable of operating in partially unknown and changing environments without human intervention.

These emergent forms and behaviours are simulated first where different simple local rules and their impact are explored. The resulting forms are evaluated for their aesthetic value as well as structural stability. This experimenting includes multiple cycles and allows programmers to observe and better understand the relationship between bottom-up local rules and their corresponding output.

Robotic Swarm

We will build on the highly popular Makeblock a kit robot system to build robots which can navigate and manipulate terrain and combine this with machine vision to allow them to make informed decisions.


The robots will be equipped with strong motors and tracks to be able to move around the site and up the structure. They will also be able to pick up and put down material with a scoop using a conveyor belt to place material.The cluster champions will propose solutions for this, and the participants will be encouraged to adjust, update and refine this.


Each robot will be equipped with sensors, cameras and a feedback system to allow it to react to the environment. There will be no global communication or central control, although additional camera’s and microphones might be used to evaluate the overall level of organization within the swarm, or lack thereof.


The robots will be made using the pretested Makeblock a kit system containing all necessary parts, sensors and motors to allow for flexibility in design within a framework which has proven its reliability. Programming will be kept simple using Arduino or even Scratch languages which are sent to the robot over a wireless connection.

Currently Happening on Mars

We are still developing and improving our robotic system. Have a look on our Blog page to see what we are up to.

Planned Schedule

Throughout the four days we will try to capture the development of the behavioural rules and the cumulative construction.

Day one

Introduction to participants of bottom up processes, explanation of local rule based behaviour. Finalising of assembly of the robots and building blocks.

Day two

Introduction of robots programming interface. Demonstration of a behaviour and programming of experimental robot’s DNA, or their behaviour dependent on local sensor input. Transferring these rules to a simulation environment to have a basic understanding of the designed behaviour.

Day Three

Tests to build structures using emergent behaviour through local rules, iterative process of simulation, improvement of behaviour and local rules.

Day Four

Construction of large structure using emergent behaviour.

Our Team

Our team is part of the multidisciplinary Specialist Modelling Group within Foster + Partners. We come from diverse backgrounds with architecture and computation being the common ground where our interests meet.
Josef Musil

Josef Musil

Cluster Champion
Josef is part of the research group SMG within Foster + Partners in London and focuses on applied research, application of new technologies and algorithmic design.
Jan Dierckx

Jan Dierckx

Cluster Champion
Jan is part of the Specialist Modelling Group at Foster + Partners where he combines an internal consultant role with research in large-scale manufacturing and the integration of robotics and automation in the office.
Richard Maddoc

Richard Maddock

Cluster Expert
Richard holds degrees in software engineering as well as architecture, and is currently an associate at Foster + Partners as part of the Specialist Modelling Group.
Miriam Dall'igna

Miriam Dall'igna

Cluster Expert
Miriam works as associate Design Systems Analyst at the Specialist Modelling Group at Foster + Partners and Computational Design Tutor at the University of Westminster.
Octavian Gheorghiu

Octavian Gheorghiu

Cluster Expert
Octavian graduated from the AADRL and now works as a design systems analyst at Foster + Partners within the Specialist Modelling Group. He is a visiting lecturer at The Bartlett UCL.
Sam Wilkinson

Sam Wilkinson

Cluster Expert
Sam works at Foster + Partners Specialist Modelling Group. His interests are in computational intelligence, additive construction, and the future.
Philip Robinson

Philip Robinson

Cluster Expert
Philip Robinson is an expert in auditory perception and architectural acoustics who develops auralization, simulation, and virtual reality systems to support architectural design.
Xiaoming Yang

Xiaoming Yang

Cluster Expert
Xiaoming Yang is an environmental designer focused on daylighting and lighting design. He studied in National University of Singapore as an electrical engineer and now works at Foster + Partners.

Get in Touch with us

Still have Questions? Don't hesitate to contact us!

Workshop locationGothenburg, Sweden

Speak to Us+44 7413 684883 / +44 7447 560970

Like us on FacebookGet social!

Follow on Twitter2.1k Followers