The aim of the I-SWARM project is the realisation of a “real” micro-robot swarm. More than one hundred micro-manufactured autonomous robots will be able for the collective execution of different tasks in the small world. This is achieved by:

• by the realisation of collective intelligence of these robots
-   in terms of co-operation and
-   collective perception
-   using knowledge and methods of pre-rational intelligence, machine learning, swarm theory and classical multi-    
    agent systems

• by the development of advanced micro-robot hardware
-   being extremely small (size of a single robot: 3 x 3 x 2 mm3)
-   by integrating novel actuators for locomotion, miniaturised powering with micro solar cells and miniaturised
    wireless communication
-   with ICs for on-board intelligence and
-   an integrated sensor and tool for the manipulation in the small world.  

The fundamental outcome of the project is a swarm of micro-robots, which is capable of performing tasks that are not possible with neither one single micro-robot nor with a small group of micro-robots. The outcome is the realisation of an observable self-organisation effect in the robot swarm similar to that seen within other ecological systems such as ant and bee colonies as well as other social insects. There are many potential benefits of such a system including greater flexibility and adaptability of the system to the environment, robustness to failures, etc. Moreover, their collective behaviour opens up new application fields that cannot be solved using today’s tools. The individual agents are able to communicate with each other with a suitable sophisticated IR-based onboard-system, and thus enable and promote the desired swarm effect. A major goal of the project is to transform knowledge gained by observations of eusocial insect behaviour, from observations of communicating insect aggregations and research already performed on swarm intelligence of robots and to apply this to the completely new swarm (concerning swarm and robot size) of micro-robots.

The work addresses the following issues:

• Hardware design of a heterogeneous robot swarm: The realisation of a large number of robot clients presents a major technical challenge and required new and novel approaches in terms of manufacturing and miniaturisation developed during the last years by the I-SWARM consortium. New techniques for the co-design of the miniaturised hardware and its embedded software ‘intelligence’ have been developed.
-   In designing the robot hardware, locomotion principles such as insect-like walking have been realised. Research
    into enhancing, this lead to novel, low-power micro-robot walking mechanisms with very good and high reliability.
-   The knowledge gained by experiments on the “Laws of the small World” significantly deepened the
    understanding of micro-physics applied to micro-handling.
-   Furthermore, the development of pre-rational intelligence modules helps us to create a swarm intelligence
    distributed over the whole system, thus making it less prone to failures and improving its capability to adapt to
    new situations.
• More importantly, systems have been developed which enable the swarms’ behaviour for solving given tasks like dispersion or aggregation to be modelled and thus predicted. This requires the development of knowledge not only about the internal systemic behaviour of a large number of heterogeneous agents, but also in describing the behaviour with specialised mathematics for this purpose. The results of this work enable the design of customised swarms which act in a predetermined way.

There are many applications for large micro-robot swarms:

• Assembly tasks in the micro-world such as assembling of gears, micro-pumps and other micro-systems
• Cleaning surfaces in a very short time
• Mechanical self configuration
• Future medical applications (e.g. examine and medicate the human body inside and outside)
  

How to view 3D Pictures?