Networked, flexible and with collective behaviour – these are the key characteristics of the projects that the engineers from the Bionic Learning Network are presenting at the Automatica 2016. Along with the BionicANTs, artificial ants with cooperative behaviour, this year’s flight objects, the eMotionButterflies, also demonstrate what networking means for Festo. The presentation is rounded off with bionic grippers, amongst them the FlexShapeGripper, designed on the natural model of the chameleon’s tongue.
The projects from the Bionic Learning Network illustrate in an entertaining way how individual systems can use communication to merge into an intelligent overall system – but also how the real and virtual worlds can grow together and what role decentralised systems intelligence plays in this process.
BionicANTs – cooperation based on the behaviour of ants
The technology carrier BionicANTs uses the cooperative behaviour of ants as a model. Engineers from Festo used complex regulation algorithms to transfer the behaviour of these insects to the world of technology: just like their models from nature, the BionicANTs cooperate in accordance with clearly defined rules. This enables the BionicANTs to react autonomously to different situations as individual units, to coordinate their behaviour with each other and to act as a networked overall system. By pushing and pulling in a coordinated manner, they shift loads that one ant could not move alone. All actions are based on a distributed catalogue of rules that was devised in advance by means of mathematical model-building and simulations and is programmed into each ant. The individual insects are thus able to make decisions autonomously, while nevertheless subordinating themselves to the common aim; each ant thus contributes its share to solving the task at hand.
The required exchange of information between the ants is effected via the radio modules in their torsos. The regulation strategy comprises a multi-agent system, in which the participants have equal rights. With the 3D stereo camera in their heads, the ants recognise an object to be grasped and can determine their own locations. The BionicANTs’ cooperative behaviour and decentralised intelligence open up interesting prospects for the factory of tomorrow. Future production systems will be based on intelligent components that can flexibly adapt to different production scenarios and process orders from the superordinate control level.
A novel application of 3D MID technology
However, not only the cooperative behaviour of the artificial ants is fascinating. Their manufacturing process is also unique: laser-sintered components are subsequently fitted with visible circuit structures in the so-called 3D MID process. The electrical circuitry is attached to the outer surface of the components, which thereby assume both a structural and an electrical function.
Grasping and walking by means of piezo technology
The BionicANTs are fitted with piezo-ceramic bending transducers in their mouthparts for gripping objects, and also for their leg movements. The piezo elements can be very precisely and rapidly controlled; they are energy-efficient and practically wear-free – and all this within a very small installation space. When the upper bending transducer is deformed, the ant raises its leg. With the lower pair, each leg can be precisely moved forward or backwards.
eMotionButterflies – ultra-lightweight flying objects with collective behaviour
The bionic butterflies likewise illustrate complex topics from the world of future production such as functional integration, ultra-lightweight construction and above all real time-optimised networked communication between individual systems.
Coordinated flying thanks to indoor GPS
An external, well-networked guidance and monitoring system coordinates the individual flying objects autonomously and safely in three-dimensional space. The communication and sensor technology used constitutes an indoor GPS system that controls the butterflies’ collision-free collective movement. The combination of integrated electronics and external camera technology with a host computer ensures process stability.
To emulate the flight behaviour of their natural models as closely as possible, the artificial butterflies are fitted with integrated electronics. These control the wings precisely and individually to effect the rapid flight movements. A human pilot is not required to control the eMotionButterflies. Pre-programmed flight paths for the bionic butterflies’ manoeuvres are stored on the host computer. By means of additionally stored behaviour patterns, however, they can also fly autonomously. In this case, there is no direct communication between the eMotionButterflies.
In devising the control system for the artificial butterflies, the developers benefited from their knowledge gained with the projects involving the BionicOpter and the eMotionSpheres. The indoor GPS was already used for the hovering spheres and was further developed for the eMotionButterflies with regard to the image rate of the cameras.
FlexShapeGripper – grasping like a chameleon’s tongue
The FlexShapeGripper has been developed in close cooperation with the University of Oslo. His working principle derived from the chameleon’s tongue. The chameleon strategically uses the tip of its tongue to catch different insects. Its tongue adapts flexibly to the particular insect.
The FlexShapeGripper can grip, collect and release several objects of very different shapes in the one process without the need for manual conversion. This is made possible by its water-filled silicone cap, which wraps itself around the items being gripped in a flexible, form-fitting manner. This ability to adapt to different shapes gives the FlexShapeGripper its name. The holding and release mechanisms both operate pneumatically. No additional energy is required for the holding process. The yielding nature of the compressible air facilitates the coordination of the handling and the gripper during the grasping process.
Soft gripping for a variety of tasks
In future, the FlexShapeGripper could be used wherever differently shaped objects are to be handled simultaneously, for example in service robotics, in assembly tasks or in the handling of small components.