In recent years, various companies and labs have developed impressive humanoid robots that walk, shuffle and even run. Some even dance in groups of up to 20, performing sophisticated choreographed routines.
This kind of synchronisation is no easy task. One way to do it is have one robot as the leader, broadcasting details of its movement and position over a network that the other robots all follow.
The trouble is that network dynamics are not as predictable as choreographers would like. Small delays of half a second or so are common while some messages can be delayed by several seconds. That’s clearly not good enough for a dance routine or any other type of synchronised behaviour.
So the approach preferred by roboticists is to program each robot with the dance routine, synchronise their internal clocks at the start of the performance and then leave them to it.
The advantage is that If the performance is reasonably short, the chances of the clocks becoming desynchronised can be made small. The disadvantage is that if the robots become desynchronised–if one falls over, for example–there is no way to regain synchronisation.
So roboticists have been searching for a better form of synchronisation that is more robust to the various trials and tribulations that befall robotic dancers. Today, Patrick Bechon and Jean-Jacques Slotine at the Massachusetts Institute of Technology in Cambridge, reveal a new approach based on the biological phenomenon of quorum sensing.
Biologists have long puzzled over the ability of bacteria and social insects to sense not only the presence of compatriots but their number and to synchronise their behaviour.
It turns out that these creatures perform this synchronisation using a process called quorum sensing. This works by constantly releasing signalling molecules into the environment while at the same time measuring the local concentration of these molecules.
This concentration rises as more creatures join the local population and so is an effective measure of population density. When the concentration rises over some threshold level, it triggers a different behaviour such cell division, pathogen production and nest building.
Now Bechon and Slotine say a similar approach provides a robust way to synchronise humanoid robots. The ideal approach to synchronisation is for each robot to have access to every other robot’s position. Instead, the quorum sensing approach gives, each robot access to a global variable such as the average position or average clock time. Each robot can also change this variable because it contributes to the average.
The idea is that if each robot attempts to synchronise with this global average, the swarm as whole should keep good time.
These guys test out their approach with a group of eight NAO robots built by the French robotics company Aldebaran. Each has an internal clock which attempts to synchronise with a global average time maintained by a central server.
It’s important to point out that the server is not acting as a master with the robots as slaves that simply follow its signal. If the connection to the central is lost, the robots simply continue with routine but without centralised synchrony.
Instead, the central server is more like a a kind of environment that the robots can sense and interact with.
This arrangement has the significant advantage that if one robot falls over it can simply get back up and join in again when it has resynchronised its movements with the group (see video).
This work is part of a broader development in robotics. The advent of relatively cheap humanoid robots from Aldebaran and other companies means that the large-scale sychronisation of humanoid swarms is now possible.
That’s interesting because while synchrony allows large numbers of robots to do the same thing at the same time–such as dancing or marching–it also allows large number so robots to do different but related tasks at the same time.
In other words, synchrony is an enabling technology for large scale co-operation. And that opens the way to an entirely new set of tasks that robots could do–think manufacturing and construction. Perhaps even nest building.
Ref: arxiv.org/abs/1205.2952: Synchronization And Quorum Sensing In A Swarm Of Humanoid Robots
The big new idea for making self-driving cars that can go anywhere
The mainstream approach to driverless cars is slow and difficult. These startups think going all-in on AI will get there faster.
Inside Charm Industrial’s big bet on corn stalks for carbon removal
The startup used plant matter and bio-oil to sequester thousands of tons of carbon. The question now is how reliable, scalable, and economical this approach will prove.
The hype around DeepMind’s new AI model misses what’s actually cool about it
Some worry that the chatter about these tools is doing the whole field a disservice.
The dark secret behind those cute AI-generated animal images
Google Brain has revealed its own image-making AI, called Imagen. But don't expect to see anything that isn't wholesome.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.