Humanoid taught synchronized movement

May 27, 2012 21:46

In recent years, a variety of companies and laboratories have developed an impressive number of humanoid robots that walk and run. Some of them even dance in groups of up to 20, performing complex dance numbers.

This synchronization — no easy task. One way to do this — to provide a robot as a leader, which will transmit the information moves across the network. As a result, the others can catch them and repeat after him.

But the problem is that the network dynamics are not as predictable as one would like to choreographers. For example, there may be a slight delay for half a second, and the transfer of some messages may be delayed even for a few seconds. It is not enough for the usual dance, or any other type of synchronous behavior.

Thus, the method preferred by robotics is to program each robot and their synchronization in the beginning.

The advantage is that if the idea is short enough, then the chances that will happen desynchronization small. The disadvantage of this method is that if the robots desynchronizing — one of them will fall, for example — then there is no clock recovery method.

Thus, the robot began to look more perfect form of synchronization, which would be more resistant to the various tests and accidents that occur with robotic dancers. So, Patrick Bechon (Patrick Bechon) and Jean-Jacques Slotin (Jean-Jacques Slotine) from the Massachusetts Institute of Technology have shown a new approach based on a biological phenomenon cooperative sensitivities or "quorum sensing» (quorum sensing).

Biologists have long puzzled over the ability of bacteria and social insects to feel not only the presence of compatriots, but their number and synchronize their behavior.

It turns out that these creatures are synchronized through a process called co-operative sensitivity or "quorum sensing." It works on the basis of continuously released into the environment of the signaling molecules and measuring the local concentration. This concentration is increased to the extent that, as new join. After reaching the threshold, they cause different behavior, such as cell division and building nests.

Bechon Slotin and argue that this approach provides a reliable way to sync humanoid robots. The ideal approach to the synchronization of robots — is to have access to the position of any other robot. Co-operative sensitivity allows each robot to access a global variable, such as the mean or average position of the day. Each robot can change this variable, because it contributes to the average.

The idea is that if each robot tries to synchronize with the global average, then the group as a whole will have a good time.

The researchers tested their approach with a group of eight robots NAO, built by the French company robotics Aldebaran. Each of them has an internal clock that synchronizes with the global average time from a central server.

It is important to note that the server simply monitors the signal. If the connection is lost, the robots just keep out of sync with it. The central server acts as a medium which robots can feel and interact with it.

This organization has a significant advantage, because if one robot falls, he can come back and join the others again, when there is a resynchronization of its movement with the group (see video).

This work is part of a development in the field of robotics. With the advent of relatively cheap humanoid robots Aldebaran and other companies are now possible, and large-scale synchronization of humanoids.

Thanks synchronicity allows more robots to perform something in one and the same time, they will be able to cope with different interconnected tasks, not just dancing.

In other words, the timing is a technology for collaboration in a large scale. And it opens the door to a whole new set of challenges that robots will be able to do, including in manufacturing and construction.

Like this post? Please share to your friends: