Physical Computing


Artificial Potential Fields – Cooperative Payload Transport by Robot Collectives by Student
September 25, 2009, 10:14 am
Filed under: Inspiration

If we look at all life on this planet, we clearly see that we work best together. This is a thought that should be transferred and used in the world of robotics. We can also see that most of the, above-mentioned, life also have some sort of intelligence. If we join the ability to work in a cooperative with some sort of intelligence, we have something very powerful. It helped the humans get where we are today. We can clearly see that the more intelligent a life form is the more individual it is. For example the humans, can without any problem solve problems alone while, for example, ants always work in groups.
So if we look at the relatively stupid robots we have today we can see that they, most of the time, work alone. They have their own task, which they perform without questioning if it is the best way or if it is necessary. One single robot is often enough, but in a lot of situations the use of multiple robots can increase the usability and also work in an error/fault minimizing way.

cooperate_robot03

A project group at University of Buffalo have been doing some interesting contribution in this area. Their vision is to create a basic framework for semi-autonomous payload transport by a collective of wheeled mobile robots. The robots would collaborate with constraints, new information, map building and reconnaissance. This would allow the robot collective to be smarter, more flexible and the combined system would be more robust. The group of robots are also able to split up into smaller groups. This means that a single robot can be sent on a different mission, while the rest of the group prepare for the single robot to return or aid from a distance.

The kinematic is handled in an efficient way, they have different rankings, there is a leader a vice leader and so on. This can be changed depending on which robot has the most critical position. The speed, angle and angular velocity are shared within the collective, which makes it possible for the other robots to calculate their trajectories.

recursive_modular_framework

The path planning is handle by Artificial Potential Fields. This is a modern way of handling the problem with finding the best (good enough) path. The Potential Fields uses a starting position and a goal position. Both of these can off course be updated during the mission. The robots then use sensors to sense for object, such as walls, pits and other robots. These objects are then added to a map, which is shared within the collective. The objects then repulse the robots, while the goal(s) attract them. The map is filled with gradients that contain information about speed and angular velocity, which is used by the robots. This gives the robot a continuously updating best path to follow.

Artificial Potential Fields Video

Leader follower example video

Decentralized Collaboration video

All information in this blog post is taken from this site. Here you can also find more information, videos and pictures.


Leave a Comment so far
Leave a comment



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s



%d bloggers like this: