Occlusion-Based Cooperative Transport With A Swarm of Miniature Mobile Robots Seminar

WELCOME
Occlusion-Based
Cooperative Transport
with a
Swarm of Miniature
Mobile Robots
Guided by,
Presented by,
Overview
Introduction
Occlusion-based cooperative transport
strategy
Mathematical analysis
Advantages
Implementation and results
Conclusion
References
3
Introduction
Strategy for transporting a large object to a

goal using a large number of mobile robots
Can transport any convex object in a planar
environment
Particularly suited for implementation on micro-
scale robotic systems
Occlusion-based Cooperative
Transport Strategy
Here, occlusion is used to organize a swarm of
robots to push a large object to a goal.

The strategy can be implemented in a fully
decentralized manner.
The robots use on-board cameras to perceive
the object and goal.
Illustration of how a swarm of robots can push a

large object in a 2D planar environment.
The robots keep pushing only along the section
of the objects perimeter that occludes their

views of the goal.
The motion of the object will be approximately
toward the goal.

Neither consistent perception of the goal nor
explicit communication are required for robots

that are pushing the object.
State machine representation of the individual

robot behaviour realizing the occlusion-based
cooperative transport strategy
8
Mathematical Analysis
M athem atical m odelling
If normal forces are uniformly applied on the blue section of the convex-shaped
objects perimeter (major arc ab in this diagram), the combined force vector F is the
vector (b a) rotated by 2 and its magnitude is proportional to the length b a
(chord ab in this diagram). Point q is an affecting point of F. c denotes the center of
mass of the object. g denotes the goal.

We
assume that each of the goals and robots are
points.
Let c R2 be the centre of mass of a rigid convex
object with respect to a coordinate frame in which

g = [0, 0]T is the goal point.
Let the perimeter of the object be described by a
closed, convex, and differentiable curve given by

p() = + c
a is the tangent point on the right side of vector
(c g), while b is the one on the left side.

10
Resultant Force Applied on the O bject
1: Assume that n robots are uniformly

Lemma
distributed along the occluded perimeter of the

object and they are the only robots asserting a
force on the object. The direction of the resultant
force asserted on the object by the robots is
equal to the direction of the vector (b a)
rotated by and its magnitude is proportional to .
Lemma 2: If the combined force contributed by
the robots F is considered as a single force, while

Q is the torque induced by F, the mid point of
segment ab is an affecting point of F.
11

The combined force is the definite integral given
by
F = p() d.
The solution of the definite integral is
F= ( b - a )
12
M otion D ynam ics ofthe O bject
As
the object is moved, a and b can change over time
Thus the solution of the definite integral is valid at
any point in time as long as g is outside p. In other

words,
F(t) =
We can apply a quasi-static analysis and obtain that
c = kF
where c is the derivative of c and k is a positive
constant that transfers F proportionally to the velocity
of the object.
13
Convergence ofthe O bjects D istance to the

G oal
Theorem 1: The distance between the objects
centre of mass (c) and the goal (g) is strictly

decreasing over time if the velocity of the object
is governed by c = kF.
As t , g will be on the object perimeter p.
14
mplementation and Resu

For the physical implementation, e-puck robot
platform is used.
e-puck robot
Top-view schematic of an e-puck

indicating the locations of its wheels,
camera, and proximity sensors.
15
e-puck emits infrared light almost continuously.

The directional colour camera is used to
recognize both object and goal.

The e-puck controller is a state machine
implementing the individual behaviours of the

transport strategy.
Experiments with three objects of different
shapes and sizes were conducted.

16
Experimental setup
Completion time of the circular object,

scalene triangular object, and elongated
rectangular object
17
Circular
object
Triangular
object
(c)
Rectangular
object
Predicted path vs. actual paths of the centroid of

the objects
18
Experim ents w ith a m oving goal.

If the goal is a
mobile robot, it
can change its
position, while the
object is being
transported.
Setup for experiments with a moving goal. The

initial position of the object was alternated
between a and b, while their corresponding
destination regions were A and B
A tele operated
goal robot was
used to guide the
pushing robots
(and thus the
object) through a
corridor with
corners.
19
20
Advantage of the
Occlusion-based
Cooperative Transport
Occlusion is used to organize a swarm of
Strategy
robots
Object always arrives at the goal.
Possible
to transport objects of various

shapes and sizes.
The robots do not need to communicate
(explicitly) with each other.

21
C onclusion
A cooperative transport strategy that uses a
large number of relatively simple and small

mobile robots to transport a large object that
can occlude the robots perception of the goal
was introduced.
A mathematical formulation of the strategy
was provided.
22
R eference
[1] Jianing Chen, , Melvin Gauci, Wei Li, Andreas Kolling, Roderich Gro,
Occlusion-Based Cooperative Transport with a Swarm of
Miniature Mobile Robots in IEEE TRANSACTIONS ON ROBOTICS,
VOL. 31, NO. 2, APRIL 2015
[2] J. Chen, M. Gauci, and R. Gro, A strategy for transporting tall objects
with a swarm of miniature mobile robots, in Proc. IEEE Int. Conf. Robot.
Autom., 2013, pp. 863869.
[3] C. Kube and E. Bonabeau, Cooperative transport by ants and robots,
Robot. Auton. Syst., vol. 30, no. 12, pp. 85101, 2000
[4] Occlusion-Based Cooperative Transport with a Swarm of Miniature
Mobile Robots-Online Supplementary Materials from
http://naturalrobotics.group.shef.ac.uk/supp/2014-002/
23
Thank You
24
Questions
25

Occlusion-Based Cooperative Transport With A Swarm of Miniature Mobile Robots Seminar

Uploaded by

Copyright:

Available Formats

Occlusion-Based Cooperative Transport With A Swarm of Miniature Mobile Robots Seminar

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Occlusion-Based Cooperative Transport With A Swarm of Miniature Mobile Robots Seminar

Uploaded by

Copyright:

Available Formats

WELCOME

Strategy for transporting a large object to a

Can transport any convex object in a planar

scale robotic systems

robots to push a large object to a goal.

the object and goal.

Illustration of how a swarm of robots can push a

The robots keep pushing only along the section

of the objects perimeter that occludes their

toward the goal.

explicit communication are required for robots

State machine representation of the individual

Let c R2 be the centre of mass of a rigid convex

object with respect to a coordinate frame in which

closed, convex, and differentiable curve given by

(c g), while b is the one on the left side.

Resultant Force Applied on the O bject

1: Assume that n robots are uniformly

distributed along the occluded perimeter of the

Lemma 2: If the combined force contributed by

the robots F is considered as a single force, while

M otion D ynam ics ofthe O bject

any point in time as long as g is outside p. In other

Convergence ofthe O bjects D istance to the

centre of mass (c) and the goal (g) is strictly

As t , g will be on the object perimeter p.

mplementation and Resu

Top-view schematic of an e-puck

e-puck emits infrared light almost continuously.

recognize both object and goal.

implementing the individual behaviours of the

shapes and sizes were conducted.

Completion time of the circular object,

Predicted path vs. actual paths of the centroid of

Experim ents w ith a m oving goal.

Setup for experiments with a moving goal. The

to transport objects of various

The robots do not need to communicate

(explicitly) with each other.

large number of relatively simple and small

A mathematical formulation of the strategy

You might also like