Scribd
Upload
160 views

Use of Swarm Intelligence Algorithms For Pattern Formation PDF

The document discusses using swarm intelligence algorithms to form patterns with mobile robots. The objective is to develop algorithms that can create any given shape using a swarm of robots with minimal sensor and computational requirements. The work done so far includes simulating flocking behavior based on Reynolds' rules and a line formation algorithm. Future work planned includes improving the line formation model and developing algorithms for circles, arcs and generic shapes.

Uploaded by

Vinit Unni
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
160 views

Use of Swarm Intelligence Algorithms For Pattern Formation PDF

The document discusses using swarm intelligence algorithms to form patterns with mobile robots. The objective is to develop algorithms that can create any given shape using a swarm of robots with minimal sensor and computational requirements. The work done so far includes simulating flocking behavior based on Reynolds' rules and a line formation algorithm. Future work planned includes improving the line formation model and developing algorithms for circles, arcs and generic shapes.

Uploaded by

Vinit Unni
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

Use of Swarm Intelligence algorithms for pattern formation in mobile robots

By Vinit Unni ED08B031 Project guide: Dr. Srikanth Vedantam

Objective
To come up with swarm intelligence algorithms that would help us create any generic shape for a given swarm Swarm Intelligence refers to algorithms/systems which work based on the interaction of many independent agents among themselves. The input requirements of the algorithm should be low as more inputs would imply that we need costlier sensor The algorithm should be scale invariant (should not be dependent on number of robots) Should use as much less global data as possible For a typical good swarm algorithm, each robot is identical to every other robot in all aspects and this aspect too should be considered while creating algorithm Possible applications of this could be in using formations of the robots to scout a given area (for example a warehouse) or creating support structures which are easier to transport

Literature Survey
One of the earliest works took inspiration from biological swarms for motion generation-was presented by Reynolds in 1987 where he proposed a method for generating visually satisfying computer animations of bird flocks, often referred as boids. Vicsek et al using the nearest neighbour update rules showed that a group of autonomous agents with different headings but same speed would eventually converge to same heading. Major works in the field of pattern formation have been by Albayrak (for circle and straight line)and Suzuki & Yamashita (for general geometric patterns). Ogren et al and Belta et al. present a different approach to the shape generation/formation control problem where they achieved motion control for decentralized system of robots in a lower dimension

Work done till now


Literature survey Simulation of Flocking algorithm
Reynolds came up with an algorithm to simulate flocking of birds with

Simulation of Line algorithm


A problem statement with a set of assumptions was created (Chapter 0) A line formation algorithm was created keeping these assumptions in mind

Flocking Algorithm
Motion of an individual robot is decided by three factors:o Collision avoidance- Avoid collision with nearby flock-mates o Velocity matching- Match velocity with velocity of nearby flock-mates o Flock centering- Attempt to stay close to nearby flock-mates

When all the robots in a swarm follow the above rules, we observe formation of small clumps (emergent behaviour) When we introduce trajectory or random movement of one particular robot (leader/apex robot) along with this, we get flock formation Size of the smaller clumps initially depends on the sensor range but later with time as all clumps merge, they form one big swarm

Results of flocking algorithm

Initial distribution of robots

Formation of smaller clumps at steady state

Line Formation (assumptions)


2 leader robots mark the two end points of the line and all the follower bots are to form a line between them It is assumed that the follower bots can distinctly identify leader bots All robots know the location and orientations of all the other robots w.r.t their local frame of reference. The sensors are assumed to have large enough range. The robots are assumed to be point sized particles with a sense of orientation in the 2D plane The robots are assumed to be Omni-directional and their minimum turn radius is assumed to be 0 We try to keep the motion of the robots jerk free (acceleration is continuous) Collisions are non-existent There are two modes of movement
Mode 1- Where only after leader robots reach final position do the follower robots start moving Mode 2- All robots start moving simultaneously

Results (Line Formation)

Mode 1

Mode 2

Future Work
Increase complexity in the problem statement of Chapter 0 to make it more realistic
Limit sensor range Include collision avoidance Make leader robots indistinguishable to others

Make and test algorithms for circles/arcs Make and test algorithms for generic shapes in the form of B-spline/Bezier curves

References
Okay Albayrak, Line and Circle Formation of distributed autonomous mobile robots with limited sensor range, June 1996 K. Sugihara and I. Suzuki, "Distributed Algorithms for Formation of Geometric Patterns with Many Mobile Robots" Journal of Robotic Systems, vol.13, no.3, 1996, pp. 127-139 I. Suzuki and M. Yamashita, "A Theory of Distributed Anonymous Mobile Robots -Formation and Agreement Problems," Technical Report TR-94-07-01, Department of Electrical Engineering and Computer Science, University of Wisconsin, Milwaukee, 1994. Craig W. Reynolds, " Flocks, Herds, and Schools: A Distributed Behavioral Model , July 1987 Ali Jadbabaie et al, Coordination of Groups of Mobile Autonomous Agents Using Nearest Neighbor Rules, June 2003 Ichiro Suzuki et al, Distributed anonymous mobile robots: Formation and geometric patterns, 1999

You might also like