[Gp][final seminar][presentation]

Computer and Information Science | Computer Systems Department

Final Seminar 24/6/2013

Co-Swarm: Cooperative area exploration using a swarm of mobile robots June 2013 1Final Seminar

• Team – Ahmed Mohamed Emad– Abdelrahman Ehab Soliman– Muhammed Aly Huessin– Anas Awad Ismail

• Academic Supervisors – Prof. Dr. Hossam El-Din Fahem– Dr. Wael S. El-Kilany– TA. Agwad Hammad

Team Members

• Overview :– Problem Definition – Previous Work

• Objective• Co-Swarm :

– System Architecture – Module_1 (Plan)– Module_2 (Sense)– Module_3 (Act)– Communication

• System Constraints• Experiments (Demo)• Conclusion and Future Work• References

Outline

Overview

ObjectiveCo-Swarm

System ConstraintsExperiments

Conclusion Future Work

References

Problem Definition“In Robotics, the exploration problem deals with the use of a

robot to maximize the knowledge over a particular area.

The exploration problem arises in mapping and search and

rescue situations, where an environment might be

dangerous or inaccessible to humans ”

Overview

Outline OverviewObjective

Co-Swarm System Constraints

ExperimentsConclusion

Future WorkReferences

Problem Overview– The exploration problems appeared when the humans think

to use robots to discover unknown areas that may be dangerous or inaccessible for humans.

– Robot needs a map to operate in a particular environment.

– Autonomous Exploration : The ability of robots to autonomously travel around an unknown environment gathering the necessary information to obtain a useful map for navigation.

– Optimization Problem: NP-Complete Problem

Overview

Co-Swarm: Cooperative area exploration using a swarm of mobile robotsFinal Seminar

Co-SwarmSystem Constraints

June 2013 5

Previous Work• In absence of global position information,

(SLAM) techniques are deployed to construct a map.

• State of the art:– Classic non-coordinated strategies– Classic coordinated strategies– Integrated non-coordinated strategies – Integrated coordinated strategies

Overview

June 2013 6

Previous Work• In absence of global position information,

(SLAM) techniques are deployed to construct a map.

• State of the art:– Classic non-coordinated strategies– Classic coordinated strategies

• Frontier-Based Exploration using Multiple Robots – Brain Yamuchi 1998

– Integrated non-coordinated strategies – Integrated coordinated strategies

Overview

June 2013 7

What’s Swarm ?! • New bio-inspired approach to control the coordination of

(Multi-Robot System)MRS, which consists of large numbers of mostly simple physical robots.

• This collective behavior leads to the field of Artificial Swarm Intelligence.

Overview

Co-Swarm System Constraints

June 2013 8

Why Swarm ?!• Simple and Elegant.– Can be represented by State Machine.

• Scalable.• Decentralized.• Usage of local interactions.

Overview

June 2013 9

“Speedup the area exploration mission using swarm of Mobile Robots”

Objective

Outline Overview

ObjectiveCo-Swarm

System ConstraintsExperimentsConclusion

June 2013 10

• Robot Control Methodology (S-P-A)

Co-Swarm |System Architecture

June 2013 11

Frontier-Based Exploration – Brain Yamuchi 1998

– Central Idea:• Gain the most new information about the world.• Move to the boundary between open space and

uncharted space.

Co-Swarm | Planning

June 2013 12

Co-Swarm | Planning

June 2013 13

Evidence Grid• Our Spatial Representation for the Environment.• A Cartesian Grid containing cells and each cell

contains its state whether it is Open , Occupied or unknown.

Frontier Detection• Frontiers are regions on the boundary between open

space and unexplored space.• A process analogous to edge detection and region

extraction in computer vision is used to find the boundaries between open space and unknown space.

Co-Swarm | Planning

June 2013 14

Navigate to Frontiers• The robot attempts to navigate to the nearest

accessible unvisited frontier.• Motion Planning with Obstacle Avoidance.• Path Finding Algorithm A*.

Co-Swarm | Planning

June 2013 15

Co-Swarm | Planning

June 2013 16

Frontier-Based Exploration (Single Robot)

Co-Swarm | Planning

June 2013 17

Frontier-Based Exploration (Mutli-Robot)

June 2013 18

June 2013 19

Co-Swarm | Planning

June 2013 20

Self-Organizing Models–Mobility Sensor Models• Fully Coordinated Control Model.• Fully Random Model.• Emergent Motion Control.

– Flocking Algorithm.– Anti-Flocking Algorithm

Co-Swarm | Planning

June 2013 21

Self-Organizing Models–Mobility Sensor Models• Fully Coordinated Control Model.• Fully Random Model.• Emergent Motion Control.

– Flocking Algorithm.

–Anti-Flocking Algorithm

Co-Swarm | Planning

June 2013 22

Anti-Flocking Algorithm

Co-Swarm | Planning

June 2013 23

Critical Cases Handling – More than one robot navigate to the same

frontier.– Robot cuts the road of another one.

Co-Swarm | Planning

June 2013 24

Co-Swarm | Planning

June 2013 25

Co-Swarm | Sensing

June 2013 26

Sensing the environment– Using PSD– Using Ultrasonic

Co-Swarm | Sensing

June 2013 27

Sensing the environment– Using PSD– Using Ultrasonic

Co-Swarm | Acting| Framework Design

June 2013 28

Motion Engine

Direction

Ultra Sonic PSD

Path Executor

Command interpreter

Serial Communication (USART)

PID Controller Motor Driver

Local Grid Generator

Message

Path[0] … Path[Length -1]

L_Grid[0] … L_Grid[Length -1]

Distance I , or Angle I Angle i

Distance

Request for Dist.

ENA. signal

Distance

Request for L_Grid .

Co-Swarm | Acting| Framework

June 2013 29

Commands Interpreter Module–Message Splitting– Function Signature– Passing Parameters

June 2013 30

Commands Interpreter Module

Path Executor

Command interpreter

Serial Communication (USART)

Message

L_Grid[0] … L_Grid[Length -1]

June 2013 31

Path Executer Module– Fetch the next step and its direction.– Call the desired function

June 2013 32

Path Executer Module

Motion Engine

Path Executor

Command interpreter

Distance I , or Angle I

June 2013 33

Local Grid Generator– Receive Request .– Uses the Motion Engine Module to move the

robot in a certain angle– Uses the PSD Sensor Adapter to explore robot

local area .– Returns the information about each cell in

robot local whether it is open or occupied.

June 2013 34

Angle i

Motion Engine

Command interpreter

L_Grid[0] … L_Grid[Length -1] Request for L_Grid .

Co-Swarm | Acting

June 2013 35

Motion Engine– Used for controlling robot motion in a specific

direction and for a given distance .– Or used to perform rotation operation on the

robot for a given angle– Motion Engine is responsible for obstacle

detection and avoidance by using the Ultra sonic sensor and PSD sensor .

Co-Swarm | Acting

June 2013 36

Motion Engine

Direction

Ultra Sonic PSD

Path Executor

PID Controller Motor Driver

Distance I , or Angle I Angle i

Distance

Request for Dist.

ENA. signal

Distance

Co-Swarm | Acting

June 2013 37

PID Controller– PID : It is a generic control loop feedback

mechanism (controller) widely used in industrial control systems, we used it to be able to control the wheels rotation of the robots with regular speed .

Co-Swarm | Acting

June 2013 38

PID Controller

Co-Swarm | Communication

June 2013 39

Due to use of External Brain:– Robot-To-Brain Communication (USART, Bluetooth)

– Brain-To-Brain Communication (Wireless Lan)

June 2013 40

Data Exchange Scheme– Timelines

June 2013 41

Robot BrainSTR (X,Y, Orientation)#

GRD (Center, Size)#

LGD (X0 – Xsize-1) $

RQP (X,Y)#

PTH [0 … n]$

June 2013 42

Brain Brain

BLG (X0 – Xsize-1) $

System Constraints

June 2013 43

• Cell Size equal to Object Size (Fixed).• 4-direction Robot motion only.• Uncertainty in Localization.

Experiments | Obstacle Detection/Avoidance

June 2013 44

Experiments |Path Executer

June 2013 45

Experiments

Conclusion

June 2013 46

• Using swarm with applying Anti-Flocking solves restrictions of frontier-based approach instead of using complex solutions like SLAM techniques.

Future Work

June 2013 47

• Robot Motion Planning in Eight Directions using Voronoi diagram.

• Dynamic Path finding using D* Lite.• Monte-Carlo localization (Markov

localization).• Migration of the AI Agent to the Cloud

(Cloud Robotics).

Tools And Technologies

June 2013 48

• Exploration Simulator – C++/CLI, C# windows Application– Microsoft Visual Studio 2010

• Embedded Software (Robot Framework):– C on AVR Studio

• ASANAA: Team Management System

• Github, Bitbucket: Source Control.

Future Work

References

June 2013 49

[1] Sebastian Thrun. Probabilistic Robotics, (1999-2000) [2] R.Murphy. An introduction to AI Robotics.MIT Press, 2000[3] Yamauchi, B. A frontier-based approach for autonomous exploration. In Proceedings of the 1997 IEEE International Symposium on Computational Intelligen ce in Robotics and Automation, (Monterey, CA, July 1997), pp. 146-151.[4] Yamauchi,B . Frontier-Based Exploration Using Multiple Robots.In the Proceedings of the Second International Conference on Autonomous Agents (Agents ‘98), Minneapolis, MN,May 1998, ACM Press.[5] Yun-Qian Miao, Alaa Khamis, Mohamed Kamel, "Applying Anti-Flocking Model in Mobile Surveillance Systems," Special Session on Distributed Surveillance Systems, International Conference on Autonomous and Intelligent Systems (AIS 2010), June 21-23, 2010, Povoa de Varzim, Portugal [Acceptance rate: 68%]. [6] Martin C. Martin Hans P. Moravec , Robot Evidence Grids,Tech. report CMU-RI-TR-96-06, Robotics Institute, Carnegie Mellon University, March, 1996.[7] Angie Shia , Survey of Swarm Robotics Techniques – A Tutorial,Special Topics Chair IEEE RAS Region 6[8] Don Miner ,Swarm Robotics Algorithms: A Survey (2007)[9] Miguel Juliá.Arturo Gil .Oscar Reinoso, A comparison of path planning strategies for autonomous exploration and mapping of unknown environments.[10] A Behavior-Based Strategy for Single and Multi-Robot Autonomous Exploration,Sensors 2012,13 July 2012.[11] Alaa Khamis , Motion Planning in robots , lectures notes at GUC.

[Gp][final seminar][presentation]

Technology

Gp - Presentation com

Gp Presentation(Imtiyaz)

gp presentation-1

GP Hot Topics Presentation

Seminar gp milan 2011 pre-match agreements

Seminar gp indianapolis_2011_perfect_judge

Gp for schools presentation

3 GP Presentation

Seminar 2(Gp 12)

Seminar presentation-seminar

Rm Presentation Gp 3

1st GP Seminar

DESP GP Presentation

Seminar gp milan_2011_innistrad_dissected

Copy of Gp Presentation

Gp&Sp Presentation Slide-bi