Incorporation of MATLAB into a Distributed Behavioral Robotics Architecture

flippinjapaneseΤεχνίτη Νοημοσύνη και Ρομποτική

2 Νοε 2013 (πριν από 3 χρόνια και 8 μήνες)

188 εμφανίσεις

Research deploy train evaluate

IROS 2004, Sendai, Japan

Incorporation of MATLAB into a
Distributed Behavioral Robotics
Architecture

A. L. Nelson, L. Doitsidis, M. T. Long,
K. P. Valavanis, and R. R. Murphy

Research deploy train evaluate

IROS 2004, Sendai, Japan

Overview


Introduction


The Distributed Field Architecture


Robots and Hardware


Example Uses


Basic robot sensor error quantification in outside
environments


Waypoint navigation with object avoidance


Conclusions

Research deploy train evaluate

IROS 2004, Sendai, Japan

Introduction


Current robot research demands a versatile
control architecture



Heterogeneous Robots


Outdoor environments


Distributed Control


Autonomous Control


Shared Autonomy

Research deploy train evaluate

IROS 2004, Sendai, Japan

Introduction


Motivation


Provide a unified versatile multirobot research
platform


Support AI and Control Theoretic work


Unify robot control research and development
phases for continuity and reduced
development time


Behavior
-
based robot control architectures


Research deploy train evaluate

IROS 2004, Sendai, Japan

Introduction: Related Work


S. Monteiro, E. Bicho, E. “A dynamical systems approach to
behavior
-
based formation control,” Robotics and Automation,
2002. Proceedings. ICRA '02. IEEE International Conference
on,

vol. 3, 2002, pp. 2606


261.



O. Ewerlid, C. Tidestav and M. Sternad, “Real Time Control
using Matlab and Java,” Nordic Matlab Conference, Stockholm,
October 27
-
28, 1997.



A. L. Nelson, E. Grant, T.C. Henderson, “Evolution of neural
controllers for competitive game playing with teams of mobile
robots,” Journal of Robotics and Autonomous Systems, vol. 46,
no. 3, pp. 135
-
150, Mar 2004.

Research deploy train evaluate

IROS 2004, Sendai, Japan


Distributed


Java
-
based


Descendant of SFX


Behavior based


Hybrid deliberative
reactive
architecture


The Distributed Field Architecture

Research deploy train evaluate

IROS 2004, Sendai, Japan

MATLAB Support


JMatLink Modular
Support for MATLAB


Decupling of client and
server


Modules


MATLAB is shown as
a driver implementation
module

Research deploy train evaluate

IROS 2004, Sendai, Japan

MATLAB Support


MATLAB runs as full
work space


Interpreted functions
and scripts


Workspace command
line strings


All tool boxes


Workspace accessed by
JMatLink with
formatted strings

...

matlink.engPutVariable(engine, “laserData” ,


laserData.readLaser);

matlink.engPutVariable(engine, “gpsData”,

gps.readGps);


matlink.engEvalString(engine, “sensorData.

laserData = laserData”);

matlink.engEvalString(engine, “sensorData.gpsData

= gpsData”);


matlink.engEvalString(engine, “result =


mFunction(sensorData.”);


resultData = engGetVariable(engine, “result”);

....





Research deploy train evaluate

IROS 2004, Sendai, Japan

MATLAB Support


MATLAB Usage modes:



Development Phase


Production Phase

Research deploy train evaluate

IROS 2004, Sendai, Japan

Robots and Hardware

Heterogeneous outdoor robots

Research deploy train evaluate

IROS 2004, Sendai, Japan

Application: Basic Sensor Error Characterization


GPS points and points
calculated from
odometry for an
example linear test
pattern.

-
10

-
5

0

5

10

15

20

-
20

-
15

-
10

-
5

0

5


Start

West
---

East (m)

South
---

North (m)

Odometry

Filtered GPS

Unfiltered GPS

Linear Test Pattern

Research deploy train evaluate

IROS 2004, Sendai, Japan

Application: Basic Sensor Error Characterization

-
18

-
16

-
14

-
12

-
10

-
8

-
6

-
4

-
2

0

2

4

-
12

-
10

-
8

-
6

-
4

-
2

0

2

4

6

8


Start

West
---

East (m)

South
---

North (m)

Rectangular Test Pattern

Odometry

Filtered GPS

Unfiltered GPS

-
30

-
25

-
20

-
15

-
10

-
5

0

5

10

15

20

-
30

-
25

-
20

-
15

-
10

-
5

0

5

10


Start

West
---

East (m)

South
---

North (m)

Circular Test Pattern

Odometry

Filtered GPS

Unfiltered GPS

Research deploy train evaluate

IROS 2004, Sendai, Japan

Ongoing Research


Go to goal with obstacle
avoidance

Research deploy train evaluate

IROS 2004, Sendai, Japan

Example Controller Block Diagram


Research deploy train evaluate

IROS 2004, Sendai, Japan

Additional Experiments


Example: Fuzzy
Control


Multiple Robots


Obstacle avoidance

Research deploy train evaluate

IROS 2004, Sendai, Japan

Conclusions


A robot control architecture for advanced
research was presented


Combined high
-
level control and modeling
environment and a distributed behavior
-
based
architecture


Example usages demonstrate utility of the
overall system presented

Research deploy train evaluate

IROS 2004, Sendai, Japan

Acknowledgements


This work was partially supported by a grant
form ONR, N 000 14
-
03
-
1
-
786 (2132
-
033
-
LO).


L. Doitsidis was partially supported by
“IRAKLITOS fellowships for research from the
Technical University of Crete, EPEAEK II


88727