w/ibo project - wi-fi over aibo => slides - reant.netreant.net/tcd/wibo_project/pdf/wibo project...

32
W/BO project W/BO project Wi-Fi over AIBO Wi-Fi over AIBO 2005-2006 2005-2006 RÉANT Gilles RÉANT Gilles French Erasmus student French Erasmus student

Upload: vuongtram

Post on 07-Apr-2018

227 views

Category:

Documents


3 download

TRANSCRIPT

W/BO projectW/BO projectWi-Fi over AIBOWi-Fi over AIBO

2005-20062005-2006RÉANT GillesRÉANT GillesFrench Erasmus studentFrench Erasmus student

SummarySummary

Gilles RÉANTThursday 6 April 2006

● aims of the project● project review● tools used● evolution of the algorithm● result● demonstration● problems encountered● next● questions

Aims of this projectAims of this project

see what's possible using Wi-Fi & the AIBO

Gilles RÉANTThursday 6 April 2006

● which equipment is compatible?

● what can we do?

● security?

Wi-Fi sniffer

the W/IBO project in itself

Gilles RÉANTThursday 6 April 2006

Aims of this projectAims of this project

Open-R SDK

another idea

finding the access point to which it is connected

report

● avoiding the inappropriate SDKs

● setting-up a connection between a computer & AIBO

● starting doing basic movements

Gilles RÉANTThursday 6 April 2006

Aims of this projectAims of this project

● good start for using the AIBO

● in 2 hours

Work from October 2005 to April 2006

● reportGilles RÉANTThursday 6 April 2006

Project reviewProject review● AIBO● Wi-Fi● tests of coding

● wireless connection for telnet

● infrastructure mode

● SDK● algorithm

Tools usedTools used

Mind 2

wireless access point

AIBO ERS-7

computer

Thursday 6 April 2006 Gilles RÉANT

R-CodeR-Code

Gilles RÉANTThursday 6 April 2006

Tools usedTools used

tests

● important result: effect(orientation) >> effect(distance)

pentagon

first implementation of the rotation

second implementation of the rotation

finding something else

Gilles RÉANT

Evolution of the algorithmEvolution of the algorithm

Thursday 6 April 2006

approximately constant angle

recording the last measurements

if in local minimum, than forward!

Evolution of the algorithmEvolution of the algorithm

Thursday 6 April 2006 Gilles RÉANT

Evolution of the algorithmEvolution of the algorithmother problems

● scattered measures

● solution 1● solution 2

● low mean

● remaining problems● what if we stay in the problem state for too long?● reflection of the waves

Thursday 6 April 2006 Gilles RÉANT

ResultResult

not so bad in noisy environment + low power signal

better results with

● one single signal

● high power

● no reflection of the waves

Thursday 6 April 2006 Gilles RÉANT

DemonstrationDemonstration

● 100% alone

● values monitored on the PC

Thursday 6 April 2006 Gilles RÉANT

???

Problems encounteredProblems encountered

Thursday 6 April 2006 Gilles RÉANT

URBI

learning from scratch

script

numerous bugs

not fully implemented

ad-hoc

connection to the CS wireless network

NextNext

what can be improved?

what will be done?

& after?

Thursday 6 April 2006 Gilles RÉANT

Slides done with Open Office.org Impress 2.0.2

THANKSTHANKS

Thursday 6 April 2006 Gilles RÉANT

02/05/06 Gilles RÉANT 1

W/BO projectW/BO projectWi-Fi over AIBOWi-Fi over AIBO

2005-20062005-2006RÉANT GillesRÉANT GillesFrench Erasmus studentFrench Erasmus student

Good afternoon, thanks for comingMy name is Gilles RÉANT, I'm a Senior Sophister

Erasmus student from France, & I'm going to talk to you about my final year project, name the W/BO project

02/05/06 Gilles RÉANT 2

SummarySummary

Gilles RÉANTThursday 6 April 2006

● aims of the project● project review● tools used● evolution of the algorithm● result● demonstration● problems encountered● next● questions

02/05/06 Gilles RÉANT 3

Aims of this projectAims of this project

see what's possible using Wi-Fi & the AIBO

Gilles RÉANTThursday 6 April 2006

● which equipment is compatible?

● what can we do?

● security?

PC under Windows & LinuxMacPocket PCLaptopSmartphones

inner web servertelnetfull pack of functionalities on Windows:

● play CDs● record videos● create precoded movements● see through the camera● text to speech● ...

02/05/06 Gilles RÉANT 4

Wi-Fi sniffer

the W/IBO project in itself

Gilles RÉANTThursday 6 April 2006

Aims of this projectAims of this project

Open-R SDK

another idea

finding the access point to which it is connected

built-in wireless 802.11b cardbased on an idea from my tutorget, no set1 single connexion defined in the memory card

02/05/06 Gilles RÉANT 5

report

● avoiding the inappropriate SDKs

● setting-up a connection between a computer & AIBO

● starting doing basic movements

Gilles RÉANTThursday 6 April 2006

Aims of this projectAims of this project

● good start for using the AIBO

● in 2 hours

in a word, preventing other people from loosing as much time as I lost

02/05/06 Gilles RÉANT 6

Work from October 2005 to April 2006

● reportGilles RÉANTThursday 6 April 2006

Project reviewProject review● AIBO● Wi-Fi● tests of coding

● wireless connection for telnet

● infrastructure mode

● SDK● algorithm

choose AIBOthemea few testsad-hoc doesn't work

02/05/06 Gilles RÉANT 7

Tools usedTools used

Mind 2

wireless access point

AIBO ERS-7

computer

Thursday 6 April 2006 Gilles RÉANT

latest generation of AIBOmind 2 just for starting test, can't use mind & coding

your own software

02/05/06 Gilles RÉANT 8

R-CodeR-Code

Gilles RÉANTThursday 6 April 2006

Tools usedTools used

finally chose URBI

02/05/06 Gilles RÉANT 9

tests

● important result: effect(orientation) >> effect(distance)

pentagon

first implementation of the rotation

second implementation of the rotation

finding something else

Gilles RÉANT

Evolution of the algorithmEvolution of the algorithm

Thursday 6 April 2006

Easy to show the live variation of the signal with URBI

first important result: the effect of the orientation of the AIBO is much more important than the distance to the access point

so, thought of a pentagon tesselation of the areastop, making a measurement, rotating of 72°, 5 times,

then decide which is the best direction to takedidn't give any good result, because the algorithm of

rotation precoded in URBI gave a approximate rotation (the robot wasn't rotating around itself)

used another algorithm, which couldn't rotate of a precise enough angle (imprecision of more than 20%)

so I had to give up this idea, & find something else

02/05/06 Gilles RÉANT 10

approximately constant angle

recording the last measurements

if in local minimum, than forward!

Evolution of the algorithmEvolution of the algorithm

Thursday 6 April 2006 Gilles RÉANT

recording the 5 last measurements

02/05/06 Gilles RÉANT 11

Evolution of the algorithmEvolution of the algorithmother problems

● scattered measures

● solution 1● solution 2

● low mean

● remaining problems● what if we stay in the problem state for too long?● reflection of the waves

Thursday 6 April 2006 Gilles RÉANT

02/05/06 Gilles RÉANT 12

ResultResult

not so bad in noisy environment + low power signal

better results with

● one single signal

● high power

● no reflection of the waves

Thursday 6 April 2006 Gilles RÉANT

02/05/06 Gilles RÉANT 13

DemonstrationDemonstration

● 100% alone

● values monitored on the PC

Thursday 6 April 2006 Gilles RÉANT

???

the access point is herethe AIBO will stop when the signal's power reaches

80% (Sony's unit)first time I do this in this room, can't assure that it will

work

02/05/06 Gilles RÉANT 14

Problems encounteredProblems encountered

Thursday 6 April 2006 Gilles RÉANT

URBI

learning from scratch

script

numerous bugs

not fully implemented

ad-hoc

connection to the CS wireless network

02/05/06 Gilles RÉANT 15

NextNext

what can be improved?

what will be done?

& after?

Thursday 6 April 2006 Gilles RÉANT

02/05/06 Gilles RÉANT 16Slides done with Open Office.org Impress 2.0.2

THANKSTHANKS

Thursday 6 April 2006 Gilles RÉANT