19.01.2014 Views

Robot Control - ICMCC

Robot Control - ICMCC

Robot Control - ICMCC

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Optimisation Issues of High Throughput<br />

Medical Data and Video Streaming Traffic<br />

in 3G Wireless Environments<br />

R. S. H. ISTAPANIAN* and N. PHILIP*<br />

* Mobile Information and Network Technologies Research Centre,<br />

Kingston University, London<br />

Kingston upon Thames, KT1 2EE, UK<br />

r.istepanian@kingston.ac.uk.


M- Health : Emerging Mobile Communications<br />

and Network Technologies for healthcare<br />

Istepanian (etal.) , IEEE Trans. Information Technologies in Biomedicine,<br />

Vol. 8, 4, DEc. 2004.


Introduction:<br />

• OTELO system Medical and non medical data.<br />

• still Images<br />

• <strong>Robot</strong> <strong>Control</strong><br />

• Medical video streaming<br />

• video conference<br />

• 3G QoS for Ultrasound requirements.<br />

• Soft Computing and Optimization issues in 3G – medical QoS<br />

(m-QoS) will be discussed here and especially Multiobjective<br />

optimization methodology based on Adaptive Simulated<br />

Annealing (ASA).


OTELO System – Introduction<br />

EU-IST Funded Project (2001-2004)


Expert station - Input Device<br />

To <strong>Control</strong> the Patient station<br />

robot:<br />

• One DoF hand free Input device<br />

• Easy to use, comparable to<br />

standard probe<br />

• 6DOF localisation sensor - Flock of<br />

Bird TM


Patient Station<br />

<strong>Robot</strong> <strong>Control</strong>ler Unit<br />

Videoconference<br />

System<br />

<strong>Robot</strong>ic Head &<br />

US Probe Holder<br />

Ultrasound<br />

Machine


EXPERT & PATIENT Stations


Ultrasound Holder (<strong>Robot</strong>ic Arm)


OTELO Medical & Non-Medical Data Requirements<br />

Expert Station<br />

• Video-conference<br />

3G Uplink<br />

• <strong>Robot</strong> <strong>Control</strong><br />

• Video-conference<br />

• <strong>Robot</strong> <strong>Control</strong><br />

• Force feedback<br />

• Dynamic US images<br />

• Still US Images<br />

3G Downlink<br />

• Fictive Probe force<br />

Patient station


Table 1. OTELO medical data requirements and corresponding data rates.<br />

Ultrasound video<br />

stream<br />

Ultrasound still<br />

images<br />

Ambient video<br />

stream<br />

Voice<br />

<strong>Robot</strong> control<br />

data<br />

Flow direction<br />

Simplex:<br />

Patient to Expert<br />

Simplex:<br />

Patient to Expert<br />

Duplex Duplex duplex<br />

Transport<br />

Protocol<br />

Speed<br />

Requirement<br />

RTP/UDP/IP TCP/IP RTP/ UDP/IP RTP/UDP/IP UDP/IP<br />

Real-time Non Real-time Real-time Real-time Real-time<br />

Payload data<br />

rate<br />

requirement<br />

over the airinterface<br />

(without protocol<br />

headers)<br />

15 frames/s<br />

@ 210 kbit/s<br />

Uplink<br />

1 frame/ 10s<br />

Uplink<br />

15 to 1 frame/s<br />

symmetrically<br />

16 kbit/s<br />

symmetrically<br />

0.3 kbit/s<br />

symmetrically<br />

The detailed medical and non-medical OTELO data traffic are shown in Table 1.<br />

From this table it can be seen that the most traffic and bandwidth demanding<br />

OTELO traffic type is the medical video streaming traffic of the system.


OTELO Functional Modalities and Data Follow<br />

Medical Data<br />

Still US Images<br />

Stream US<br />

Images<br />

Stream US<br />

Images<br />

<strong>Robot</strong>ic Arm control<br />

Frequency<br />

Medium Qlty.<br />

Videoconf<br />

High Qlty.<br />

Videoconf.<br />

Data<br />

Description<br />

Gray Scale,<br />

512x512 Pixel<br />

Gray Scale,<br />

CIF (Regon Of Interest),<br />

200x200 Pixel<br />

Gray Scale,<br />

CIF, 352x288 Pixel<br />

Data Rates<br />

&<br />

Resolution<br />

14-97<br />

Kbyte<br />

Data Flow<br />

Direction<br />

Up-link<br />

P-to-E<br />

10 fps Up-link<br />

P-to-E<br />

7 fps Up-link<br />

P-to-E<br />

100-200 Hz 3-4 Kbps UP or Down<br />

Link, E-to-P, P-to-E<br />

QCIF<br />

176x144<br />

QCIF<br />

176x144<br />

CIF<br />

352x288<br />

5 fps Up & Down<br />

P-to-E, E-to-P<br />

7.5 fps Up & Down<br />

P-to-E, E-to-P<br />

5 fps Up & Down<br />

P-to-E, E-to-P


Transmitted image<br />

Received image<br />

2 cm hyperechoic focal lesion in<br />

deeper part of the right liver


Therefore we should implement ‘some control mechanism’ that will<br />

adhere within these bounds:<br />

1- OTELO’s user (QoS) requirement<br />

2- the Network QoS requirements<br />

From the Clinical Trials of the system and the experts<br />

evaluation of the diagnostic quality of the received US<br />

image and stream, that the communication bounds<br />

for efficient diagnosis using the OTELO system are:<br />

1. Frame rate of the received images >= 5fps ( kbps)<br />

2. Peak Signal to Noise Ratio (PSNR) >= 36dB<br />

3. The average delay of Probe movement-to-Received<br />

image at the Patient station is < 350 ms<br />

These are called Medical ‘user QoS’ requirements


3G OTELO Performance analysis studies :<br />

1. Evaluate the performance of the OTELO <strong>Robot</strong>ic system<br />

over UMTS network.<br />

2. To study the effect of multiplexed data transmission on<br />

the Real-time performance.<br />

3. Subjective and objective evaluation of the received data<br />

Garawi, (etal.) , ‘ Performance Analysis of a Compact <strong>Robot</strong>ic Tele-Echography<br />

E-Health System over Terrestrial and Mobile CommunicationLinks’, Proc.<br />

5th. IEE International Conference on 3G Mobile Communication<br />

Technologies- 3G 2004, London, 18-20, October, pp.118-122, 2004.


OTELO System on UMTS network<br />

GGSN<br />

SGSN<br />

UTRAN<br />

3G<br />

Core Network<br />

GMSC<br />

PSTN<br />

ISDN or ADSL<br />

RNC<br />

Node B<br />

Node B<br />

- Ambient Video<br />

- Voice<br />

- <strong>Robot</strong> <strong>Control</strong><br />

Downlink:<br />

3G mobile handset<br />

(or PCMCIA card)<br />

(Patient) Mobile<br />

OTELO Station<br />

GateAway<br />

<strong>Robot</strong> <strong>Control</strong><br />

Unit<br />

Modem<br />

Uplink:<br />

- US video stream<br />

or Ambient Video<br />

- US still images<br />

- Voice<br />

- <strong>Robot</strong> <strong>Control</strong><br />

Probe Holder<br />

<strong>Robot</strong><br />

<strong>Robot</strong>ic<br />

<strong>Control</strong> Input<br />

Device<br />

Audio-Video<br />

Device<br />

Expert Station<br />

Audio-Video<br />

Device<br />

Echograph<br />

Device


Ultrasound data stream throughput<br />

Intantaneous recieved throuhput Vs time<br />

Average total<br />

throughput:<br />

19.7 Kbps<br />

Average US video<br />

stream throughput:<br />

12.7 Kbps<br />

Throughput (Kbps)<br />

100<br />

90<br />

80<br />

70<br />

60<br />

50<br />

40<br />

30<br />

20<br />

10<br />

0<br />

<strong>Robot</strong>ic<br />

Data<br />

<strong>Robot</strong>ic Data +<br />

US stream<br />

Average throughput<br />

0 20 40 60 80 100 120 140<br />

Time (s)<br />

Total throughput for the control & ultrasound data received<br />

at the Expert station, UMTS Uplink Data rate = 64Kbps.


Supporting robust video communications over 3G wireless<br />

networks is a significant problem, primarily because of two<br />

factors:<br />

• Low bandwidth.<br />

• Time varying error characteristics of the transmission channel.<br />

This will cause congestion and then degradation in the network<br />

performance:<br />

• utilization<br />

• packet losses<br />

• higher delay<br />

• delay jitter<br />

These are called generally 3G Network QoS.


Adaptive simulated Annealing in the medical 3G network optimization<br />

Initialization (algorithm parameters)<br />

Generate a new<br />

frame rate<br />

Accept or reject<br />

no<br />

accept<br />

yes<br />

Adapt the next<br />

walk to the topography<br />

of the required 3G m-QoS<br />

Reduce temperature<br />

No<br />

3G m-QoS<br />

condition<br />

satisfied<br />

Yes<br />

End


The proposed Optimization Technique:<br />

In this work we introduce Adaptive Simulated Annealing (ASA)<br />

Optimization technique and apply it for this multiobjective<br />

optimization problem. In particular we consider the packet<br />

loss and delay Jitter objectives function (Фi), therefore the<br />

algebraic inequality here are:<br />

Packet loss<br />

Delay Jitter<br />

φ<br />

φ<br />

r ) 1<br />

( ≤ ε<br />

r)<br />

2<br />

( ≤ ε<br />

1<br />

2<br />

ε 1 ε 2<br />

, are the limits of these inequalities.


The proposed Optimization Technique:<br />

These can be combined as one equation that the optimization<br />

techniques will work on to find the optimum rate.<br />

⎧ ⎧φi(<br />

r)<br />

−εi<br />

⎫ ⎫<br />

E( r)<br />

= max⎨max⎨<br />

,0⎬<br />

: i = 1, 2⎬<br />

⎩ ⎩ ωi<br />

⎭ ⎭<br />

Ф is the objective function,<br />

i is the number of QoS metrics elements,<br />

r is the rate control value,<br />

ω is the weightings.


End-to-End System Architecture


Conclusions and future work:<br />

(This is ongoing research)<br />

• The OTELO ultrasound stream is the most traffic demanding .<br />

• New QoS metrics for medical applications<br />

• Multiobjective optimization technique base (ASA) is considered<br />

to find the optimal frame rate that satisfy the QoS of the<br />

received images.<br />

• Network monitoring is used here to feedback the packet loss<br />

and delay jitter using the RTCP receiver report.<br />

• Packet Loss and delay jitter is considered here as our objective<br />

functions. PSNR will be added later to the optimisation problem


Acknowledgments:<br />

EU-IST for funding the OTELO project<br />

Vodafone R&D, UK


THANK YOU<br />

Robert S. H. Istepanian<br />

r.istepanian@kingston.ac.uk<br />

www.kingston.ac.uk/MINT

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!