TELEOPERATION OF A MOBILE ROBOT VIA UMTS LINK

Florian Zeiger, Markus Sauer, Lothar Stolz

Zentrum f¨ur Telematik e.V., Allesgrundweg 12, Gerbrunn, Germany

Klaus Schilling

Institute of Computer Science, University of W¨urzburg, Am Hubland, W¨urzburg, Germany

Keywords:

Networked robotics, Teleoperation, UMTS, Wireless teleoperation, Wireless networks of robots.

Abstract:

Nowadays available broadband wireless communication technologies offer a broad spectrum of applications

and the todays availability of UMTS technology has a great potential in the area of networked robotics. This

work investigates teleoperation of mobile robots via UMTS communication links. First, several scenarios

with artiﬁcially generated command trafﬁc between mobile robot and operator are analyzed in order to gain

knowledge of the link behavior. In a next step real mobile robot hardware is remote controlled via UMTS.

Thereby, the robot transmits sensor data and a video stream of its onboard camera while receiving commands

from the operator. The focus of this work is set on the measurement and analysis of the round trip time and the

packet inter-arrival time of data in different scenarios like the connection of two UMTS nodes, the connection

of an Internet PC and an UMTS node, or the behavior of several UMTS nodes generating a large amount

of data. The results clearly show how applications for mobile robot teleoperation can efﬁciently use UMTS

communication in order to allow seamless teleoperation between operator and hardware in distant locations.

1 INTRODUCTION

Communication is a very important issue in the area

of mobile robotics. This is especially the case in

networked multi-robot systems or joint human-robot

teams. Although many advances in autonomy have

been made in several applications like e.g. search

and rescue still direct teleoperation might be needed.

For direct teleoperation (Fong and Thorpe, 2001) the

communication link characteristics are even more im-

portant than for other applications where data might

not be as time-critical. From the human factors point

of view situational awareness (Endsley, 2000) is a

very critical aspect for robot teleoperation and can

be signiﬁcantly inﬂuenced by communication param-

eters e.g. available bandwidth, delays, packet inter-

arrival time, and jitter for the payload data. For direct

teleoperation, often the camera image from the robot

is one of the major feedback elements for the human

operator from the remote environment. Therefore, a

relative high communication bandwidth, low delays

and a small jitter is desirable for these connections

in order to support maintaining situational aware-

ness. Nowadays, in robotics research either wire-

less LAN according to 802.11 (Ollero et al., 2003)

(Zeiger et al., 2008a) or proprietary communication

systems are used (Musial et al., 2001) (Pezeshkian

et al., 2007). The upcoming high-bandwidthnetworks

for mobile phones or mobile Internet like UMTS offer

a new widely used and commercially available tech-

nology with high potential for the application in mo-

bile robot teleoperation. Up to now, the coverage of

these networks has increased in a way that at least

all bigger cities have access to broadband networks.

This everywhere availability in large areas is a ma-

jor advantage for any telematic application compared

to a solution where infrastructure initially has to be

built up and maintenance effort is necessary. In par-

ticular, the application area of service robotics can

largely beneﬁt from these characteristics. However

, the mobile phone networks like UMTS are designed

for different purposes and under different constraints.

Therefore, it is important to investigate the critical pa-

rameters of a communication technology like UMTS

in order to adjust the possible communication param-

eters on the application layer in a way to realize the

optimum usage of this technology.

The remainder of this work is structured as fol-

lows. Section 2 gives a brief introduction to the

UMTS technology. In Section 3, the test setup is de-

121

Zeiger F., Sauer M., Stolz L. and Schilling K.

TELEOPERATION OF A MOBILE ROBOT VIA UMTS LINK.

DOI: 10.5220/0002164901210126

In Proceedings of the 6th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2009), page

ISBN: 978-989-674-000-9

scribed. In Sections 4 and 5 the evaluations of the

tests are given. Also the results of the real mobile

robot hardware teleoperation test is presented. This

work concludes with a discussion of the results and

an outlook in Section 6.

2 UMTS BASICS

The Universal Mobile Telecommunications System

(UMTS) is a European standard for third generation

mobile telephony (3G). It is based on W-CDMA tech-

nology, standardized by the 3GPP, and implemts Eu-

ropean speciﬁcations of IMT-2000 ITU for third gen-

eration cellular radio systems. Although it is not a

worldwide standard, UMTS will probably become the

most popular one among its competitors. UMTS rig-

orously divides its infrastructure into two parts (cf.

Fig. 1). The UMTS Terrestrial Radio Access Net-

work (UTRAN) handles all tasks related to radio and

wireless networking and the Core Network (CN) pro-

vides all user services running via UMTS.

The UTRAN maintains the radio connection to

mobile operators for circuit and packet switched

links. A key to UMTS radio technology is the use

of the Wideband Code Division Multiple Access (W-

CDMA) user multiplexing procedure. W-CDMA

is responsible for ensuring that several participants

can communicate simultaneously using the same fre-

quency channel without interfering each other. The

UTRAN radio provides two different modes of oper-

ation: using UTRA-FDD mode, up- and downlink of

data run on separate frequencies. This mode supports

data rates up to 384 kbit/s. In UTRA-TDD mode,

up- and downlink take place at the same frequency,

but are separated by time slots. UTRA-TDD supports

data rates up to 2 Mbit/s but is quite uncommon com-

pared to UTRA-FDD. Since demand for higher data

rates emerged, UTRAN was enhanced by High Speed

Downlink Packet Access (HSDPA) as well as High

Speed Uplink Packet Access (HSUPA) to support data

rates up to 14.4 Mbit/s on downlink and 5.76 Mbit/s

on uplink. The UTRAN is composed of two basic

architectural components. Base stations (Node Bs)

provide the radio signaling. In addition, there are Ra-

dio Network Controllers (RNCs). Managing all radio

resources, these RNCs administrate a group of Node

Bs. The area covered by all Node Bs connected to

one RNC form a Radio Network Subsystem (RNS).

These RNSs are linked by cross connections of the

RNCs. sThe RNC ﬁnally makes the connection to the

Core Network (CN), too. Inside the CN, there are

several nodes providing various high level services

needed for mobile telecommunication. The Mobile

Switching Center (MSC) serves as a switching node

and gateway for circuit switched connections (phone

calls). In case of packet switched data (IP data), these

functionality is provided by the Serving GPRS Sup-

port Node (SGSN) and Gateway GPRS Support Node

(GGSN). For more details on UMTS please refer to

(Holma and Toskala, 2007).

Figure 1: UMTS architecture.

3 SCENARIO SETUP

The focus of this work is set on teleoperation of a

mobile robot via UMTS communication link and the

characterization of the communicationlink in order to

allow for a seamless teleoperation of the robot. There-

fore, three different test setups are analyzed:

• The connection between one robot connected to

UMTS and a PC connected to the Internet via

DSL (16 MBit/s) (Mode 1).

• The connection of two mobile robots via UMTS

(Mode 2). In this case, only the two robots gener-

ate trafﬁc.

• The connection of two mobile robots via UMTS

while a third UMTS node in the same cell is

transmitting a large amount of data to the Inter-

net (Mode 3).

The mobile clients are represented by mini PCs which

are connected to the Internet via a USB UMTS device

(Huawei 3G modem). This device supports HSDPA

and HSUPA broadband data transmission besides the

normal UMTS mode. This broadband communica-

tion is used for all tests presented in this work. Cur-

rently, all UMTS access providers do not provide pub-

lic IP addresses or in case they do, only a small set

of services is supported. Therefore, each UMTS node

joins a virtual private network using openvpn(cf. Fig-

ure 1). Thus, a physical and hardware component is

present, and a logical overlay is set on top of this (cf.

ICINCO 2009 - 6th International Conference on Informatics in Control, Automation and Robotics

122

Fig. 1) which enables easy addressing of each mo-

bile node. Also the provided data services can be de-

ﬁned according to the correspondingtest. Usually, the

UMTS speciﬁcation promises a reliable packet trans-

mission every 20 milliseconds. To get an idea of the

link behavior, for each of the three above mentioned

modes, data streams of different sizes are generated.

As measurement categories, the packet inter-arrival

time is analyzed as this is an important criterion for

video and sensor data transmission and the round trip

time (rtt) is investigated as this two-way delay is also

very important in case the mobile robot is teleoperated

directly by a human operator. For the packet inter-

arrival time analysis, data is transmitted only one-way

and the packet inter-arrival time is plotted at the des-

tination node. Therefore, data is generated with a

packet size of 2048 bytes and a packet inter-departure

time of 10, 20, 50, and 100 milliseconds which results

in packet streams of 20, 40, 100, and 200 kb/s. The

round trip time measurements use ICMP ping packets

with the size manually set to 2040 bytes of payload

and 8 bytes ICMP header. Here, the send intervals

for ping packets are also set to 10, 20, 50, and 100

ms which generates data streams comparable to the

above described tests for the packet inter-arrival time.

Of course, the data stream for the round trip time tests

are transmitted in both directions.

4 EVALUATION

4.1 Packet Inter-arrival Time

For this analysis, the packet inter-arrival time at the

destination node is measured. Also the number of

packets is counted. The results are displayed in his-

tograms with a bin size of 2 milliseconds. The small-

est bin holds all values between 0 and its own value,

and the bin with the highest value holds also all larger

values up to inﬁnity. The x-axis shows the packet

inter-arrival time in seconds and the y-axis shows the

relative frequency of occurrence. As the data set of

the measurements contains enough data to prove that

the resulting distribution is stable, the relative fre-

quency of occurrence can be considered as equal to

the probability of occurrence of the corresponding

packet inter-arrival time.

4.1.1 Connection between Internet PC and

UMTS Node

For this setup, the packet inter-arrival time is mea-

sured in both directions for packet streams of 20, 100,

and 200 kb/s. Figure 2a shows the results for the 20

kb/s stream. The packet inter-departure time for this

stream is 100 milliseconds. The upper subplot of Fig-

ure 2a shows, that more than 50% of the packets arrive

with an inter-arrival time of 100 ms. Approximately,

another 10% arrive with an inter-arrival time of 90

ms and 110 ms respectively. For the opposite direc-

tion – from UMTS node to Internet PC (cf. lower

subplot of Figure 2a)– the result is completely dif-

ferent. Here, the packet inter-arrival time is closely

distributed around the expected packet value of 100

milliseconds.

0.08 0.09 0.1 0.11 0.12

0.2

0.4

0.6

Internet to UMTS 20 kbytes/sec

packet inter−arrival time [s]

probability of occurence

0.08 0.09 0.1 0.11 0.12

0.2

0.4

0.6

UMTS to Internet 20 kbytes/sec

packet inter−arrival time [s]

probability of occurence

(a)

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

0.2

0.4

0.6

Internet to UMTS 100 kbytes/sec

packet inter−arrival time [s]

probability of occurence

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

0.2

0.4

0.6

UMTS to Internet 100 kbytes/sec

packet inter−arrival time [s]

probability of occurence

(b)

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

0.1

0.2

0.3

0.4

Internet to UMTS 200 kbytes/sec

packet inter−arrival time [s]

probability of occurence

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

0.1

0.2

0.3

0.4

UMTS to Internet 200 kbytes/sec

packet inter−arrival time [s]

probability of occurence

(c)

Figure 2: Packet inter-arrival time between Internet PC and

UMTS node and a data bandwidth of 20 kb/s (Fig. 2a) , 100

kb/s (Fig. 2b) and 200 kb/s (Fig. 2c).

For the stream with a bit rate of 100 kb/s, the re-

sults between the both investigated transmission di-

rections vary signiﬁcantly (see Figure 2b). At the

TELEOPERATION OF A MOBILE ROBOT VIA UMTS LINK

123

UMTS node, more than 50% of the transmitted pack-

ets arrive with an inter-arrival-time of 20 ms and two

other peaks with 10% each are visible at about 10

ms and 30 ms. For the packet inter-arrival time for

the transmission direction from UMTS node to Inter-

net PC, only about 13% of the packets have an inter-

arrival time of 20 ms. Here, also small peaks at 10ms

and 30 ms (> 6% each), as well as a high amount of

packets with 5 ms or less (> 13%) and with 40 ms

or more (> 7%) are present. Also for the 200 kb/s

stream, the results for both directions are quite differ-

ent (cf. Figure 2c). For the transmission from Inter-

net PC to UMTS node more than 40% of the packets

have an inter-arrival time of 10 ms which corresponds

to the packet inter-departure time. More than 20% of

the packets have an inter-arrival time of 4ms and less.

For the other transmission direction – from UMTS to

internet PC – only about 20% of the packets arrive

with the expected inter-arrival time of 10 ms. A sec-

ond peak of about 18% is present for an inter-arrival

time of 4 ms and less. Most of the other packets are

distributed to inter-arrival times between 4 and 36 ms.

4.1.2 UMTS to UMTS Connection

This section shows the results of transmissions be-

tween two UMTS nodes. Each upper subplot of Fig-

ures 3a, 3b, and 3c show the packet inter-arrival times

when trafﬁc is transmitted only between the two in-

volved nodes. The lower subplots of these Figures

show the inter-arrivaltime of the packets while a third

UMTS node transmits a large ﬁle to an Internet PC.

Thus, this data stream must share the UMTS cell

capacity with the measured data stream. Figure 3a

shows the results for the 20 kb/s stream. Here, both

setups show similar results. The majority of the pack-

ets has an inter-arrival time of 100 ms (33% without

additional trafﬁc and 28% with additional trafﬁc) and

almost all other inter-arrival times are distributed in

10 ms intervals at 80 ms, 90 ms, 110 ms, and 120 ms.

For this stream, the additionally generated trafﬁc re-

duces the amount of the packets with an inter-arrival

time of 100 ms for approximately 5%. In Figure 3b

the results are displayed for the 100 kb/s stream with

a packet inter-departure time of 20 ms. In both situa-

tions, almost 20% of the packets have an inter-arrival

time of 20 ms. In case of no additionally generated

trafﬁc, more than 55% of the packets arrive with an

inter-arrival time of less than 10 ms. In case addi-

tional trafﬁcis generated, only about 42% of the pack-

ets have an inter-arrival time of 10 ms or less. The

remaining packet inter-arrival times are distributed at

peaks around 30 ms and 40 ms. The results of the 200

kb/s stream displayed in Figure 3c are similar to the

above described observations for the 100 kb/s stream.

0.08 0.09 0.1 0.11 0.12

0.1

0.2

0.3

0.4

UMTS to UMTS 20 kbytes/sec

packet inter−arrival time [s]

probability of occurence

0.08 0.09 0.1 0.11 0.12

0.1

0.2

0.3

0.4

UMTS to UMTS 20 kbytes/sec and additional traffic

packet inter−arrival time [s]

probability of occurence

(a)

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

0.1

0.2

0.3

0.4

UMTS to UMTS 100 kbytes/sec

packet inter−arrival time [s]

probability of occurence

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

0.1

0.2

0.3

0.4

UMTS to Internet 100 kbytes/sec and additional traffic

packet inter−arrival time [s]

probability of occurence

(b)

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

0.1

0.2

0.3

0.4

UMTS to UMTS 200 kbytes/sec

packet inter−arrival time [s]

probability of occurence

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

0.1

0.2

0.3

0.4

UMTS to Internet 200 kbytes/sec and additional traffic

packet inter−arrival time [s]

probability of occurence

(c)

Figure 3: Packet inter-arrival time between UMTS nodes

and a data bandwidth of 20 kb/s (Fig. 3a), 100 kb/s (Fig.

3b), and 200 kb/s (Fig. 3c).

The additionally generated trafﬁc reduces the amount

of packets at the expected inter-arrival time of 10 ms

for about 5% and the remaining packets are again lo-

cated at the bins with 10 ms inter-bin distance and

inter-arrival times of less than 4 ms, 20 ms, 30 ms,

and 40 ms.

Finally, the effective receiving bit rates of the pay-

load data is shown in Tables 1a and 1b. Table 1a

shows the results of the test runs between UMTS node

and Internet PC and Table 1b shows the results of the

test between two UMTS nodes without additionally

generated trafﬁc (Mode 2) and between two UMTS

nodes with additionally generated trafﬁc (Mode 3).

These two tables give an idea of the present packet

ICINCO 2009 - 6th International Conference on Informatics in Control, Automation and Robotics

124

Table 1: Resulting effective payload bit rates at the receiv-

ing node between UMTS node and Internet PC (Table 1a)

and for Mode 2 and Mode 3 (Table 1b).

sending data rate receiving at UMTS node receiving at Internet PC

20 kbytes/s 19.20 kbytes/s 19.58 kbytes/s

100 kbytes/s 98.09 kbytes/s 92.62 kbytes/s

200 kbytes/s 194.33 kbytes/s 162.82 kbytes/s

(a)

sending data rate receiving Mode 2 receiving Mode 3

20 kbytes/s 16.19 kbytes/s 19.31 kbytes/s

100 kbytes/s 94.03 kbytes/s 92.53 kbytes/s

200 kbytes/s 141.80 kbytes/s 125.61 kbytes/s

(b)

loss during the test runs. Surprisingly good results

are achieved for all data streams during the transmis-

sion from Internet PC to UMTS node. In the oppo-

site direction, an increased packet loss is observed for

the 200 kb/s stream. For Mode 2 and Mode 3, where

two UMTS nodes communicated with each other, the

packet loss was signiﬁcantly higher.

4.2 Round Trip Times

Figure 4 shows the measured rtts for the generated

packet streams. Mode 1 corresponds to the Internet-

UMTS node scenario, Mode 2 is the transmission be-

tween two UMTS nodes without additionally gener-

ated trafﬁc and for Mode 3, the data exchange of two

UMTS nodes is measured while a third node trans-

mits additional data to an Internet PC. These box-

plots show the median of the measured values (hor-

izontal line) and the box shows 50% of the values

boundedby the lower 25% quartile and the upper 75%

quartile. Furthermore, lines indicate the most extreme

values within 1.5 times the interquartile range from

the ends of the box and extreme values lying out-

side this borders are marked with crosses. For the

three streams with 20 kbyte/s, 50% of the measured

rtt values are distributed very close to the correspond-

ing median. A similar observation can be made for

the 40 kbyte/s stream in Mode 1 and Mode 3, and for

the 100 kbyte/s stream in the Mode 1 scenario. As ex-

pected, the largest variations appear for the high band-

width streams with 200 kbyte/s and the 100 kbyte

stream in Mode 3 with the additionally generated traf-

ﬁc. In general, 50% of the measured rtt values of each

test run are located close to the corresponding me-

dian. Only few extreme values are measured below

the lower 25% quartile border and sometimes, very

high rtt values are measured above the upper border of

the 75% quartile (e.g. for 100 kb/s in Mode 3). Thus,

the later used application to mobile robotics must con-

sider these high round trip times which occur some-

times and must be able to handle this large variability.

5 REAL HARDWARE TEST

To analyze the UMTS link under real conditions for

mobile robot teleoperation, a Pioneer 3-AT is tele-

operated from a Laptop via UMTS. The robot plat-

form is equipped with a mini PC and a network video

camera AXIS 221 and transmits a motion JPEG video

stream with a resolution of 320x240 and a frame rate

of 15 frames per second. The Player version 2.0.4

software is used to interface the hardware and for

providing communication between the client and the

hardware. The client is a Java program which pro-

vides a video image of the mobile robot’s onboard

camera and displays sensor data. During the test runs,

the mobile robot generates a sensor data stream and

sends it to the Laptop. Additionally, video data is sent

each 66.66 ms. The Client program is sending control

commands with a bit rate of 1.2 kb/s. In Figure 5, the

behavior of the video data which arrives at the oper-

ator’s PC is shown. The video source generates rela-

tively large packets which must be fragmented. Thus,

the send buffers are continuously ﬁlled and packets

are sent as often as possible. On the receiver’s side,

the already above observed peaks of the packet inter-

arrival times each 10 ms is also visible.

0 0.02 0.04 0.06 0.08 0.1 0.12

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

packet inter−arrival time [s]

relative frequency of occurence

Figure 5: Packet inter-arrival of the video data coming from

the mobile robot.

6 DISCUSSION OF THE RESULTS

The above presented results show clearly that UMTS

is a well suited communication technology for the

teleoperation of mobile robots. The results obtained

in the initial test runs with the artiﬁcially generated

command trafﬁc give a detailed overview of the chan-

nel behavior. When data is transmitted from the

UMTS node via the UMTS infrastructure and the In-

ternet, the arriving packets at the PC have a higher

variance in the packet inter arrival time as typically,

the Internet data transfer induces such a variance. In

the opposite direction, the packets arrive at the UMTS

infrastructure via the Internet. Then the UMTS infras-

tructure is responsible for further data transmission

TELEOPERATION OF A MOBILE ROBOT VIA UMTS LINK

125

200 kb/s (1) 200 kb/s (2) 200 kb/s (3) 100 kb/s (1) 100 kb/s (2) 100 kb/s (3) 40 kb/s (1) 40 kb/s (2) 40 kb/s (3) 20 kb/s (1) 20 kb/s (2) 20 kb/s (3)

200

400

600

800

1000

1200

1400

1600

1800

2000

round trip time [ms]

generated traffic stream

Figure 4: Round trip times in milliseconds (Mode (1): Internet-UMTS; Mode (2): UMTS-UMTS; Mode (3): UMTS-UMTS

and additional trafﬁc).

and takes care that packets are transmitted with de-

ﬁned packet inter-departuretimes. These packet inter-

departure times are depending on the used bandwidth

and the used HSDPA/HSUPA mode and some other

aspects depending on the provider. Nevertheless, the

measured packet inter-arrival times are now bound to

ﬁxed times which reduces the variance. In general, it

can be seen that as soon as the broadband communi-

cation mode is entered, data is delivered at the mobile

robot and at the operator’s PC constantly. Noticeable

outcomes are the frequent packet inter-arrival times

at 10 ms and multiple of 10 ms with a relatively high

peak at the 20 ms inter-arrival times. Also, the re-

ceived effective payload data rates give a clear view

of the present packet loss during the tests. Finally, the

measured round trip times and their distribution lead

to the following hints for implementing mobile robot

teleoperation. The ”just plug and try” setup which

is described in Section 5 shows to be already usable.

Nevertheless, a deﬁned trafﬁc shaping might be a suit-

able approach to use the characteristics of the UMTS

link more efﬁcient and to increase the quality of tele-

operation (e.g. better video quality or less packet

loss). Approaches for this idea are already published

in another context (Zeiger et al., 2008c)(Zeiger et al.,

2008b) and shows to be a useful technique. Neverthe-

less, UMTS is a promising technology to allow broad-

band communication for hardware teleoperation and

will be in the research focus of networked robotics

in near future. Future work will be set on analyzing

more details of the UMTS link in combination with

mobile robot teleoperation. Open issues are how to

use the behavior of the communication channel in or-

der to allow for high quality onboard video transmis-

sion together with reliable control data transmission

in both directions.

ACKNOWLEDGEMENTS

This work is partially supported by the European

Union and the Federal State of Bavaria in the frame

of the EFRE program.

REFERENCES

Endsley, M. R. (2000). Theoretical underpinnings of sit-

uation awareness: A critical review. In Endsley,

M. R. and Garland, D. J., editors, Situation aware-

ness analysis and measurement, chapter 1, pages 3–

26. Lawrence Erlbaum Associates.

Fong, T. and Thorpe, C. (2001). Vehicle teleoperation inter-

faces. Auton. Robots, 11(1):9–18.

Holma, H. and Toskala, A., editors (2007). WCDMA for

UMTS: HSPA Evolution and LTE. John Wiley& Sons,

Ltd, fourth edition.

Musial, M., Brandenburg, U. W., and Hommel, G. (2001).

Success of an Inexpensive System Design: The Flying

Robot MARVIN. In 16th Int. Unmanned Air Vehicle

System Conf. (UAVs).

Ollero, A., Alcazar, J., Cuesta, F., Lopez-Pichaco, F., and

Nogales, C. (2003). Helicopter teleoperation for aerial

monitoring in the comets multi-uav system. In 3rd

IARP Workshop on Service, Assistive and Personal

Robots (IARP 2003), Madrid (Spain).

Pezeshkian, N., Nguyen, H. G., and Burmeister, A. (2007).

Unmanned ground vehicle radio relay deployment

system for non-line-of-sight operations. In Proc. of

the 13th IASTED Int. Conf. on Robotics and Applica-

tions, August 29-31, W¨urzburg, Germany, RA2007.

Zeiger, F., Kr¨amer, N., and Schilling, K. (2008a). Com-

manding mobile robots via wireless ad-hoc networks

- a comparison of four ad-hoc routing protocol imple-

mentations. In IEEE Int. Conf. on Robotics and Au-

tomation (ICRA 2008).

Zeiger, F., Sauer, M., and Schilling, K. (2008b). Intelligent

Shaping of a Video Stream for Mobile Robot Tele-

operation via Multihop Networks in Real-World Sce-

narios. In Proc. of IEEE/RSJ Int. Conf. on Intelligent

Robots and Systems (IROS 2008).

Zeiger, F., Sauer, M., and Schilling;, K. (2008c). Video

Transmission with Adaptive Quality based on Net-

work Feedback for Mobile Robot Teleoperation in

Wireless Multi-Hop Networks. In 5th Int. Conf. on In-

formatics, Automation and Robotics (ICINCO 2008).

ICINCO 2009 - 6th International Conference on Informatics in Control, Automation and Robotics

126