The 18th Annual IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC'07)

SIMULATION AND BASIC EXPERIMENT OF INTER RADIO SYSTEM

HANDOVER FOR COGNITIVE RADIO
Seishi Hanaoka Junji Koji Satoshi Tetsuhiko
Yamamoto Wakayama Yoshizawa Hirata
Hitachi Ltd., Central Research Laboratory
Tokyo, Japan

ABSTRACT approach and system concept of our cognitive radio system.

In section 3, we describe the system architecture of the
Cognitive radio system consists of multiple wireless accesses
simulator and the testbed system. In the proposed
that cover overlapping area, and cognitive terminals that use
architecture, each cognitive terminal, that has multiple radio
one or more of the wireless accesses simultaneously. In this
systems, operates with a single local IP address. In section 4,
paper, we propose the architecture of the cognitive radio
we show the result of simulation, and finally in section 5, we
system, and the inter-system handover protocols. In the
show the result of experiments, conducted on the testbed
proposed architecture, each cognitive terminal, that has
system, using streaming service and UDP packet
multiple radio systems, operates with a single local IP
transmission.
address. The control sequence and packet format are designed
to achieve fast handover among the radio systems. Based on
II. CONCEPT OF COGNITIVE RADIO
the proposed architecture, we have developed both a
simulator and a testbed system with WiMAX and wireless Recently, research on cognitive radio has become popular
LAN. Through simulation and testbed, we have proved that both in Japan and abroad [7, 8]. Concept of cognitive radio is
stream data can be delivered continuously without any effects shown in Figure 1. In this figure, concept of cognitive radio is
of the handover. Also, we present the evaluation result of the expressed using both frequency domain and time domain.
end-to-end latency on the testbed system. For example, we assume we can communicate with three
systems: System (A), System (B) and System (C).
I. INTRODUCTION In this figure, when the current time is t2, the cognitive
terminal (CT) D communicates using the frequency of System
It is well known that frequency below the 6GHz band is
(A), and when the current time becomes t3, terminal D
suitable for mobile communication system, hence many
changes the wireless communication system from System (A)
wireless communication systems such as 3rd generation
to System (B) to communicate using the frequency of
cellular system and wireless LAN are assigned within this
System(B). In the same way, when the current time becomes
frequency band in Japan. As a consequence, there are not
t4, terminal D changes the wireless communication system
enough spare bands for the future wireless broadband system.
from System (B) to both System (A) and System (C). In this
In this situation, it is needed to use these frequencies, which
example the number of wireless communication systems used
are “finite resources,” in a more efficient manner, including
simultaneously is not limited to one, and the cognitive system
the utilization of multiple wireless communication systems
can transmit and receive data with multiple wireless
with intelligence.
communication systems simultaneously.
In the national report “e-Japan Priority Policy Program -
In our cognitive radio system, we assume the use of multiple
2004”[1], a system that recognizes the surrounding radio
frequencies of multiple wireless communication systems.
condition and that adapts proper radio resources, such as
Japanese regulation assigns a frequency band to a particular
bandwidth, modulation, and duplex, based on the radio
wireless system uniquely, both for licensed and unlicensed
environment is described. It is also recommended that in
bands. However, time ratio of frequency utilization varies
order to realize this kind of system by 2011, a wide range of
widely according to location, time, a day of week, wireless
research and development is needed to be promoted.
communication system, communication carrier company, etc.
Cognitive radio has two trends. One is the so-called “Multiple
By using these spare radio resources adaptively, the time ratio
System”, which switches wireless communication systems
of frequency utilization can be increased.
according to the radio conditions [2, 3]. The other is the so-
For example, assuming that EVDO as System (A), Wireless
called “Dynamic Spectrum Access”, which recognizes spare
LAN as System (B) and WiMAX as System (C), when the
frequencies of a primary system and allocates them to be used
current time is t2, terminal D communicates using EVDO,
for communication of a secondary system, to such an extent
and when the current time becomes t3, terminal D changes
that it would not affect the primary system.
the wireless communication system from EVDO to wireless
Based on the “Multiple system” concept, we have studied a
LAN. In this sense, to realize system switching as depicted
cognitive radio system that covers plural wireless
above, a fast system handover is required.
communication systems. We have built the basic system
Further, the terminal of cognitive radio system (cognitive
architecture that realizes the cognitive radio system [4, 5, 6].
terminal) switches the wireless communication system
In this paper, we propose the architecture of the cognitive
frequently according to the radio condition as previous said.
radio system and describe the testbed system, built based on
the proposed architecture. In section 2, we describe the

1-4244-1144-0/07/$25.00 ©2007 IEEE.

The 18th Annual IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC'07)

Frequency Frequency Frequency the data received from these two systems, monitor terminal to
of System (A) of System (B) of System (C) display the streaming data delivered.

ch1 ch2 ch1 ch2 ch3 ch1 ch2

Server
Frequency
Network
Time t1 CT-A CT-B CT-C

Cognitive BS
Time t2 CT-D CT-A CT-B CT-C Monitoring Node

Time t3 CT-A CT-D CT-B CT-D CT-C Control Node

Time t4 CT-D CT-A CT-B CT-C CT-D

CT: Cognitive Terminal WiMAX Wireless

AP LAN
Figure 1: Concept of Cognitive Radio. AP

Therefore in the cognitive radio, it would be required that the

corresponding node needs not to know which wireless system
is being used. In a conventional system, each wireless module
has an IP address, and a terminal that has multiple wireless
communication modules would be assigned with multiple IP
addresses individually. In our cognitive radio system, each
terminal has only one local IP address regardless the number
of the wireless communication systems that the terminal
communicates with.
Based on the behavior of cognitive radio, we provide four
requirements to realize cognitive radio below:
1. System architecture for fast system handover.
2. One local IP address assignment to the terminal regardless
the number of the wireless communication systems that the
terminal communicates with.
3. The area of cognitive base station is the same as the area of
the WiMAX system, which covers the widest area among the
other wireless systems.
4. Cognitive base station consists of the function of access
point of WiMAX and wireless LAN in the cognitive base
station area, and control node to integrate these functions.
III. SYSTEM ARCHITECTURE
The system architecture based on the concept described in
section II is shown in Figure 2. As wireless communication
systems, we adopted two systems: One is WiMAX (3.5GHz),
one of the candidates of wireless broadband system for
metropolitan area, and the other is wireless LAN (2.4GHz),
the most popular wireless communication system in office
and indoor environment. Cognitive Base Station consists of
multiple access points of WiMAX and wireless LAN,
monitoring node and control node. Monitoring node collects
various kinds of radio information from each wireless
communication system, and calculates statistical value to
recognize the radio condition. Control node switches wireless
communication system according to the information from the
monitoring node.
Cognitive terminal consists of WiMAX CPE function, access
terminal function of wireless LAN, control node to converge
WiMAX Wireless LAN
CT Change in
Location
Figure 2: System Architecture of our system.
The terminal of cognitive radio system (cognitive terminal)
switches the wireless communication system frequently
according to the radio condition. Therefore in the cognitive
radio, it would be required that the corresponding node needs
not to know which wireless system is being used. In our
proposed system, each terminal has only one local IP address
regardless the number of the wireless communication systems
that the terminal communicates with.
IV. SIMULATION
A. Simulator
Based on the system architecture in section III, we have
developed a simulator that supports both WiMAX and
wireless LAN.
As described in section III, cognitive node switches the
wireless system to communicate with according to the radio
condition. In the simulator, monitoring node monitors RSSI
value of wireless LAN and based on this value, control node
switches the system, because WiMAX service is provided in
all area of the simulator.
The threshold value to switch wireless systems is set to
certain value (RSSI_th), and when RSSI value reported from
monitoring node exceeds RSSI_th, control node switches the
communication system from WiMAX to wireless LAN, and
when RSSI value reported from monitoring node is below
RSSI_th, control node switches the communication system
from wireless LAN to WiMAX.
System parameters of each wireless system for this simulator
are shown in Table 1.
The 18th Annual IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC'07)

Table 1: Simulation Specifications.

Item Specifications 2) Scenario 2
WiMax System IEEE802.16-2004
Max. rate 75Mb/s In scenario 2, the terminal communicates with WiMAX area
QoS Mode rtPS first, and then moves to wireless LAN area as shown in
Radio Info. RSSI Figure 4, and switches wireless system according to the RSSI
Wireless LAN System IEEE802.11b of wireless LAN.
Max. rate 11Mb/s
Radio Info. RSSI
Data Scenario 1 UDP (DL: 3.2Mb/s)
Scenario 2 UDP (DL: 8kb/s)
Mobility Scenario 1 6.0km/h WiMAX Area
Scenario 2 2.7km/h
Cognitive
WiMAX AP Terminal
B. Simulation Results
WLAN Area
Control Node
1) Scenario 1
In scenario 1, a terminal moves along a line between WLAN AP
cognitive base station and cognitive terminal, and the terminal
Direction
is in both WiMAX and wireless LAN area. To realize the
system switching between two systems, we set the threshold
to switch the wireless system according to RSSI value of
wireless LAN. Figure 4: Simulation Overview (Scenario 2).
Simulation Result is shown in Figure 3. The top two figures
of Figure 3 show the history of user throughput with using As described in Figure 4 and Table 1, the cognitive terminal
wireless LAN and RSSI of wireless LAN. When the distance moves in WiMAX service area through wireless LAN service
from cognitive base station and cognitive terminal becomes area with slow speed and constant bit rate data is transmitted
longer, user throughput becomes lower due to the degradation from the server to the cognitive terminal.
of propagation loss that is related with RSSI of wireless LAN. The result is shown in Figure 5. It shows that the control node
And from the bottom left figure of Figure 3, user throughput switches the wireless system according to RSSI value of
improvement can be seen by switching to WiMAX according wireless LAN correctly, and data transmission rate is still kept
to RSSI level of wireless LAN. And more we can maximize regardless the system switching, and we have found the effect
the user throughput by adjusting the threshold of switching of using cognitive radio system.
wireless system as shown in the bottom right figure of Figure
3. WiMAX Received Signal Level Wireless LAN Received Signal Level
-78 0
RSSI (dBm)

-82 -40
RSSI (dBm)
RSSI of Wireless LAN (dBm)

5 Communicates with wireless LAN only -91

User Throughput (Mbit/s)

4 -92 -86 -80

3 -93
-90 -120
2 -94 0 1000 2000 3000 4000 0 1000 2000 3000 4000
Simulation Time (seconds) Simulation Time (seconds)
1 -95
0 60 120 180 240 0 60 120 180 240
User Throughput Selected System Status
Simulation Time (seconds) Simulation Time (seconds) 10
Throughput (kbit/sec)

5 WiMAX + wireless LAN WiMAX + wireless LAN WLAN

8
User Throughput (Mbit/s)
User Throughput (Mbit/s)

Communicates 3.4
with WiMAX
4 6
3.2
3 4
WiMAX
3.0 2
2
Threshold = -93.1dBm 0
0 1000 2000 3000 4000 0 1000 2000 3000 4000
1 2.8
0 60 120 180 240 -93.2 -93.0 -92.8 -92.6 -92.4 -92.2 Simulation Time (seconds) Simulation Time (seconds)
Simulation Time (seconds) Threshold (dBm)

Figure 5: Simulation Results.

Figure 3: Simulation Result (Scenario 1).
The 18th Annual IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC'07)

V. EXPERIMENTS result, we set the threshold value to switch wireless system to

-65dBm, and when RSSI value reported from monitoring
A. Testbed System node exceeds this threshold, control node switches the
communication system from WiMAX to wireless LAN, and
Based on the system architecture in section III, we have also when RSSI value reported from monitoring node is below -
developed a preliminary testbed system that supports both 65dBm, control node switches the communication system
WiMAX and wireless LAN. Figure 6 shows the overview of from wireless LAN to WiMAX.
the testbed system. In the testbed system, we connect the base
station side and the terminal side via RF cables, with variable 10 10
attenuators inserted in the middle of RF cables instead of WiMAX
8Mb/s
Wireless LAN
8Mb/s
wireless radio emission. The attenuation level can be changed 8 8

Throughput (Mb/s)

Throughput (Mb/s)
manually and continuously, simulating the fluctuation of
radio condition. 6 6
4Mb/s 4Mb/s
4 4
2Mb/s 2Mb/s
2 2

0 0
-90 -80 -70 -90 -80 -70
Streaming Server Monitoring Node
Control Node RSSI level (dBm) RSSI Level (dBm)
Control Node WiMAX BS
/ WLAN AT
/ WLAN AP Terminal
Figure 7: Throughput performance of each system.
WiMAX CPE
The reason for setting the threshold to -65dBm is that white
Gaussian noise level is known as -95dBm and -65dBm is
Figure 6: Overview of the Testbed System. 30dB higher than the Gaussian noise. This value is equivalent
to the SNR (Signal to Noise Ratio) of 30dB, and it is expected
System parameters of each wireless system for this that high performance can be achieved with wireless LAN.
experiment are shown in Table 2. The result of the streaming experiment demonstrates control
As described in Table 2, RSSI level of wireless LAN can be node switches the communication system according to the
adjusted in the range of -95dBm to -50dBm using variable threshold, namely the RSSI of wireless LAN, accurately, and
attenuator. Streaming data is transmitted from the streaming a continuous streaming service regardless the frequent
server shown in the left side of Figure 6, to the cognitive switching between the communication systems.
terminal shown in the right side of Figure 6. Next, we have measured delay performance on the testbed.
Concretely, we have measured the end to end delay of
Table 2: Testbed Specification. streaming service for the system shown in Figure 6. The
result is shown in Table 3.
Item Specifications
WiMax Based on IEEE802.16-2004
Max. rate 75Mb/s Table 3: End-to-End Packet Delay.
Freq. Band 3.5GHz System Delay (average)
Tx Power 28dBm WiMAX 1.21 seconds
QoS Mode Continuous Grant Wireless LAN 1.22 seconds
Wireless LAN Based on IEEE802.11g
Max. rate 54Mb/s
Freq. Band 2.4GHz From Table3, there is no obvious difference between two
Tx Power 18dBm systems. The reason is that we use streaming service as an
RSSI level -95dBm to -50dBm application, in which buffers for reconstructing images in the
Data UDP DL: 2-3Mb/s application is predominant.
Mobility Speed 0km/h From this result, delay difference due to the switching of
(insert variable att.) wireless communication systems barely affects the
application, and we have proved that stream data can be
delivered continuously with cognitive radio system.
B. Results of Experiment
Before system switching, we evaluated the throughput of each
wireless system. The result is shown in Figure 7, and we have
found that both WiMAX and wireless LAN have capability to
transmit up to 8Mb/s on the testbed. Based on this evaluation
The 18th Annual IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC'07)

VI. CONCLUSION
In this paper, we proposed the architecture of the cognitive
radio system, and described the simulator and testbed system
with WiMAX and wireless LAN based on the proposed
architecture. Through simulation and testbed, we have proved
that stream data can be delivered continuously without any
effects of handovers. Also, we presented the evaluation result
of the end-to-end latency on the testbed system.
For the next step, we will add CINR (Carrier to Interference +
Noise ratio) as radio environmental information, and will
evaluate this testbed system under mobile environment.

ACKNOWLEDGMENT
Part of this project is funded by Ministry of Internal Affairs
and Communications of the Japanese Government.

REFERENCES
[1] Ministry of Internal Affairs and Communications, “e-Japan Priority
Policy Program – 2004,” Jun. 2004.
[2] M.Nohara, et al., ”Study on Cognitive Radio Communications for
Radio Resource Enhancement,” IEICE, SR2006-11, Apr. 2006.
[3] K.Kashiki, et al., “Research and Development on Cognitive Radio
System -Current Research Results,” IEICE, SR2006-22, Jul. 2006.
[4] S.Yoshizawa, et al., ”A Study of Cognitive Radio –System Concept for
Muti-System Integration,” IEICE General Conference 2006, B5-124,
Mar. 2006.
[5] S.Hanaoka, et al., ”A Study of Cognitive Radio –Environment
Recognition for Wireless LAN,” IEICE General Conference 2006, B5-
125, Mar. 2006.
[6] S.Hanaoka, et al., ”A Study of Cognitive Radio –System Concept for
Experimental System,” IEICE Society Conference 2006, B17-6, Sep.
2006.
[7] H.Harada， ”A Study on Cognitive Radio and its applications,” IEICE,
SR2005-18, May 2005.
[8] J.Mitola, “Cognitive radio for flexible mobile multimedia
communications,” 1999 IEEE Int Workshop on Mobile Multimedia
Communications Digest, Nov. 1999.
[9] S.Hanaoka, et al., ”Proposal of Radio System Handover for Cognitive
Radio,” IEICE, SR2006-42, Nov. 2006.
[10] J.Mitola, et al., “Cognitive Radio: Making Software Radios More
Personal,” 1999 IEEE Personal Communication, Vol.6, No.4, 1999.
[11] H.Harada, et al., “The Overview of the New Generation Mobile
Communication System and the Role of Software Defined Radio
Technology,” IEICE Trans. Communication, Vol. E86-B, No.12, Dec.
2003.
[12] D.Bourse, et al., “E2R Project: Major European Initiative on
Reconfigurability,” Proceeding 2004 SDR Technical Conference 1.3-3,
Nov. 2004.
[13] H.Harada, “Software defined radio prototype for W-CDMA and
IEEE802.11a wireless LAN,” IEEE VTC’04 Fall, Nov. 2004.
[14] Hung-Yun Hsieh, et al., ”An end-to-end approach for transparent
mobility across heterogeneous wireless networks,” Mob. Network
Application Vol.9 No.4, 2004.