The present invention relates to a packet transmission technique for realizing wireless transmission of packets between network segments, and more particularly, to a wireless basestation functioning as a wireless bridge and an optimum routing method used in an ad hoc wireless network with reduced load concentration.
A network that allows two or more devices (or terminals) to link up spontaneously to conduct mutual communication is called an ad-hoc wireless network, or a multi-hop wireless network. Such a network does not require a specific control station, and the devices define a self-organized local network by themselves. In an ad hoc wireless network, terminals that cannot communicate directly with each other can transmit and receive packets via a third terminal located between them, while keeping the transmit power levels low. The communication range can be expanded with this arrangement. In fact, network layer routing protocols are proposed as the Internet standard of the ad hoc wireless network. See, for example, S. Corson and J. Macker, “Mobile Ad hoc Networking (MANET): Routing Protocol Performance Issues and Evaluation Considerations”, Internet standard RFC 2501, January 1999.
On the other hand, in a wired network, packets are transmitted between different network segments using bridges, and a packet transmission technique based on a spanning tree protocol is known. See, for example, “Information Technology, Telecommunications and Information Exchange between Systems, Local Area Networks, Media Access Control (MAC) Bridges”, ISO/IEC 10038, ANSI/IEEE Std 802. ID, 1993. With this technique, a specific bridge or basestation is selected as a root bridge, and a transmission tree extending from the root bridge is created using a spanning tree protocol to prevent a loop and realize a fault-tolerant system.
An example of the transmission tree in a wired network is illustrated in
In order to apply the above-described packet transmission technique using wired bridges to a wireless network, multiple wireless interfaces have to be provided to each of the nodes through which packets are transmitted. This attempt further requires an antenna and a modulation and demodulation circuit to be added to each of the interfaces; and is disadvantageous on the cost front.
To overcome this problem, JP 2000-69046A proposes a technique for virtually providing multiple wireless interfaces by regarding the address of the correspondent node as a virtual interface. This method allows the concept of wired network spanning tree to be applied to packet transmission in a wireless network, using substantially a single wireless interface. Each of the wireless basestations is furnished with an address table, in which table the address of the correspondent node is recorded in association with the next node to which the packet is to be transmitted.
Another publication JP 2000-78147A discloses a technique for reducing redundancy of packet transmission routes, making use of the feature of wireless communications, that is, the fact that packets can reach a wireless station as long as that wireless station is located within the communication range, even if the station is not the correspondent node. With this technique, each wireless station monitors packets approaching that station along a tree-like transmission route. A table is created to record the source device indicated by the source address contained in the packet, in association with the previous wireless node (or basestation) indicated by the address of the transmitting station. Using the table, a shorter route can be selected in packet transmission.
Still another publication JP 2003-188811 proposes a method for improving the communication quality by determining whether the power level of the signal received from a basestation is at or above a threshold level when creating a transmission tree, and by not establishing a link if the detected power level is below the threshold.
By the way, it is necessary for an ad hoc wireless network to select the optimum route in a short time when transmitting packets or creating a transmission tree if such a tree is used. However, in wireless communication channels, the network environment and conditions are likely to change, as compared with wired communication channels. Accordingly, JP 2003-152786A proposes to introduce a weighting value reflecting the bit error rate and the transmission rate between adjacent wireless devices (or wireless basestations) in determining the optimum route. The weighting values set in the links between adjacent devices are successively added up to determine the optimum route.
The packet transmission routing protocol proposed by the above-described Internet Standard RFC 2501 requires all the devices forming the network to be furnished with packet relay functions. In other words, devices with less sophisticated functions cannot participate in the network, and extra funds for adding functions or equipment are needed.
The above-described publication JP 2000-69046 tries to apply the concept of wired packet transmission disclosed in ANSI/IEEE Std. 802 to a wireless packet network, and a single transmission tree is used in a wireless network, as illustrated in
With the technique disclosed in JP 2000-78147A, the packet transmission route can be shortened by monitoring at a node those packets approaching that node. However, load concentration occurs at the root bridge of the transmission tree. Consequently, the entire network efficiency is lowered.
In the packet transmission system disclosed in JP 2003-188811, the determination whether to establish a channel between adjacent wireless basestations is made based on the power level of the received signal. However, the actual wireless communication environment changes easily, and wireless interfaces that adaptively change the communication rates according to the environment are also used. The technique shown in this publication cannot estimate the link cost reflecting the communication rate. Consequently, the throughput of the network is lowered as a whole, and the network connectivity is degraded.
The route optimization method disclosed in JP 2003-152786A considers the change in the wireless environment. However, it does not consider overhead with respect to the data transmission time, which overhead changes depending on the size (or the length) of the payload of a packet.
With the conventional packet transmission methods in the third layer (the network layer), wireless devices not having bridging functions cannot participate in the network. On the other hand, packet transmission methods using bridges in the media access control (MAC) layer are still in the process of development for application to wireless networks. When using a transmission tree in a wireless network, localized load concentration cannot be avoided.
Therefore, it is an object of the present invention to provide a packet transmission technique in an ad hoc wireless network that can achieve improved network efficiency through determination of the optimum route, while avoiding localized load concentration, regardless of the functionalities of wireless devices.
One method to achieve the object is using two or more transmission trees in an ad hoc wireless network to determine the optimum route, while appropriately distributing the workload across the network.
Another method to achieve the object is determining the optimum route taking into account the payload size of a packet, in addition to the continually changing wireless environment.
With the former method, two or more wireless basestations (bridges) behave as root bridges to create multiple transmission trees simultaneously, and each of the wireless basestations determines which tree is to be used for the currently being conducted packet transmission.
In the network, wireless basestations (regardless of whether a mobile type or a fixed type) having bridging abilities function as wireless bridges, while terminal devices (such as wireless devices) with lower abilities connect themselves to the nearest wireless basestation to participate in the network. An arbitrary mobile terminal with a bridging function may participate in the network to function as a wireless bridge.
With the latter method, the payload size of a packet is reflected in the link cost when creating and/or updating a transmission tree, regardless of existence or absence of a transmission tree, or regardless of the number of transmission trees. The optimum transmission route can be determined according to the actual conditions of the radio channel for packet transmission.
In the first aspect of the invention, a packet transmission system that includes a plurality of wireless basestations and one or more terminal devices belonging to one of the wireless basestations is provided. In this system, multiple transmission trees are employed. Each of the wireless basestations has a tree table in which identification (ID) information items of the transmission trees are recorded in association with the corresponding root bridges. When receiving a packet, each of the wireless basestations determines which transmission tree is to be used for the packet transmission, and transmits the packet to the next bridge according to the determined tree.
In determining the appropriate transmission tree, either the following means (1) or (2) may be employed.
The transmission tree used in the packet transmission system may be created based on a link cost reflecting the radio channel conditions, such as the power level of the received signal, the error rate, or the delay.
In the second aspect of the invention, a wireless basestation used in a wireless packet transmission system is provided. The wireless basestation includes:
In the third aspect of the invention, a method for optimizing a packet transmission route in a wireless network is provided. The method includes the steps of:
Since the route is determined taking into account the packet size, in addition to the transmission rate and/or other factors, the packet can be transmitted along the optimum route with reduced overhead.
In this context, the term “transmission tree” includes a routing table, spanning tree, and any suitable routing means.
Other objects, features, and advantages of the invention will become more apparent from the following detailed description when read in conjunction with the accompanying drawings, in which
The preferred embodiments of the present invention are now described in detail in conjunction with the attached drawings.
First, the first embodiment of the invention is described with reference to
In
Although, in the example shown in
Each of the wireless basestations (Bridges) a-f has a tree table, in which the identification (ID) information of each of the transmission tress is recorded in association with the root bridge, together with the adjacent bridges (nodes) along the transmission tree.
When using multiple transmission trees in a wireless network, each of the wireless basestations (bridges) determines which transmission tree is to be used for the currently transmitted packet when a packet is received. There are two methods for determining which transmission tree.
The first and second methods are described in more detail below. When the transmission tree is determined, the wireless bridge transmits the packet to the next node (bridge) according to the transmission tree. By using multiple transmission trees in the wireless network, route optimization is performed efficiently, and the packet transmission route can be shortened. In addition, load concentration on a specific wireless bridge can avoided, and the network efficiency can be improved as a whole.
The source address represents the address of wireless communication equipment that first generates and transmits the packet. The destination address represents the address of wireless communication equipment to which the packet is finally addressed. The address of a transmitting node represents the address of wireless communication equipment that transmits the packet along the transmission tree, and the address of a receiving node represents the address of wireless communication equipment that receives the packet along the transmission tree.
The symbol “DS” represents wireless communication equipment, “To” indicates a direction to the receiving side, and “From” indicates a direction from the transmitting side. If the field “To DS” is “0”, a terminal device (wireless terminal) is on the receiving side. If the field “To DS” is “1”, a wireless bridge is on the receiving side. If the field “From DS” is “0”, a wireless terminal is on the transmitting side, and if the field “From DS” is “1”, a wireless bridge is on the transmitting side. If both fields “To DS” and “From DS” are “1”, then the packet is transmitted between adjacent wireless bridges. By inserting the “To DS” field and the “From DS” field in the packet, it can be determined whether the packet is currently being transmitted between bridges.
If, in the network shown in
Bridge “b” that has received the packet from Bridge “a” adds the self address in the address field of the transmitting node, and adds the address of Bridge “c” (which is the next node) in the address field of the receiving node.
Next, explanation is made of the above-described two methods for identifying the currently used transmission tree, using the network topology shown in
When a packet is transmitted from Station S to Station D in the network shown in
Upon receiving the packet from Station S, Bridge “b” transmits the packet to the next node(s) along the transmission tree extending from Bridge “b” itself as the root. Each of the nodes (Bridges) having received the packet checks the address fields of the packet, and determines that the source terminal is Station S. These Bridges also check the location table to determine that Station S currently exists under Bridge “b”, select the transmission tree extending from Bridge “b” as the root from the tree table, and transmit the packet to the next node along the selected transmission tree.
For example, Bridge B transmits the packet to Bridges x7, x8, and x9 along the transmission tree indicated by the dashed arrows. Bridge x7 selects the transmission tree extending from Bridge b as the root, based on the location table and the address information of the packet, and transmits the packet to Bridges x4 and x6. Bridge x8 also selects the transmission tree, and determines that the next node does not exist on the selected tree. In this case, Bridge x8 discards the packet. Bridge x9 performs the same process as Bridge x7. In this manner, the packet is finally relayed to Bridge “c”, and delivered to Station D located under Bridge “c”.
If a transmission tree extending from the bridge of the transmitting side is employed, as in the above-described example, the packet is transmitted to two or more nodes at a branch point. This arrangement may be acceptable for broadcasting a packet without a specific destination. However, when unicasting a packet addressed to a specific destination along the transmission tree extending from the bridge of the transmitting side, the packet is transmitted to bridges irrelevant to the destination. To avoid such inefficiency, the following measures may be employed.
With measure (a), a packet is transmitted from Station S addressed to Station P in the network shown in
This arrangement can further reduce the network workload. On the other hand, when broadcasting a packet without specific destinations, or when transmitting a packet addressed to a terminal device whose basestation is unknown, a transmission tree extending from the root bridge under which the source terminal is located is selected to relay the packet to the destination(s).
A wireless bridge located at the end of the network may not have a transmission tree extending from this wireless bridge itself as the root. For example, Bridge “c” shown in
With measure (b), a learning table is used, in addition to the location table. Each of the wireless bridges records the previous node from which the packet generated by the source terminal specified by the source address is currently relayed. For creation of the learning table, any known method may be employed.
Station D then sends a reply packet to Station S (as indicated by the arrow (4)). The reply packet is transmitted from Bridge “c” to Bridge “x4” (as indicated by the arrow (5)). Bridge x4 determines that the source terminal is Station D and the reply packet is relayed from Bridge “c”, based on the address information contained in the reply packet, and records these information items in the learning table (step (6)), as exemplified in
The reply packet is further transmitted from Bridge x4 to Bridge “b”, via Bridge x7 (as indicated by the arrow (7)). Bridge “b” records the information pair about the source terminal (Station D) and the previous node (Bridge x7) in the learning table (step (8)), and supplies the packet to Station S.
Although not illustrated in
When Bridge “b” again receives another packet addressed to Station D, Bridge “b” transmits this packet only to Bridge x7, without transmitting the packet to Bridge x8 or other branches, because Bridge “b” already has information that Station D is located in the direction of Bridge x7. Similarly, when Bridge x4 receives the packet addressed to Station D next time, it transmits the packet only to Bridge “c”, without transmitting to other branches. With this arrangement, wasting of packets can be greatly reduced even if a transmission tree extends from the root bridge on the transmitting side.
With method (1) providing a location table to each of the wireless basestations (bridges), when a terminal device is newly connected to one of the wireless basestations, or when a terminal device moves and connects itself to another wireless basestation, the wireless basestation transmits a message packet reporting the newly connected terminal device along the transmission tree by broadcasting. Upon receiving the message packet, each of the wireless basestations records the newly connected terminal device in the location table. At this time, the wireless basestation may be configured to record the source bridge that transmitted the message packet, in association with the previous node, in the learning table.
With this arrangement, the location table and the learning table are simultaneously updated every time a terminal device is newly connected to a wireless basestation, and the optimum routing along the transmission tree can be performed more efficiently.
Next, explanation is made of the second method (2) for determining the currently used transmission tree, with reference to
It is assumed that a packet is transmitted from Station S located under Bridge “b” to Station D located under Bridge “c” in the network shown in
In the example shown in
Bridge x7 located on the transmission tree receives the packet from the root bridge, and checks the extra field of the packet to determine that the packet is being transmitted along the transmission tree extending from root Bridge “b”. Bridge x7 adds the address of the next node, which is Bridge X4 in this example, as the receiving node address in the address field 1 of the packet, and adds the self address as the transmitting address in the address field 2.
If a wireless basestation (bridge) is located at a branch of the transmission tree, the packet is to be transmitted to all the branched paths if only the tree information or the root bridge information contained in the packet is used. To avoid this inefficiency, it is preferable for the second method to use a learning table shown in
In the example shown in
Next, explanation is made of creation of a transmission tree in conjunction with
Under these circumstances, in the first embodiment, the link cost is adjusted according to the conditions of the wireless channel between wireless interfaces or the network traffic when creating a transmission tree.
For example, the signal power level received from the adjacent bridge and/or the error rate are reflected in the cost estimation. In the cost table shown in
Using these parameters, the cost between the target bridge and the adjacent (nearby) bridge is estimated by the following equation, using normalization factors α, β, and γ.
Cost=α*(Signal)+β*(Queue size)+γ*(Error Rate)
Upon receiving the tree creation packet, Bridge “b” calculate a link cost “ab” between Bridge “a” and Bridge “b”, based on the cost table shown in
Similarly, Bridge “c”, which has received the tree creation packet from Bridge “a”, calculates a link cost “ac” between Bridge “a” and Bridge “c”, adds the calculated cost “ac” to the cost field of the packet, and transmits the packet to adjacent nodes (as indicated by the arrows (3)).
When Bridge “c” receives the tree creation packet from Bridge “b”, Bridge “c” compares cost “ac”, which is the link cost between Bridge “c” and the root Bridge “a”, and cost “ab+bc”, which is the link cost calculated from the tree creation packet from Bridge “b”, and discards the link with a higher cost. For example, if ac<ab+bc, then the path extending from Bridge “a” via Bridge “b” to Bridge “c” is not employed. By repeating this process at Bridge “d” (the arrow (4)), Bridge “e” (the arrows (5)), and Bridge “f” (the arrow (6)), a non-loop tree indicated by the solid line in
The transmission tree can be created or updated at prescribed time intervals, or whenever a wireless terminal device with a bridge function participates in the network to behave as a bridge. The information about the newly created or update transmission tree is supplied to the wireless basestations (bridges) on the network, and each of the basestations updates the tree table. By reflecting the wireless channel conditions in the link cost, a transmission tree can be created adaptively in agreement with the actual traffic or the network topology.
In the structure shown in
In the structure shown in
Each of the wireless basestations 10A and 10B also has a cost estimation unit 20. The cost estimation unit 20 includes a cost table 21 and a channel monitoring unit 22. The channel monitoring unit 22 monitors the channel conditions and updates the parameters (cost values) recorded in the cost table 21. When the transmitting and receiving unit 12 receives a tree creation packet, the cost estimation unit 20 adds the cost from the previous node to the wireless basestation 10A (or 10B) to the tree creation packet, consulting the cost table 21. The transmitting and receiving unit 11 then transmits the tree creation packet to adjacent nodes. If the wireless basestation 10A (or 10B) becomes a root bridge, the transmitting and receiving unit 11 generates and transmits a tree creation packet with cost zero.
Although only a single interface and a single transmitting and receiving unit 11 are depicted in
The first embodiment has been described based on the example of a complete wireless network. However, a portion of the network may be wired. For example, a terminal device may be connected to one of the wireless basestations by a cable. A mobile terminal with a bridge function may be incorporated as a wireless basestation in the network. When such a mobile terminal participates in the network, a transmission tree is created dynamically and adaptively, reflecting the current wireless channel conditions. Using two or more transmission trees, route optimization and load distribution can be performed more efficiently.
Although the first embodiment has been explained using the wireless LAN based on the IEEE 802.11 standard, the invention is not limited to this example and is applicable to a wireless network of a WCDMA scheme or the next generation wireless communication scheme. In addition, by adding an interface and a protocol converting function to each of the wireless basestations, the first embodiment can be applied to a wireless packet network in which network segments with different communication schemes coexist.
A portion or all of the transmission trees may be put together to reduce the workload required for maintaining multiple transmission trees. In this case, information about the combined trees is added to the extra field of the packet and/or the tree table.
In the example shown in
[(20+16+16+101.5)+(8x+134)/k] μs.
The value k [Megabit] expressing data transmission rate varies depending on the modulation scheme and/or the coding rate employed according to the radio wave environment between wireless basestations (or access points). For example, with high-intensity radio waves, a higher bit rate is employed, and with lower intensity, a lower bit rate is employed. In the example of
If the payload of a packet is 1000 bytes (x=1000), time required for packet transmission in the 6M mode is about 1510 μs, according to the above-described formula. In the 27M mode, it become about 454 μs.
If the payload of a packet is 100 bytes (x=100), then 6M-mode packet transmission time becomes 310 μs, and 27M-mode packet transmission time becomes 189 μs.
If 1-hop transmission is performed for 1000 bytes payload on the 6 Mbps channel, the transmission time (required for one frame) is about
1510[μs]*1[hop]=1510[μs].
This can be converted to data rate of about 5.3 Mbps.
If 2-hop transmission is performed for the same 1000-byte payload on the 27 Mbps channel, one-frame transmission time becomes
454[μs]*2[hop]=908[μs],
which can be converted to data rate of about 8.6 Mbps.
This means that it is advantageous for a long packet (with a greater payload size) to select a higher bit-rate route even if the number of hops increases.
On the other hand, when 1-hop transmission is performed for 100-byte payload on the 6 Mbps channel, one frame transmission time becomes about
310[μs]*1[hop]=310[μs],
which is converted to data rate of about 2.6 Mbps. If 2-hop transmission is performed for the same 100-byte payload on the 27 Mbps channel, one-frame transmission time becomes
189[μs]*2[hop]=378[μs],
which is converted to data rate of about 1.9 Mbps.
Since, for a short packet (with a smaller payload size), the overhead ratio needed for header transmission or the like becomes higher, it is advantageous to select a route with fewer hops even if the bit rate is lower.
In this manner, when a wireless network employs adaptive modulation and adaptive coding, the optimum route varies depending on the payload size of the packet to be transmitted. Accordingly, in the second embodiment, each of the wireless basestations in the network is furnished with a packet size determination unit and multiple routing tables corresponding to different packet sizes, to take into account the packet size, in addition to the transmission rate, when creating a transmission table.
Although in
Upon receiving a link cost request, such as a route search packet or a tree creation packet, the cost estimation unit 60 calculates a link cost between the previous node and the wireless basestation 50 itself, taking into account the current transmission rate, for each of the short-packet case and the long-packet case. Then, the cost estimation unit 60 adds the two estimation results to the route search packet or the tree creation packet, and transmits the packet to adjacent wireless basestations. The wireless basestation 50 updates the short packet table 57 and the long packet table 58 based on the routes selected from the cost information in the network.
The wireless basestation 50 may be either a mobile type or a fixed type, as in the first embodiment.
When transmitting a packet from wireless basestation A to wireless basestation E, the wireless basestation A transmits a route search packet to adjacent nodes. Node (wireless basestation) F and node B receive the route search packet, and each of these nodes calculates a link cost from the previous node (node A) to this node for each of the short packet case and the long packet case, and adds the calculation results in the packet. At node F, a short packet link cost of 310 μs and a long packet link cost of 1510 μs are estimated based on the current transmission rate between node A and node F. Similarly, at node B, a short packet link cost of 189 μs and a long packet link cost of 454 μs are estimated based on the transmission rate between node A and node B.
Then, node E receives the packet from node F, and calculates a short packet link cost and a long packet link cost for the channel between node F and node E, and adds the calculation results to the A-F link cost. Similarly, node C, which has received the packet from node B, calculates a short packet link cost and a long packet link cost for the channel between node B and node C, and adds the calculation result to the A-B link cost. By successively repeating this process, several candidate routes from node A to node E can be obtained. In this example, the following candidates of
Route 1: A→F→E;
Route 2: A→B→C→E; and
Route 3: A→B→C→D→E are obtained.
When the route search packets reache node E, the total link costs of short packet transmission and long packet transmission are obtained for each of the candidate routes. For short packet transmission, cost 1 of Route 1 is 620 μs and the cheapest. Accordingly, Route 1 with the fewest hop number is selected even if the transmission rate is low. For long packet transmission, cost 3 of Route 3 is 1816 μs and the cheapest. Accordingly, Route 3 with the highest transmission rate is selected even if the number of hops is large.
The routes selected for both short packet transmission and long packet transmission are reported to each of the nodes (wireless basestations) in the network, and the short packet table and the long packet table are updated at each node. When wireless basestation A transmits a data packet of short size next time, it checks the short packet table and transmits the packet to the next node E designated in the table. When wireless basestation A transmits a long data packet, it transmits the packet to the next node B by consulting the long packet table.
Although, in
Although the second embodiment has been described using the example based on the IEEE 802.11a standard, the invention is not limited to this example. The link cost can be estimated according to the above-described techniques in an arbitrary transmission system to determine the optimum route taking the transmission rate and the packet size into account. Link cost estimation may be performed using three or more reference values about the packet size or the payload size.
With the arrangement of the second embodiment, the optimum packet transmission route can be appropriately selected taking the packet size into account in an ad hoc wireless network under frequently changing network topology and radio environment over time and locations.
This patent application is based on and claims the benefit of the earlier filing dates of Japanese Patent Application Nos. 2004-041604 and 2004-277495 filed Feb. 18, 2004 and Sep. 24, 2004, respectively, the entire contents of which are hereby incorporated by reference.
Number | Date | Country | Kind |
---|---|---|---|
2004-041604 | Feb 2004 | JP | national |
2004-277495 | Sep 2004 | JP | national |
Number | Name | Date | Kind |
---|---|---|---|
5655219 | Jusa et al. | Aug 1997 | A |
5872773 | Katzela et al. | Feb 1999 | A |
6785552 | Shinozaki et al. | Aug 2004 | B2 |
6810259 | Zhang | Oct 2004 | B1 |
6961575 | Stanforth | Nov 2005 | B2 |
7061876 | Ambe | Jun 2006 | B2 |
20010027107 | Shinozaki et al. | Oct 2001 | A1 |
20030002482 | Kubler et al. | Jan 2003 | A1 |
20030053422 | Ambe | Mar 2003 | A1 |
20030091001 | Watanabe | May 2003 | A1 |
20030112810 | Nakabayashi et al. | Jun 2003 | A1 |
20030165117 | Garcia-Luna-Aceves et al. | Sep 2003 | A1 |
20050243757 | Yagyu et al. | Nov 2005 | A1 |
Number | Date | Country |
---|---|---|
0 849 974 | Jun 1998 | EP |
1-129550 | May 1989 | JP |
2-137544 | May 1990 | JP |
02-214349 | Aug 1990 | JP |
4-284749 | Oct 1992 | JP |
08-037535 | Feb 1996 | JP |
9-182146 | Jul 1997 | JP |
2000-69046 | Mar 2000 | JP |
2000-78147 | Mar 2000 | JP |
2000-261499 | Sep 2000 | JP |
2002-118597 | Apr 2002 | JP |
2002-232448 | Aug 2002 | JP |
2002-247634 | Aug 2002 | JP |
2002-534842 | Oct 2002 | JP |
2002-369239 | Dec 2002 | JP |
2003-069625 | Mar 2003 | JP |
2003-152786 | May 2003 | JP |
2003-188811 | Jul 2003 | JP |
2003-348130 | Dec 2003 | JP |
2005-6264 | Jan 2005 | JP |
2001-0062686 | Jul 2001 | KR |
WO 0039967 | Jul 2000 | WO |
Number | Date | Country | |
---|---|---|---|
20050243757 A1 | Nov 2005 | US |