WO2006095432A1

WO2006095432A1 - Transmission method

Info

Publication number: WO2006095432A1
Application number: PCT/JP2005/004247
Authority: WO
Inventors: Rahul Malik; Pek Yew Tan
Original assignee: Matsushita Electric Industrial Co., Ltd.
Priority date: 2005-03-10
Filing date: 2005-03-10
Publication date: 2006-09-14

Abstract

A transmission method for transmitting a plurality of data units either in one complied frame or altogether. In this method, a scheduling control unit arranges, based on the length of each compartment data unit, the individual compartment data units in the smaller order from the head of a frame body. In case the data units are to be complied into one frame and transmitted wirelessly, a component MPDU of a smaller size is scheduled in that compiled frame to such a positional portion of the body portion of the complied MPDU as is estimated to have a higher reliability, i.e., to a positional portion nearer to the head portion. As a result, a larger number of MSDUs can succeed in their transmissions. This results in an effect that re-transmission overheads are reduced.

Description

Specification

Sending method

Technical field

[0001] The present invention relates to a transmission method in which a plurality of data units are aggregated into one frame or collectively transmitted.

Background art

[0002] Conventionally, the OFDM transmission method has been used in the field of continuous stream transmission such as digital audio broadcasting (DAB), but it is not patentable to apply OFDM to packet-type transmission. This is a relatively new technology, starting with Document 1.

[0003] As a result of being affected by multipath, a radio channel has frequency selectivity.

The state of the radio channel changes over time. By using OFDM as a modulation scheme, a radio channel with frequency selectivity can be divided into a number of subcarriers, each of which can be regarded as receiving frequency non-selectivity Z flat fading, It is possible to use a simple 1-tap equalizer for channel compensation.

A conventional general packet structure will be described with reference to FIG. The figure shows the structure of a packet that is ready for transmission by the physical layer, as defined in the wireless AN (registered trademark) Physical Layer Protocol Data Unit (PPDU), or Non-Patent Document 1 mentioned above. . This packet also has three main section capabilities with different functions, as described in the following paragraphs. Furthermore, as shown in Non-Patent Document 2, the figure shows between MAC layer service data unit (MSDU), MAC layer protocol data unit (MPDU), physical layer service data unit (PSDU), PPDU and The relationship between is also illustrated.

[0005] As the first section, PLCP preamble 110 is a short training sequence

It consists of 10 iterations of (STS) 111 and 2 iterations of long training sequence (LTS) 112. The STS is used by the receiver to perform the functional operations of AGC confluence, diversity selection, timing acquisition, and coarse frequency estimation. LTS uses fine frequencies to correctly equalize received packets against the effects of the channel. Used to perform number estimation and channel estimation.

[0006] As a second section, the signal field 120 has a robust modulation scheme, that is, it is hardly affected by the communication environment !, and the modulation scheme (6 Mbps in the minimum speed mode defined in Non-Patent Document 1 described above). The data power transmitted using is also configured. This consists of parameters (eg, modulation and coding scheme, packet length, etc.) used to initialize the receiver to correctly demodulate the data (payload) segment 130 of the packet.

[0007] As a third section, the data field 130 is the payload of the PPDU. This consists of a physical layer service data unit (PSD U) 132 passed from the MAC layer to the physical layer, a service field 131, a pad field 133, and a tail field 134. The PSDU is called a MAC protocol data unit (MPDU) 150 in the MAC layer. The MPDU also includes a MAC header field 151 and an FCS field 1 53, and a MAC service data unit (MSDU) 152 encapsulated by them. The data field 130 is transmitted using the modulation code method defined in the signal field 120.

[0008] Recently, with the rapid increase in demand for wireless bands, IEEE established a high-throughput task loop TGn for developing next-generation wireless LAN systems with throughput exceeding 100Mbps. To achieve higher throughput, higher physical speeds, combined with multi-antenna transmission and high-level modulation schemes, are expected to be part of the solution, but also at the MAC layer. Some improvement is expected.

[0009] Non-competitive polling mode operation in which each terminal receives an inquiry (polling) from the AP for the presence of data to be transmitted, and the polling terminal starts transmission if there is transmission data However, the above-mentioned Non-Patent Document 2 is mainly directed to the asynchronous contention-based network. As a result of the basic behavior of contention-based access, the likelihood of transmission data collisions is reduced. A high percentage of “access overhead”, including access delay and back-off delay for each occurrence of media access. Will occur. This large overhead makes media use inefficient, which is undesirable for high throughput. Polling mode In order to make it possible for terminal equipment that conforms to this new method and legacy terminal equipment that does not conform to this to coexist in the same network, each individual device In order to allow the station to correctly set the network allocation vector (NAV), which is a local counter used to track the usage of shared media, poll frames are transmitted using the “basic rate”. This is a significant overhead for data transmission.

[0010] In order to improve MAC efficiency by reducing "media access overhead", several proposals that propose the use of MAC frame aggregation (Non-patent Document 3, Non-patent Document) Reference 4) is sent to TGn.

[0011] Here, the conventional aggregation technique will be described in some detail while referring to the contents of the above-mentioned documents. Non-Patent Document 3 describes MAC frame aggregation as a means for reducing access overhead and improving network throughput. This results in a reduction in the number of media access events as a result of aggregation and saves additional overhead such as IFS, backoff period, training sequence, PLCP header, etc. It is intended to improve this. As the frame structure of this document, the frame is composed of a MAC header and a frame body composed of a plurality of MSDU compartments. First, the MAC header characteristically includes an aggregation control field and a header FCS field. The aggregation control field has one compartment count information (subfield) at its head, and then the number of compartments corresponding to the number of compartments (value corresponding to the length of each compartment) compartment length) information. An aggregate data frame can be viewed as a “carrier” for multiple packets, which facilitates point-to-multipoint transmission. The number of compartments is the number of MSDU compartments carried by one aggregated data frame. Next, the frame body will be described. Each MSD U compartment of the body includes a header, a data unit body, and an FCS. By using the above-described aggregate data frame format, the terminal can perform more efficient data transmission when there are multiple data packets to be transmitted when a transmission opportunity is obtained. You can. Note that this document also mentions fragmentation technology, but this technology is intended to deal with the effects of rapid changes in channel conditions by dividing packets into smaller fragments. On the other hand, the aggregation technology is expected to improve the MAC transmission efficiency by transmitting several frames simultaneously in a relatively low channel fluctuation situation. The location of each MSDU component in the frame body and its effects are not particularly touched.

[0012] Special Reference 1: Local and Metropolitan Area Networks-specific Requirements-Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications: Higher-Speed Physical Layer Extension in the 5 GHz Band ", IEEE Std 802.11a—1999, IEEE, September 1999.

Non-Patent Document 2: "Local and Metropolitan Area Networks-Specific Requirements-Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications", IEEE Std 802.11-1999, IEEE, August 1999.

Non-Patent Document 3: T. Fukagawa, et al, 'Partial Proposal for TGn, doc: IEEE

802.11- 04 / 956r0, September 2004.

Non-Patent Document 4: A. S. Mujtaba, et al, TGnSync Proposal Technical Specification, doc: IEEE 802.11- 04 / 889r0, September 2004.

Disclosure of the invention

Problems to be solved by the invention

[0013] As described above, in the conventional frame aggregation technique represented by the above-mentioned literature, there is not much consideration for fluctuations in the radio channel. It is inevitable that retransmission overhead will be generated for errors caused by the change of state, and there is a place where the effect cannot be sufficiently brought about by improving the throughput by reducing the access overhead using the conventional technology.

[0014] The object of the present invention is to reduce the retransmission overhead, consider the priority, satisfy the error threshold, etc. in the case where a plurality of data units are aggregated into one frame or collectively transmitted. It is to provide a transmission method for the effective arrangement of data units and transmission. Specifically, the present invention provides a data based on data unit length. Data unit placement techniques, and further, combined placement based on weights and lengths. Another object of the present invention is to provide a dynamic fragmentation method in which a plurality of fragments are dynamically sized according to their position in the packet in order to achieve a constant fragment error rate. That is.

Means for solving the problem

[0015] The transmission method of the present invention is a transmission method in which a plurality of data units are aggregated into one frame and transmitted in packet-type wireless transmission, and each data unit is extended from the head of the frame body. The method includes an arranging step of arranging the data units in the shortest order and a transmitting step of collecting and transmitting the arranged data units.

[0016] The transmission method of the present invention is a transmission method in which a plurality of data units are aggregated into one frame and transmitted in packet-type wireless transmission, and each of the above-mentioned data units is based on the priority of each data unit. A weighting step of calculating weights for the data units and assigning the calculated weights to the data units; and an arraying step of arranging the data units from the top of the frame body in descending order of weight. And a transmission step of collectively transmitting the arranged data units.

[0017] The transmission method of the present invention is a transmission method in which a plurality of data units are aggregated into one frame and transmitted in packet-type wireless transmission. A weighting step for calculating a weight for each data unit and assigning the calculated weight to each data unit; and for each data unit in the descending order of the assigned weight, A second arrangement step of arranging the first arrangement step to be arranged and two or more data units determined to have the same weight in the first arrangement processing in the order of short length; And a transmission step of collectively transmitting the arranged data units.

The invention's effect

[0018] According to the present invention, when a plurality of data units are combined into one frame for wireless transmission, the predetermined threshold fragment error rate requirement is satisfied for the plurality of fragment data units. Determine the optimal size and ensure that the MPDU fragment error probability is below a given threshold, regardless of its position in the aggregate MPDU If you can do it!

Brief Description of Drawings

[0019] [FIG. 1] A diagram showing a packet structure of a wireless LAN physical layer protocol data unit.

[Fig.2] Diagram showing frame format of aggregated data frame

[Figure 3A] Diagram showing bit error rate (BER) at various locations within a packet under ideal noise-free conditions

[Figure 3B] Diagram showing BER at various locations in the packet under noise conditions in the presence of practical additive white Gaussian noise (AWGN)

[Fig. 4] Diagram showing the channel drift phenomenon of interest in the present invention

[Figure 5] Diagram showing how important information is placed at the beginning of the packet body

[Figure 6] Diagram showing aggregate format

[FIG. 7] Overall diagram of a configuration example for realizing the present invention in a radio transceiver

FIG. 8 is a flowchart showing a procedure for allocation allocation for the second embodiment of the present invention.

FIG. 9 is a diagram showing a dynamic fragmentation mechanism as a third embodiment, which is specified by the present invention.

BEST MODE FOR CARRYING OUT THE INVENTION

[0020] First, some terms used in this specification are defined here.

[0021] The term "aggregation" is not limited to the general meaning of these terms. In this specification, a plurality of data units are aggregated in a frame, that is, taught and paraphrased. For example, it shall imply a summarized state or result or effect. For the strict definition of these terms, the contents defined in some of the reference documents listed in this specification are incorporated by reference where appropriate.

In addition to the general meaning of “robust” as used herein, it is difficult to be affected by the communication environment, as described in Non-Patent Document 4 above. In addition, even if a loss or error occurs in one or more data units within the aggregate, the aggregate structure can be restored without necessarily affecting the entire multiple data units (pages 32 and 33). ) Is also implied and the contents thereof are incorporated in appropriate places in the present specification.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. (Embodiment 1)

In the first embodiment of the present invention, data unit allocation based on packet length will be described.

[0025] This embodiment improves the overall throughput by selectively arranging individual data units in the order of smaller size in the context of the transmitted packet. It is something to be made. The occurrence of a single bit error in a data unit causes an error in the entire data unit. In this embodiment, a larger number of data units are likely to be successfully transmitted. is there. Furthermore, throughput can be improved in the form of a reduction in the number of retransmissions as overhead associated with each data mute.

[0026] First, the frame structure of the prior art will be briefly described again. FIG. 2 illustrates a conventional MAC aggregate frame format described in Non-Patent Document 3 mentioned above. The format described here includes an aggregated MAC header that describes the location of multiple compartments MPD U in the frame body, and each of these multiple compartment MPDUs is individually FCS. Have Reference numeral 210 in the figure represents the MAC header of the aggregated MPDU, and 220 represents the frame body of the aggregated MPDU. It will be apparent to those skilled in the art that for power-intensive MPDUs for which FCS was previously shown in data 130 in Figure 1, an “overall” FCS is not required.

[0027] Unlike the conventional one described in Non-Patent Document 2, the format proposed here uses the header FCS212 to protect the contents of the MAC header information 211 in the aggregated frame MAC header 210. Shall be.

[0028] The frame body 220 of the aggregated MPDU is configured with several compartment MPDU forces, denoted in this example as reference numerals 230, 240, 250 and 260. Each conversion MPDU has its own MAC header (231, 241, 251, and 261), its own frame body (ie, MSDU) (232, 242, 252, and 262), and its own FCS (233, 243, 262). 253 and 263).

[0029] Upon receiving the aggregated MAC frame described in Non-Patent Document 3 above, the receiver First, the checksum is compared with the header FCS 212 to confirm that the MAC header 210 is affected by the error (sanctity). This aggregate MAC header is used to facilitate the determination of the boundaries between the individual MPDU compartments (230, 240, 250 and 260 in this example) within the aggregate frame body. Contains multiple pointers to the MPDU companion of the current.

[0030] Each individual MPDU compartment has its own FCS (233, 243, 253, and 263), which allows the receiver to detect errors in that individual MPDU separately. By using a selective retransmission protocol such as a block ACK mechanism, it is possible to request retransmission of the MPDU compartment in which an error has occurred, while passing an MPDU that has been successfully received as a successful reception to the upper layer. This is a significant impact on throughput. In short, block ACK can be described in a word. This technology improves channel utilization efficiency by consolidating or consolidating several acknowledgments (ACK) into one frame.

[0031] As described above, the radio channel changes over time. A standard wireless LAN system is designed for indoor use where there is no mobility or low mobility of terminals, and the channel variation with respect to the time length of the packet itself is very slow. The power seen As a result of using techniques such as frame aggregation, aggregated frames are all received by using longer packets or by using a MAC mechanism that supports point-to-multipoint transmission. Need to be sent at a common minimum rate supported by the subscriber. This means that the packet transmission time is longer than that of the current 802.11a system.

[0032] The packet transmission time is nevertheless expected to be less than the channel coherence time, and theoretically results in a static channel. Assumptions do not hold. As an example, we simulated a 802.11a system with a standard Doppler frequency of 6 Hz for indoor conditions in the 5 GHz band. Using MSD of 2304 bytes (this is the maximum specified packet size in Non-Patent Document 1 mentioned above), and 54Mbps transmission mode (64QAM and r = 3Z4 convolutional coding), the bit error rate for all different packets Individual bits over minutes It was determined by averaging the errors. The simulation was performed on a 5-tap exponentially selective selective fading channel with 25 ns delay dispersion. Figure 3 shows the result.

FIG. 3A illustrates the bit error rate according to the bit position in the ideal case without thermal noise under the above-described simulation conditions. In this setting, the overall packet error rate is determined to be 2.36%. Since there is no AWGN (additive white Gaussian noise), the only source of bit errors (and packet errors) in Figure 3A is due to the Doppler shift caused by the channel.

[0034] On the other hand, FIG. 3B is a force that illustrates the bit error rate corresponding to the relationship with the bit position under the simulation conditions described above. Here, since there is AWGN, the packet error rate Is 7.63%, which is a typical PER (packet error rate) for wireless LAN. Here, there are two sources of errors: errors caused by the channel Doppler effect and errors caused by AWGN. The effects of Doppler are more pronounced as they are directed toward the tail end of the packet, as can be seen from those noted by ranges 301 and 311, while the errors due to AWGN are noted in range 312. Thus, the packets are uniformly distributed.

[0035] It should be noted that, in generating the results of FIG. 3A and FIG. 3B, according to Non-Patent Document 1, it was assumed that 23 04 bytes / packet were transmitted at 54 Mbps. As a result, the packet transmission time is about 364 sec. As PHY (physical layer) speed and packet size increase and increase at the same time, the transmission time of the next generation network is expected to be comparable.

[0036] FIG. 4 shows the effect of Doppler on the channel estimation as illustrated in FIG. 3, ie, the actual channel 420 relative to the channel estimation 410 derived from the training sequence 401 at the beginning of the bucket 400. This is a description of the drift. This figure illustrates the problem that is addressed by the present invention, that is, the improvement of transmission reliability and, as a result, the improvement of throughput in an aggregated frame transmitted via a radio channel. Show.

[0037] Channel estimation drift phenomenon by using high multi-level modulation schemes such as 256QAM It is expected that the difference between the bit error possibilities at the beginning and tail of the packet will be even greater. All MSDUs delivered to the upper layer are the force that contributes to the throughput. A single bit error in a single MPDU can cause a transmission failure for the entire MPDU. This is subtracted from the achievable MAC throughput.

[0038] The training sequence reinsertion method, which is well described in the prior art, particularly in the field of continuous transmission methods such as DAB, has a training sequence so that the error rate does not exceed a predetermined threshold. The ability to deal with such issues by adjusting the reinsertion period.Such schemes have a wide range of packet sizes, operating environments, uses, and are sufficient for backward compatibility with legacy systems. It cannot be easily adapted to the 802.11 wireless LAN architecture with the established frame format. Needless to say, the weighting overhead associated with the re-insertion of the training sequence and the signal notification during the re-insertion period.

[0039] Based on the background described above, data unit allocation based on packet length according to the present embodiment will be described below.

[0040] First, as a premise example for facilitating the distribution of the present embodiment, FIG. 5 shows important information (for example, information that is extremely important for accurate reception of the entire packet or a packet later. (Information important for correctly decoding the following part or information important for the entire packet, etc.) Force The figure shows how it is placed at the beginning of the packet body. This indicates, for example, that the important information power S that is essential for accurate decoding of the entire frame and that it is placed at the beginning of the packet where the possibility of bit errors is the lowest. Reference numeral 500 in FIG. 5 represents a packet including a training sequence 510 and a packet body 520. The important information 521 is located at the head of the packet body where the error due to channel drift is expected to be minimal, and is closest to the training sequence 510. The likelihood of bit errors is expected to increase as the distance from the training sequence increases. Scheduling highly important information in the portion of the packet where the BER is expected to be low makes it possible to improve the overall throughput of the packet-type communication system. In such a case, aggregate MPDU According to Non-Patent Document 3, the force that the aggregated information is placed in the MAC header at the beginning of the packet (see 211 in FIG. 2). This is in contrast to the distributed format described.

[0041] In the technique for improving throughput proposed by the present embodiment, a plurality of compartment MPDUs are included in the aggregate MPDU, and the shortest of these is the start of the aggregate MPDU. They are scheduled in ascending order of their packet length in the form of being placed in the close part (at the start of the aggregated MPDU). As a result, in the aggregated frame, the relatively small size of the compre- hensive MPDU power is scheduled to a position portion (zones) in which the BER of the packet is expected to be relatively low. According to FIG. 3 and FIG. 4 described above, these position portions are portions that are relatively close to the head portion of the packet, and the possibility of bit error increases as the distance from the head portion increases. Expected. Since each MS DU has a certain overhead associated with it, the present invention can result in a greater number of successful MSDU transmissions, thereby reducing retransmission overhead. Become.

[0042] This will be described with reference to FIG. Reference numeral 600 in FIG. 6A illustrates a conventional aggregate MPDU in which a plurality of component MPDUs A to D represented by 602 force 60 5 are arranged in a random order, where “random”. This phrase suggests something that has nothing to do with the embodiment of the present invention. Note that the term “component” is used, which means the component of the frame and is synonymous with the compartment. Reference numeral 610 in FIG. 6B indicates an aggregate M PDU generated using the technique of the present invention. Here, a plurality of component MPDUs A to D represented by 612 to 615 are represented. These are arranged in the order of short MSDU length. The aggregated frame header is represented by 601 and 611 in the aggregated MPDU 600 and aggregated MPDU 610, respectively.

[0043] A schematic configuration for realizing the allocation assignment as described above will be described with reference to the drawings. Figure 7 illustrates a wireless LAN transceiver (Transino system 700 as a wireless network adapter for network node communication, with reference numeral 701 The MAC processor 702 represents a PHY (physical layer) modem. Reference numeral 703 denotes an interface between the MAC and the PHY, and reference numeral 704 denotes an antenna array (which functions as an interface between the modem and the transmission medium (line)). Similarly, reference numeral 705 denotes an interface between the MAC processor and an upper layer of the network node, and a control signal, transmitted data, and received data are transmitted through the interface.

[0044] The aggregation mechanism described above and described in Non-Patent Document 3 is implemented in the layer 2 as a module of the MAC processor 701 when viewed in accordance with the OSI network architecture. Will be done. This embodiment describes a mechanism for scheduling a plurality of MPDUs in an aggregated frame, and its configuration is a module of the MAC processor 701 that embodies this embodiment, that is, It will be realized as a scheduling control unit 706.

Furthermore, the processing functions of the scheduling control unit 706 as a configuration example for realizing the present embodiment in the wireless transceiver will be schematically outlined.

[0046] First, in the compartment length information acquisition determination process, the scheduling control unit 706 acquires information on the compartment length indicating the length of each compartment data unit, and based on this acquired information, each compartment data Unit size Z length is determined, and the compartment data unit size is arranged in ascending order. Based on the determined size, each compartment data unit is assigned in ascending order, that is, in the ascending order. Arrange. This process can be handled programmatically.

As described above, according to the scheduling control unit 706 of the present embodiment, when a plurality of data units are aggregated into one frame and wireless transmission is performed, a plurality of component MSDUs are included in the aggregated MPDU. The shortest of these, the compartment M PDU, is scheduled in ascending order of their packet lengths in such a way that the compartment M PDU is arranged at the top of the aggregate MPDU. As a result, the component MPDU having a relatively small size within the aggregated frame is relatively relative to the position portion where the reliability is expected to be relatively high in the main body portion of the aggregated MPDU, that is, relative to the head portion thereof. Scheduled to the nearest part. In view of the fact that each MSDU has a certain overhead associated with it, according to the present embodiment, as a result, successful transmission can be obtained for a larger number of MSDUs. As a result, it is possible to reduce the retransmission overhead. This is an effective way to deal with radio channel fluctuations that have been carefully considered in the conventional frame aggregation technology by allocating data units that take this into account. It reduces the retransmission overhead for errors caused by changes in channel conditions and reduces the access overhead by using frame aggregation technology.

[0048] (Embodiment 2)

In the second embodiment of the present invention, the allocation and allocation of data units based on the priority, and further the allocation and allocation of data units based on this priority are allocated and allocated to the data units based on the packet length of the first embodiment. Arrangement allocation combined with the above will be described.

[0049] In the former allocation allocation according to the present embodiment, a data unit having a higher priority has higher reliability in a transmitted packet (also in view of the potential for bit error). Multiple data units to be transmitted within the same wireless packet are scheduled according to their priority in such a way that they are allocated to segments.

[0050] The latter allocation assignment according to the present embodiment is the method described in the above-described two embodiments, that is, a method of integrating an array based on weights based on priority and an array based on data unit lengths. In consideration of several factors, the data unit is weighted and both the weight and size of the data unit are transmitted. It ’s like a target.

[0051] First, a case will be described where a plurality of component MSDUs to be aggregated have different priorities. In such cases, multiple MSDUs are weighted based on their relative priority. In addition to or in addition to priority, the weight is the number of retransmission attempts for the current MSDU (the current MSDU), its length, the corresponding MSD It is not limited to the service type of the traffic stream to which u belongs and this embodiment! It can be calculated based on any other means!

[0052] The following is only an example that is not intended to limit the present embodiment in order to understand the present embodiment, and an appropriate response is given as an example of such optional weighting. In order to generate a high-weighted control frame that requires a certain amount of processing delay on the receiver side, it may be possible that the transmission is performed in the future network in time. It may be such that a frame that affects the state (which is often a control frame, but not necessarily a control frame) is given a high weight. Therefore, such a transmission will require higher reliability and, on the other hand, will guarantee it.

[0053] Multiple compartment MSDUs, of which the lower weight is assigned to the less reliable part of the packet, whereas the higher weighted compartment MSDU Arranged so that it is assigned to the more reliable part of the aggregate MPDU. Multiple compartments with the same weight MSDU can be arranged in ascending order of MSDU size!

[0054] FIG. 8 is a flowchart describing an algorithm for generating one aggregated MP DU from a plurality of compartments MSDUs according to the present embodiment. This figure describes one of the technologies according to the present invention for dealing with the arrangement Z scheduling of multiple components MSD U in an aggregate frame structure for the purpose of improving throughput. It is directed to prioritizing traffic by both type Z class of multiple component packets and their size.

[0055] When a transmission opportunity (TXOP) is acquired or granted in step 801 in the figure, a weight for an MSDU that can be transmitted in the current transmission opportunity is calculated in step 802, and They are assigned. Based on the rule that higher weights are assigned to MSDUs with higher priorities, in step 803, the MSD Us are sorted in order of their weights, from highest to lowest. For a plurality of MSDUs given the same weight in step 802, the second sort operation is executed in ascending order of the length of the corresponding MSDU. As a result, multiple compartment MPDUs Depending on their relative weights, they will be placed in reliability dependent zones, with shorter weights being closer to the PLCP preamble in the relationship between each weighted location. Thus, the packets are arranged in ascending order of packet length. After completion of the process described in step 804, in step 805, a compartment MPDU is generated for each MS DU, and these compartment MPDUs are aggregated to generate an aggregate MPDU for output generation. .

[0056] If the priority allocation method is not adopted! /, In some cases, the weight allocation process described in step 802 may be replaced with a process that gives the same weight characteristic to all packets. This will be apparent to those skilled in the art. As a result, the first sort operation 803 has no effect, and as described in the first embodiment of the present invention, the plurality of compartment MSD U forces arranged in ascending order of the MSDU size in step 804 are also included. Aggregate MPDU power that results in the generated. Alternatively, the same effect can be achieved by omitting steps 802 and 803 in the flowchart 800. That is, this flowchart is a process chart that is a combination of sorting by weight and sorting by length, and is illustrated as a supplement to the description in the present embodiment. As described above, in step 802 By assigning the same weight characteristic or simply omitting steps 802 and 803, this figure also shows the sort processing corresponding to the first embodiment.

[0057] Furthermore, the processing functions of the scheduling control unit 706 as a configuration example for realizing the present embodiment in the wireless transceiver will be schematically outlined.

[0058] First, the scheduling control unit 706 performs, in the weight calculation process, the priority, the number of retransmission attempts of the current target MSDU, the length thereof, the service type of the traffic stream to which the MSDU belongs, and the present embodiment. Calculate and assign a weight for each compartment data unit based on any other means not limited. In the weight descending order arrangement processing, the compartment data units are arranged in descending order of the assigned weights (first sort processing). On the other hand, the scheduling control unit 706, in the compartment length information acquisition determination process, acquires information on the compartment length, which indicates the length of each compartment data unit, and acquires this information. Based on the information, determine the size Z length of each compartment data unit. In the ascending order processing for the compartment data unit size, the data after weighted descending order processing is used as the second sorting process for two or more compartment data units that have the same order of arrangement by weights for the first sort. For the units, the compartment data units are arranged in ascending order based on the size of the determined size. These processes can be handled programmatically.

As described above, according to the scheduling control unit 706 of the present embodiment, when a plurality of data units are aggregated into one frame and wireless transmission is performed, the plurality of compartment MPDUs are relative to each other. Depending on the target weight, it is placed in the position part that has a relationship with the reliability, and in the relation between each part with the same weight, the packet length is set so that the shorter packet is closer to the PLCP preamble. Arranged in ascending order. As a result, the component MPD U force having a relatively high priority within the aggregated frame is relative to the position of the main part of the aggregated MPDU that is expected to be relatively reliable, i.e., relative to the head of the position. Is scheduled to the closest location. As a result, according to the present embodiment, it is possible to obtain more stable transmission reliability and robustness for a component MPDU having a relatively high priority. This is because the conventional frame aggregation technology effectively copes with the fluctuation of the radio channel with much consideration, by arranging the data units in consideration of this, This improves transmission reliability against errors caused by changes in channel conditions and reduces the access overhead by using frame aggregation technology.

[0060] (Embodiment 3)

In Embodiment 3 of the present invention, a dynamic fragmentation method will be described, and more specifically, for a plurality of fragment data units, their optimum sizes that will satisfy the requirements of a predetermined threshold fragment error rate are set. Describes techniques that can be used to ensure that the probability of MPDU fragment errors is below a given threshold, regardless of location within the aggregate MPDU. [0061] The present embodiment relates to a dynamic fragmentation scheme, in which the fragmentation threshold of one data unit is the position of the data unit in the context of the transmitted packet. That is, a smaller threshold is taken as the channel drift increases. The purpose of this embodiment is to maintain a uniform fragment error rate within the context of a large data unit, which ensures that transmission succeeds for data units that are defragmented at the receiver side. It is intended to be

[0062] First, in summary, FIG. 9 illustrates a dynamic fragmentation mechanism embodied by the present invention, which provides a common threshold for all fragments. To achieve the error rate, it is shown that the fragmentation threshold is dynamically reduced as the training sequence separation increases (that is, the channel drift increases and the BER increases). Yes.

[0063] In the present embodiment, the transmitted MSDU is fragmented before the generation of the aggregate MPDU. Non-Patent Document 2 described above describes a mechanism for fragmentation. In the description of the present invention, the purpose of fragmentation is to “aggregate” the same aggregated MPDU. , Mapping one MSDU to multiple fragment MSDUs. As will be described subsequently in the present embodiment, the present technology differs from the technology of Non-Patent Document 2 in that fragmentation is performed using a dynamic fragmentation threshold. RU

[0064] As described above, each compartment MPDU has an individual FCS, which enables selective retransmission, and the dynamic aggregation proposed in the present technology. The objective is to ensure that the MPDU fragment error probability is below a predetermined threshold, regardless of its location in the aggregated MPDU.

[0065] The error probability of one MPDU fragment is expressed by the following equation (1).

[Number 1]

,

\ -BER pktPosn) j (1) [0066] In the above equation, P is based on the training sequence Z packet head

fragErr

Fragment position is a function of pktPosn. Similarly, BER (or its average value [ mean]) is also a function of pktPosn, as demonstrated in Figure 3. fragSize defines the number of bits in the fragment.

[0067] Equation (1) is needed to maintain a certain threshold error probability P for fragments. FragErrThresh

Can be rewritten to represent the correct fragment size, in which case the following equation (2) is obtained.

[Equation 2] j- _c . LOg ^ l-^ fragErrT resh) r> \ fragSize = —— / ~~ r, \

"log l-BERipktPosn))

[0068] By using the method described in Equation (2), it is possible to determine the optimal size of each of the MPDU fragments that will satisfy the threshold fragment error rate requirement. However, as demonstrated in Figure 3, it should be properly understood that the average BER expressed in equation (2) is a function of the packet position. BER also depends on environmental conditions (multipath) and terminal mobility. Furthermore, as shown in Fig. 3A and Fig. 3B, the absolute value of the BER also depends on the SNR at the receiver. Therefore, since the transmitter does not have an accurate statistical value of the received BER, this value may be selected empirically using external information feedback or reciprocity, or the transmission It can be foreseen that this selection may be made using a look-up table of pre-calculated values on the machine side. It is expected that, among other parameters, such information will be related to frame error rate statistics or fragment error rate statistics and SNR.

[0069] Here, referring again to FIG. 9, this figure shows the MPDU900 composed of the header 901, the MSDU 902, and the FCS 903 force described in the present embodiment in FIG. The effect when applied to is described. By applying this technique, a new set of aggregated fragment MPDUs 911 to 914 (these are collectively represented by 910) can be configured based on MSDU 902 of MPDU 900. Note that according to the technique and equation (2), 911 has the largest size, while 914 is the smallest.

[0070] In order to fully understand the effect of the dynamic fragmentation technique of the present embodiment, the channel drift effect shown in Fig. 4 is also related to each MPDU (911 to 914). Show. 920 in the figure represents a channel estimation value derived from the training sequence at the beginning of packet transmission (not shown in the figure, see 401 in FIG. 4), and 921 represents the estimated value power. Represents the actual channel drift.

[0071] Shown to use fragment size that tends to be smaller at the tail of aggregated packets, taking into account the previous description, causing additional overhead, ie FCS and compartment headers required for each fragment Such application of the dynamic fragmentation method according to the embodiment is expected to make a wise decision.

[0072] In the present embodiment and the description of Fig. 9, it is suggested that dynamic fragmentation is applied to an existing MPDU. This description is written to make it easier to visualize the effects of dynamic fragments and represents the effects of additional header overhead and additional FCS overhead associated with dynamic fragments. In actual implementation, this embodiment can be realized by omitting the step of generating M PDU 900 and by fragmenting component MSDU 902 itself before configuring the aggregate 910 of multiple fragment MPDUs. That should be understood.

[0073] As described in the dynamic fragmentation method, and the method power of the fragment size that exponentially decreases based on the equation (2). In doing so, some sort of compromise can be made, for example, a single MSDU! /, And even fragment sizes can be used to generate multiple fragment MPDUs. In keeping with the spirit of the present invention, the fragmentation threshold taken for the following multiple component MSDUs will be smaller.

[0074] As another form of implementation, the MSDU size sorting method described in the above-described embodiment may be used. In this case, while using this, the expression (1 ), The dynamic fragmentation scheme may be applied to larger MSDUs that are placed in locations that are expected to have a higher probability of error.

The configuration embodying the present embodiment will also be realized as a scheduling control unit 706 that is a module of the MAC processor 701.

[0076] In the wireless transceiver, the processing function of the scheduling control unit 706 as a configuration example for realizing the present embodiment will be schematically outlined. [0077] The scheduling control unit 706 first uses the method described in Equation (2) in the optimum sizing operation processing based on the dynamic fragmentation threshold shown in FIG. Determine their optimal size that will meet the threshold fragment error rate requirement. Next, in the fragmentation process by the optimum size, the fragmentation by the decision size is performed. Furthermore, in the aggregate configuration process of multiple fragments, a plurality of fragment MPDUs are “aggregated”.

As described above, according to the scheduling control unit 706 of the present embodiment, when a plurality of data units are collected in one frame and wireless transmission is performed, a predetermined threshold fragment is obtained for the plurality of fragment data units. Determine the optimal size of those that will meet the error rate requirements and ensure that the MPDU fragment error probability is below a predetermined threshold, regardless of its position in the aggregate MPDU. ! /, The effect is obtained. The present embodiment effectively copes with radio channel fluctuations by performing fragmentation that takes this into account, and improves transmission reliability against errors caused by channel state changes. As a result, it has the effect of improving the throughput.

[0079] Although the above description has been given as a preferred embodiment of the present invention, the present invention is not limited to the embodiment disclosed herein, and various modifications, In addition to this, it is possible to implement modifications. The scope of the present invention is determined by what is indicated in the description of the scope of claims and the scope of equivalents thereof. Industrial applicability

[0080] The present invention can be applied to a transmission method used for packet-type wireless terminal equipment that aggregates a plurality of data units into one frame or performs wireless transmission collectively.

Claims

The scope of the claims

[1] For packet-type wireless transmission, a transmission method in which a plurality of data units are aggregated into one frame and transmitted.

An arraying step in which each data unit is arranged in order from the top of the frame body in the order of short length;

A transmission step of collectively transmitting the arranged data units.

[2] In packet-type wireless transmission, a transmission method in which a plurality of data units are aggregated into one frame and transmitted.

A weighting step of calculating a weight for each data unit based on the priority of each data unit, and assigning the calculated weight to each data unit;

An arrangement step of arranging each of the data units in order from the top of the frame body in order of the weight!

A transmission step of collectively transmitting the arranged data units.

[3] In the weighting step, in place of or in addition to the priority, in addition to the priority, the number of retransmission attempts of the data unit that is the current transmission target, the length of the data unit that is the current transmission target 3. The transmission method according to claim 2, wherein the weight is calculated based on at least one of service types of a traffic stream to which a data unit currently being transmitted belongs.

[4] For packet-type wireless transmission, a plurality of data units are aggregated into one frame and transmitted.

A first arrangement step of arranging each of the data units in the descending order of the assigned weight, A second arrangement step of arranging the two or more data units determined to have the same weight in the first arrangement process in order of increasing length;

A transmission step of collectively transmitting the arranged data units.

[5] In the weighting step, instead of or in addition to the priority, in addition to the priority, the number of retransmission attempts of the data unit that is the current transmission target, the length of the data unit that is the current transmission target 5. The transmission method according to claim 4, wherein the weight is calculated based on at least one of service types of a traffic stream to which a data unit currently being transmitted belongs.

[6] For packet-type wireless transmission! A transmission method that combines a plurality of fragments into one frame,

For each fragment, a decision step to determine the size of each fragment to achieve a uniform fragment error rate;

A transmission method comprising: a fragmentation step for fragmenting a data unit according to the determined size; and an assembly step for consolidating a plurality of fragments resulting from the fragmentation into one frame.

7. The transmission method according to claim 6, wherein the determining step determines a larger size for a fragment closer to the head of the frame body.