WO2010117646A1

WO2010117646A1 - Retransmission technique for a communication network

Info

Publication number: WO2010117646A1
Application number: PCT/US2010/028568
Authority: WO
Inventors: Naveen Arulselvan
Original assignee: Motorola, Inc.
Priority date: 2009-04-09
Filing date: 2010-03-25
Publication date: 2010-10-14

Abstract

A system and method for retransmitting data in a communication network includes a first step (500) of establishing packet transmissions arranged into transmission and retransmission intervals, and a repair-packet permutation rule wherein repair packets are derived from packet data of the transmission interval XOR'ed in a permuted sequence. A next step (502) includes transmitting the packet data. A next step (503) includes receiving feedback about packets received in error. A next step (504) includes sending R repair packets. A next step (506) includes receiving repair packets. A next step (508) includes recovering the error packet from one of the repair packets.

Description

RETRANSMISSION TECHNIQUE FOR A COMMUNICATION NETWORK

RELATED APPLICATION

This application is related to co-pending and co-owned patent application number 2125/DEL/2008, filed September 10, 2008, which is incorporated by reference in its entirety herein.

FIELD OF THE INVENTION

The present invention relates generally to wireless communication networks, and in particular, a technique to provide a retransmission of data in a communication network.

BACKGROUND OF THE INVENTION

Multimedia and group communications are becoming more important aspects of telecommunication networks, and the demand for such services will continue to increase. For instance, there are presently many different systems and networks driving 3GPP/3GPP2/IEEE to provide group communication and efficient broadcast support in a growing popularity of applications in next generation wireless access network such as LTE, UMB, HSDPA, DO-A, and 802.16x, etc. Mobile broadcast services such as mobile television, live sports/ news-casting are expected to be popular in 4G networks. Accordingly, a suite of protocols has been developed for use in broadcast communications. These protocols are used to control multimedia broadcast and multicast service (MBMS) communication sessions that include data streams such as audio (voice), video, text messaging, and internet protocols, for example between, or to, user equipment (also referred to as subscriber stations or mobile stations) in a communication network.

A broadcast group communication has the efficiency of delivering one set of informational streams to multiple UEs depending upon the broadcast radio link characteristics and capabilities of the UEs in the call. This makes MBMS very efficient on the downlink. Users subscribed to a broadcast service, may be distributed throughout the cell and as a result experience varying channel conditions. Conservative broadcast coding schemes which cater to the users with worst channel condition will invariably affect the overall system throughput. For example, an evolved Node B (eNB) or base station, knowing the minimum capabilities of the user equipment (UE) it serves, can broadcast a current operating modulation and coding scheme (MCS) to the UEs in its cell directing the UEs to use that MCS to decode the MBMS service provided by the eNB. However, the MBMS system has limited bandwidth available for feedback information from all of the UEs, i.e. only a few dedicated uplink feedback control channels are provided, which can not accommodate complete feedback from all the UEs.

Retransmission schemes have been suggested to improve reliability in broadcast transmissions. However, these schemes rely on uplink feedback and have various implementation issues. One issue is the provisioning of an uplink feedback channel in broadcast settings which enables users to send information about the packet delivery status. In this case, although uplink feedback is accommodated, there is a lack of uplink bandwidth as previously described. Another issue is devising a strategy to minimize feedback overhead while collecting information representative of all users in the cell. This also suffers from limited feedback bandwidth.

Therefore, a need exists to provide a retransmission technique for broadcast data stream delivery in a communication network to enable efficient error correction for a call. It would also be of benefit to provide this error correction using limited uplink resources.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention is pointed out with particularity in the appended claims. However, other features of the invention will become more apparent and the invention will be best understood by referring to the following detailed description in conjunction with the accompanying drawings in which:

FIG. 1 illustrates a simplified block diagram of a network, in accordance with the present invention;

FIG. 2 is a table of potential packet errors and correctable errors, in accordance with the invention in its most basic form;

FIG. 3 is a table of potential packet errors and correctable errors, in accordance with a first embodiment of the present invention;

FIG. 4 is a table of potential packet errors and correctable errors, in accordance with a second embodiment of the present invention; FIG. 5 is a table of potential packet errors and correctable errors, in accordance with a third embodiment of the present invention;

FIG. 6 illustrates a method, in accordance with the present invention; and

FIG. 7 shows a graph of an improvement provided by the present invention.

Skilled artisans will appreciate that common but well-understood elements that are useful or necessary in a commercially feasible embodiment are typically not depicted or described in order to facilitate a less obstructed view of these various embodiments of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

The present invention provides a system and method for providing a retransmission technique for broadcast data stream delivery in a communication network to enable efficient error correction for a call. The present invention provides this error correction benefit using limited uplink resources such that uplink feedback overhead is minimized while collecting feedback information representative of all users in the cell. In particular, the present invention provides a retransmission scheme based on network-coding for MBMS services that enhances the transmission reliability. The evolved NodeB (eNodeB or eNB) transmits N data packets at a time over the broadcast channel, followed by R repair packets in a repair interval. The repair interval consists of R repair slots, which can be slots in time and/or frequency. In every repair slot, the eNodeB transmits a pre-determined combination of a subset of the TV packets transmitted earlier. Each of the TV packets goes out exactly once in the repair interval. Users can correct their erroneous packets in a memory-less manner using these repair packets, as will be detailed below.

The wireless access network is designed to support the delivery of a broadcast stream to a group of user equipment (UE) nodes through a serving base station. Preferably, this is accomplished in a secure manner supporting confidentiality, authentication, and integrity of the multicast data stream. Although described herein in a 4G Long Term Evolution (LTE) embodiment, it should be noted that the present invention is also applicable to other communication technologies including, but not limited to, High Speed Downlink Packet Access (HSDPA) and WiMAX. It is possible to implement the present invention in a separate entity, as part of a Radio Network Controller (RNC), or in a base station (eNB) as is described for the example herein.

FIG. 1 illustrates a multicast communication used as an example of the present invention described herein. An evolved Node B (eNB) 106 or base station serves a plurality of various user equipment (UE) 100, 102, 104 in the cell site and sends broadcast communications 108 to the UEs. Each eNB and UE includes a transmitter, receiver, memory, and processor utilized in the implementation of the present invention. The base station 106 can also send a group management message 110 containing data transfer parameters and rules to the UEs such as to make the UEs aware of the network coding being used in the broadcast. In this example, the network provides limited uplink resources 112 to receive feedback from the UEs 100, 102, 104.

The present invention provides various techniques which works effectively with limited resources for uplink feedback, which is expected for some types of MBS usage scenarios, such as interactive television games, pervasive multiplayer games, mobile television for file downloading with voting, betting broadcast services, auction broadcast services, trading broadcast services, etc. In particular, the present invention is based on network coding/fountain coding and exploits the fact that error packets tend to be distributed among the various cell UEs. The network coding includes systematic codes in which the original packets are transmitted followed by repair packets which are pre-determined combinations of the packets already transmitted, in accordance with predefined permutation rules, as will be defined below, and in response to various forms of uplink feedback.

A first embodiment utilizes shared feedback using a common feedback channel where the total number of errors for each packet is known via a polling mechanism. A second embodiment is a limited enhanced scheme that requires unicast UL feedback from users to report only the number of received errors. A third embodiment is a full enhanced scheme that requires unicast UL feedback from users to report the actual incorrectly received packets. Any of these embodiments can be used and are only dependent upon the particular uplink resources that are defined in the communication system. All of the embodiments use a permutation rule using various combinations of originally transmitted packets, wherein each combination is re-transmitted in a repair slot. The permutation rule allows a user equipment to recover one packet error among each combination. If the user equipment originally received more than one error packet out of each retransmitted combination, the user equipment will not be able to recover either error. The present invention provides various techniques to maximize the amount of recoverable errors by user equipment using the permutation rules defined below.In the operation of the present invention, if UE A receives packet P₁ in error and P₂ correctly, while UE B receives packet P₂ in error and packet Pj correctly, transmitting one repair packet of Pj XOR' ed with P₂ will enable both users to decode their respective error packets correctly. For example, eNodeB can transmit JV packets given by Q= (Qi, Q₂, ...., Q_NJ- After the TV packets are transmitted, there is a repair interval of R packets (for some value of R less than N). A pre-determined permutation rule (PR) maps the original sequence Q to P where the original data packets are permuted and then XOR' ed. Different permutation rules can be used that are pre-defined in the system and are known, defined, and/or report to user equipment. In a basic permutation scheme //-packet permuted sequence P is sent out by the eNodeB as follows:

P ® P, © .... © P,

— as the first repair packet,

i^> M ©i ^x> M ₇ © ®P ^x 2N_. as the second packet and so on till,

R R R

p (R-UN "-+ ,1 ^{Φ P}UUW +2₀ ® ®^PN in the R^th repair packet

R

The present invention offers several advantages; a) decoding at the UE is memory-less, which results in greatly reduced complexity on the UE side, as there is no need to buffer repair packets, b) there is low over-the-air overhead, where parameters such as the number of transmitted packets N, repair interval packets R, and the permutation rule PR can be transmitted just once over the broadcast control channel (BCCH), c) like in any systematic code, good users do not have to wait till the last encoded packet and can decode quicker, d) provides significant performance improvement compared to schemes with no network coding, e) scores higher over a conventional code's performance when file sizes are small or when streaming applications demand a short playback buffer, and f) can be flexibly implemented in the eNB with limited uplink resources.

The performance of the above technique degrades slightly in case error rates are very high or the retransmission interval is tight. Therefore, a variation to the basic scheme is suggested then, wherein d (a random number of) packets are dropped from the original JV packets for the repair interval transmissions. The truncated packet sequence Q^-_d is permuted as before to obtain the retransmission sequence P^-_d- This is transmitted in R repair slots, (N-d)/R packets at a time.

Referring to FIG. 2, a chart is shown demonstrating errors that are correctable by the present invention in its most basic form when there are no uplink resources available in the system. Given ten UEs to which one hundred packets are to be sent, the repair interval includes five repair slots. For ease of illustration, identity permutation was applied. The eNodeB XOR' s groups of packet numbers 1-20, 21-40, 41-60, 61-80, 81-100 together into five repair packets respectively, which are sent in the repair interval. A user can recover a packet received in error from one of the repair slots by knowing from the permutation rule that defines which slot contains the packet received in error. For example, UE A has received the eighteenth packet (out of one hundred sent in the transmission interval) in error. UE A knows that there are five repair slots in the repair interval, and that the eighteenth packet was XOR'ed in the first repair packet from the permutation rule. UE A can then XOR the correctly received packets 1-17 and 19-20 with that first repair packet to recover the correct eighteenth packet, as long as no other packets used in that repair packet were received in error by the receiver. Therefore, UE A will not be able to recover packets 23 and 26 received in error by using the second repair packet. As a result, UE A will only be able to correct three out of its five received error packets, as shown.

The column "Correctable Errors" shows the number of errors that can be corrected among the total number of errors reported by all the users. If two or more packets in the same group are in error, they cannot be corrected. This example shows that for the forty-seven received in error, the present invention in its basic form will correct twenty errors in all using only five repair slots.

In the first embodiment, a common feedback channel is assigned for all broadcast users. This feedback can be used to improve the number of correctable errors. In particular, a voice-vote polling mechanism is used by the eNB to estimate the number of users (i.e. receivers) receiving a particular packet in error, the same as or similar to a not acknowledged (NACK) message. In other words, all users are polled on a common feedback channel after every packet is sent, and the users are able to communicate on the common feedback channel how many of the users received that packet in error. Referring to this operation in FIG. 3, packet 55 was reported in error by 3 users while packet 1, packet 18 etc were reported in error by 2 users and packet 2, packet 3, packet 4 etc were reported in error by 1 user.

In this embodiment, the eNB sorts the reported errors on every packet from high to low, and calculates the number of packets that will be transmitted in every repair slot. For example, when there are 40 packets in all that were reported in error and 5 repair slots, the eNB attempts to transmit 8 packets in every repair slot. Then the eNB XOR's the packet having the highest number of reported errors (e.g. 55) and a number of packets with the lowest number of reported errors (e.g. 23, 53, 66,69,71, 95 and 85), and then sends this permuted combination to the user equipment in a first repair slot. This would then be repeated for the packet next highest number of reported errors along with packets with low number of errors reported (eg., 7 packets with 1 error reported) in the next repair slot, and so on, until all the error packets have been permuted into repair packets. For example, N packets are broadcast transmitted by the eNB to K users, followed by a retransmission interval of R packets. Using voice-vote polling, via common feedback channel, the K users inform the eNB how many of the K users have received a particular packet in error. For example, packet Pi is received erroneously by K₁ users, packet P₂ was received erroneously by K₂ users, and so on. Assuming that there are a total of N' (<= N) packets received in error, the eNB can sort that the number of users receiving each packet in error in a list; K₂ <=...<=K_N\ The eNB can then schedule network-coding based retransmissions using the NACK count for each error packet, P₁ through P_N- If t is the number of packets eNB will send in each repair slot, packets Pi and (P^; P_N'- i,..P_N:_t-i} would be XOR' ed into a first repair packet; packets P₂ and {PN'-t--PN'-2t-i} would be XOR'ed into a second repair packet, and so on. These repair packets would be communicated to the user equipment along with the packet numbers used in each respective repair slot. In this embodiment, the total number of errors in the system is

N' given by the sum of the list, E = ∑K_t . Using the above technique with the error in i=l

FIG. 3 has shown an improvement in the number of correctable errors to 25 out of 47, as will be shown in FIG. 6 below. It should also be noted that idle-mode user equipment can participate without the need for a Cell Radio Network Temporary Identifier. In the second embodiment, and referring to FIG. 4, unicast UL feedback is received from users to report only the number of received errors in a packet group in a multicast. As in the basic permutation scheme, the eNB transmits N packets to K users: P= [Qi, Q2, - -, Q_N] - After the N packets are transmitted, there is a retransmission interval of R packet-slots (R<N). Pi, P ₂, ... P_R are the set of packets that are to be network coded and sent in each retransmission slot. Each retransmitted packet P₁ is the XOR of exactly N/R original packets. Each original packet Q₁ appears only once in the network-coded retransmission sequence [P₁] . Different permutation rules can be used and compared to select the rule providing the most correctable errors. The selected permutation rule PR:Q->P is pre-defined and is known or reported to all users.

In this second embodiment, every user 7 informs the eNB of only the number of errors in the group of constituent packets of the network-coded packet P₁ : N(J₇P₁). The total number of errors to be corrected from the retransmitted packet P₁ is

K

N(P₁ ) = ∑ N(j, P₁ ) , and the maximum uplink feedback needed is KR log₂(ΛW?).

Based on error count, algorithm at the BS will send retransmission packets with appropriate network coding. For example, referring to the last column of FIG. 4, "Group Feedback #", User 1 reports that it had one error within the group of packets 01-20, two errors in group 21-40, one error in group 41-60, one error in group 61-80, and no errors in group 81-100. Two of its errors (in Group 21-40) can not be corrected in the basic permutation scheme.

However, in accordance with this second embodiment, the eNB can then compare the basic group scheme of permutation (see FIG. 2) to the different permutation schemes to see if a higher number of correctable errors can be achieved. For example, instead of Group 01-20 and Group 21-40, the eNB could define a first group of odd number packets between 01 and 40 and a second group of even number packets between 01 and 40, and so on, to provide interleaved groups. Alternatively, various random interleaving schemes can be used. If such an odd/even or interleaved group scheme provides more overall corrections, then this permutation scheme (e.g. coding scheme #2) would be used, defined for, and reported to the user equipment to use instead of the basic scheme (e.g. coding scheme #1). It should be recognized that many different schemes can be predefined, wherein the eNB would just provide the repair packets along with the scheme number used to combine those repair packets to user equipment.

In operation, for the proposed random interleaved permutation scheme, CQ, P₁) is defined as number of correctable errors for usery after receiving packet P₁. In fact C(j, P₁) =1 if N(j,PJ = l and CQ, P₁) =0 otherwise. The total number of correctable

K errors for all users after receiving P₁ is C(P₁). C(P₁ ) = ^ C(j, P₁) is deterministic.

The eNB then sorts the proposed retransmission P₁ packets in ascending order of C(P₁)ZN(P₁). Any ties are broken by the higher number of reported errors i.e., N(P₁). The reordered packet sequences are denoted as { T₁) where i=\, 2, ,..R. It should be noted that each packet T₁ is going to be made up of N/R packets from the original packets [Q₁]. The eNB then defines an interleave operation on T₁ and T_h where {U) and {t_j} are the original packets that constitute retransmission packets T₁ and T₁ respectively. The eNB interleaves by randomly spreading the two sequences {U) and {t_j} to form two equal-length packet sequences {tj '} and {t, '} . The new packets {tj '} and {t_j '}will be network-coded packets to form T₁ ' and T/, respectively. It should be noted that the number of correctable errors in interleaved packets T₁ ' and T₁ 'is random. Starting with T₁, the eNB computes the set S₁=[J ] E[C(T_j ')+C(Tj ')] > C(T_j) +C(T₁) }, where T₁ is a network-coded packet consisting of N/R original packets. If₁S; is a null set, only T₁ is sent as a retransmission. The best packet sequence to interleave with T₁ is given by i* = arg max, _{m S}iEfC(T_j ')+C(Tj ')]. [t}} and [U_*] are interleaved before the eNB retransmit to the user equipment. T₁ and T_1* are eliminated for successive iterations, which are repeated until no sequence is left. Using the above random interleaving technique with the error in FIG. 4 has shown an improvement in the number of correctable errors to 30 out of 47, as will be shown in FIG. 6 below.

In the third embodiment, unicast UL feedback is received from users to report all the incorrectly received packets for each user in a multicast. As in the basic permutation scheme, the eNB transmits N packets to Abusers: P= [Q₁, Q2, - -, Q_N] - After the N packets are transmitted, there is a retransmission interval of R packet-slots (R<N). P₁, P ₂, ... P_R are the set of packets that are to be network coded and sent in each retransmission slot. Each retransmitted packet P₁ is the XOR of exactly N/R original packets. Each original packet Q₁ appears only once in the network-coded retransmission sequence [P₁). The fixed permutation rule of this embodiment PR'.Q- >P is pre-defined and is known to all users. In this third embodiment, the user with the most errors is given a higher priority for retransmitting errors. The repair packets are permuted such that no more than one error packet of an individual user is present in each repair packet. In addition, duplicate error packets are removed from each repair packet. Also, if a retransmitted packet is sent for one user the eNB can remove that packet from the error list of any other user. In other words, a retransmitted packet is only sent once to all users.

In operation, and referring to FIG. 5, using the same packet error example as is used in the first and second embodiments (FIGs. 3 and 4), the eNB first sorts the users with the most packets received in error from high to low. In this case, User E₈ has the most errors. Sequential error packets from each user list is placed into a respective transmission set AP. In other words, the first error packet of each user is then examined as a set AP, wherein duplicate entries are eliminated from this set and any succeeding sets. In addition, this set is examined to ensure that not more than one of the error packets in the set is present in any error packet list S of a user. If more than one error packet is found in set AP that also appear in a user list, then that packet is moved to a next set and a packet from that next set can be substituted into the first set (as long as it also is not present along with one of the error packets in the set in an error packet list of any user. The remaining packets in the set are then XOR' ed into a first repair packet. This process repeats for all the remaining error packets.

Referring to FIG. 5, the eNB first sorts the users in order of number of error packets; Eg, E_3j E₄, E₆, E_1Oj Ei₁ E_7j E₂, E_5j Eg, which are redefined as error sets Si to S₁₀. Taking the first entry in each list S gives the set {1, 2, 28, 18, 3, 18, 1, 57 8, 53}. Examining this list shows two duplicate packets; 1 and 18, which are removed from the set AP. None of the packets in this set appear in any next set so no packets are removed therefrom. It is also observed that none of the lists S₁ to S₁₀ share more than one entry with the set AP. Set AP is then XOR'ed into the first repair packet. Taking the second entry in each list S gives the set {17, 30, 41, 19, 7, 23, 4, 92, 66, 85}. Examining this list shows no duplicate packets, and none of the lists S₁ to S₁₀ share more than one entry with the set AP. None of the packets in this set appear in any next set so no packets are removed therefrom. Set AP is then XOR'ed into the second repair packet. Taking the third entry in each list S gives the set {49, 34, 51, 20, 15, 26, 29}. Examining this list shows no duplicate packets, and none of the lists Si to S₇ share more than one entry with the set AP. None of the packets in this set appear in any next set so no packets are removed therefrom. Set AP is then XOR'ed into the third repair packet. Taking the fourth entry in each list S gives the set {68, 55, 69, 62, 59, 55, 55}. Examining this list shows two duplicate packets; 55, which are removed from the set AP. It is also observed that none of the lists Si to S₇ share more than one entry with the set AP. None of the packets in this set appear in any next set so no packets are removed therefrom. Set AP is then XOR'ed into the fourth repair packet. Taking the fifth entry in each list S gives the set {71 73, 75, 65, 63, 63, 60}. Examining this list shows a duplicate packet; 63, which is removed from the set AP. It is also observed that list Si shares more than one entry, 71 and 75, with the set AP. A packet from the set must then be substituted with an entry from a next set. Packet 73 can not be used as this exists in the next set. Packet 75 can not be used as this also exists in the next set. Packet 65 can be used as it is not in the next set. Substituting packet 65 with packet 87 is not allowed as it is observed that list S₂ shares more than one entry, 73 and 87, with the set AP. However, substituting packet 65 with packet 91 is allowed as none of the lists Si to S₇ share more than one entry with the set AP. It is also noticed that packet 73 is in the next set, and as a duplicate it is removed from the next set. Set AP is then XOR'ed into the fifth repair packet. Taking the sixth entry in each list S gives the set {75, 87, 65, 95}. Examining this list shows no duplicate packets, and none of the lists Si to S5 share more than one entry with the set AP. None of the packets in this set appear in the last set so no packets are removed therefrom. Set AP is then XOR' ed into the sixth repair packet. The seventh packet 78 stands alone and has no conflicts, and it itself is sent as the seventh repair packet (i.e. not XOR' ed with anything. Comparing with first and second embodiments, the third embodiment corrects 42 out of 47 errors shown in FIG. 5 when the number of the repair slots is 5.

In practice, for the K users in the system, the set of erroneous packets are denoted by E₁, E₂, ...E_κ. Designing the optimal retransmission scheme with an interval equal to maxj IE₁I is known to be NP-complete. Sub-optimal memoryless decoding heuristic can easily implemented at the eNB. In operation, erroneous packets sets of Abusers: E₁, E₂, ...E _κ are arranged in descending orders of length. If the sorted queues are S₁, S₂, ... S_κ and S/i) denote the i^th packet in the sorted queuey, the eNB will initialize an active set (AS) and an active packet sequence (AP), where AS= [S₁) and AP=[S₁(I)). As i varies from 2 to K, if for any 7, S₁Q) is in the AP then include i in the AS. Otherwise, the eNB findsy * = min_} (S₁Q) not in set S_k) for all k in the AS. If a validy * is obtained, include i in AS andy * in AP before the eNB retransmit to the user equipment. AP is then removed from all queues S₁ for all i=\, 2,...,K, and the NACK queues S₁, S₂, ...,Sκ are sorted in descending orders of length. This operational procedure repeats till no packet is left.

FIG. 6 illustrates a method for retransmitting data in a communication network, in accordance with the present invention, which includes a first step (500) of establishing parameters for packet transmissions arranged into a sequence of N packets in a transmission interval followed a sequence of R repair packets in a retransmission interval, where R is less than N. The parameters including at least one permutation rule wherein repair packets are derived from the N packets by dividing the N packets into R groups and XOR'ing (i.e. modulo-2 adding) together the packets in each group to define respective repair packets in the retransmission interval. Each repair packet is assigned to respective packet slots in the retransmission interval. Preferably, this step includes transmitting the parameters over a Broadcast Control Channel to the users as a one-time configuration to save header overhead. The user is capable of distinguishing network-coded retransmission packet and ordinary packets from header information.

A next step (502) includes transmitting the N packets to a plurality of receivers in the transmission interval, wherein at least one of the transmitted packet is received in error.

A next step (503) includes receiving from the users (receivers) feedback about the at least one transmitted packet received in error. In the first embodiment, this step includes receiving feedback from the receivers on a shared feedback channel, which includes the number of receivers receiving a particular transmitted packet in error. In the second embodiment, this step includes receiving feedback indicating a number of received packet errors from the transmitting step in a packet group. In the third embodiment, this step includes receiving feedback from each receiver indicating which transmitted packets were received in error by that receiver.

A next step (505) includes implementing the at least one permutation rule to provide repair packets. In the first embodiment, this step includes the substeps of sorting the number of receivers receiving each transmitted packet in error into a list, XOR'ing the packet having the highest number of reported errors and the packet having the lowest number of reported errors into a first repair packet, XOR'ing the packet having the next highest number of reported errors and the packet having the next lowest number of reported errors into a next repair packet, and repeating the above step until all the error packets have been permuted into repair packets.

In the second embodiment, this step includes the substeps of interleaving packets into different packet groups in accordance with a first permutation rule. Optionally, this step includes randomly interleaving packets into different groups accordance with a second permutation rule, and estimating whether the first or second permutation rule provides the most correctable packet errors, wherein the repair packets are permuted using the permutation rule with the most correctable packet errors.

In the third embodiment, this step includes a permutation rule where no more than one error packet of an individual user is present in each repair packet. In particular this step includes the substeps of giving the receiver having the most packet errors the highest priority when permuting repair packets, sorting users per the number of reported packets received in error, each user have a list of error packets, placing sequential error packets of each user list into a respective set, eliminating duplicate error packets within the set, ensuring that not more than one of the error packets in the set is present in any error packet user list (where if more than one error packet is found in a set that also appear in a user list, then that packet is removed to a next set and a packet from that next set can be substituted into the first set as long as it also is not present along with one of the error packets in the set in an error packet list of any user), removing duplicate error packets that exist in any next set, repeating the eliminating, ensuring, and removing steps for each next set in turn until there are no more sets, and permuting each set into a respective repair packet. A next step (504) includes sending the R repair packets to the plurality of receivers in the retransmission interval after the N packets have been transmitted. In particular, each of the N packets goes out exactly once in the repair interval to enable memoryless decoding.

A next step (506) includes receiving the R repair packets by the receivers. In some embodiments, the sending and receiving steps include an indication in the retransmission interval of the particular packets being permuted in each repair packet..

A next step (508) includes recovering the at least one transmitted error packet from one of the repair packets by associating the position of the error packet in the transmission interval with the respective position of that packet in the repair packet group in the retransmission interval, and XOR' ing the correctly received packets in that group with that repair packet, as long as no other packets used in the repair packet of that group were received in error by the receiver.

Example

To illustrate the improved performance provided by the present invention, a simulation was performed using an arbitrary permutation rule. Referring to FIG. 6, the simulation results show that up to 100% of transmission errors can be corrected using the third embodiment of the present invention, and over 90% of transmission errors can be corrected using the first or second embodiments of the present invention. The first embodiment performs almost as well as the second embodiment, but has the advantage that the common feedback bandwidth is independent of the number of user equipment. Note that the present invention works equally well when users have differing packet error rates. Performance degrades slightly when the error rates are very high or retransmission interval is tight.

It should be recognized that the diagrams herein are simplified for purposes of illustrating the present invention. However, those of ordinary skill in the art will realize that many other network entities and processes may be part of the communication system, which have not been shown for the sake of simplicity. For example, the present invention can be incorporated in one or more of a radio network controller, base station (eNB), session controller, a group database manager, a registration manager, an application layer router, a group entity manager, a broadcast and unicast address manager, a policy manager, a flow controller, a media manager, and a bandwidth manager, among others, all of which are known in the art. It should be appreciated that the above described entities can be integrated in the same physical or logical network element or provided as distributed or individual physical or logical network elements.

The present invention can be implemented in the eNB, and UEs should be able to distinguish between the original packets and repair packets which are network- coded. Legacy UEs will simply ignore the repair packets. It would also be of benefit for the eNodeB to transmit network coding parameters: transmission interval, repair interval, and permutation sequence. This ensures that the header overhead in repair packets is kept to a minimum.

Advantageously, the present invention is effective and provides packet decoding without using memory. In addition, the present invention provides low UE complexity: packets can be corrected after each and every retransmission slot. Further the present invention provides low overhead: the permutation sequence PR can be fixed and conveyed to the users statically. Over-the-air header length is low as a result. The permutation sequence can be chosen such that burst errors can be corrected.

It will be understood that the terms and expressions used herein have the ordinary meaning as is accorded to such terms and expressions by persons skilled in the field of the invention as set forth above except where specific meanings have otherwise been set forth herein.

The sequences and methods shown and described herein can be carried out in a different order than those described. The particular sequences, functions, and operations depicted in the drawings are merely illustrative of one or more embodiments of the invention, and other implementations will be apparent to those of ordinary skill in the art. The drawings are intended to illustrate various implementations of the invention that can be understood and appropriately carried out by those of ordinary skill in the art. Any arrangement, which is calculated to achieve the same purpose, may be substituted for the specific embodiments shown.

The invention can be implemented in any suitable form including hardware, software, firmware or any combination of these. The invention may optionally be implemented partly as computer software running on one or more data processors and/or digital signal processors. The elements and components of an embodiment of the invention may be physically, functionally and logically implemented in any suitable way. Indeed the functionality may be implemented in a single unit, in a plurality of units or as part of other functional units. As such, the invention may be implemented in a single unit or may be physically and functionally distributed between different units and processors. Although the present invention has been described in connection with some embodiments, it is not intended to be limited to the specific form set forth herein. Rather, the scope of the present invention is limited only by the accompanying claims. Additionally, although a feature may appear to be described in connection with particular embodiments, one skilled in the art would recognize that various features of the described embodiments may be combined in accordance with the invention. In the claims, the term comprising does not exclude the presence of other elements or steps.

Furthermore, although individually listed, a plurality of means, elements or method steps may be implemented by e.g. a single unit or processor. Additionally, although individual features may be included in different claims, these may possibly be advantageously combined, and the inclusion in different claims does not imply that a combination of features is not feasible and/or advantageous. Also the inclusion of a feature in one category of claims does not imply a limitation to this category but rather indicates that the feature is equally applicable to other claim categories as appropriate.

Furthermore, the order of features in the claims do not imply any specific order in which the features must be worked and in particular the order of individual steps in a method claim does not imply that the steps must be performed in this order. Rather, the steps may be performed in any suitable order. In addition, singular references do not exclude a plurality. Thus references to "a", "an", "first", "second" etc do not preclude a plurality.

Claims

CLAIMSWhat is claimed is:

1. A method for retransmitting data in a communication network, the method comprising the steps of: establishing parameters for packet transmissions arranged into a transmission interval followed by a retransmission interval, the parameters including at lease one permutation rule wherein repair packets are derived from packet data to be transmitted by permuting the data packets and XOR' ing them into a permuted sequence; transmitting the packet data to a plurality of receivers in a sequence of N packets in the transmission interval, wherein at least one of the transmitted packet is received in error; receiving feedback from the receivers about the at least one transmitted packet received in error; implementing the at least one permutation rule to provide repair packets; sending the repair packets to the plurality of receivers in R repair packet slots in the retransmission interval after the N packets have been transmitted; receiving the R repair packets by the receivers; and recovering the at least one transmitted error packet from one of the repair packets.

2. The method of claim 1, wherein the receiving feedback step including receiving feedback indicating a number of receivers receiving a particular transmitted packet in error.

3. The method of claim 2, wherein the implementing step includes the substeps of: sorting the number of receivers receiving each transmitted packet in error into a list; XOR'ing the packet having the highest number of reported errors and the packet having the lowest number of reported errors into a first repair packet; XOR'ing the packet having the next highest number of reported errors and the packet having the next lowest number of reported errors into a next repair packet; and repeating the above step until all the error packets have been permuted into repair packets.

4. The method of claim 1, wherein the sending and receiving steps include an indication in the retransmission interval of the particular packets being permuted in each repair packet.

5. The method of claim 1, wherein the receiving feedback step including receiving feedback indicating a number of received packet errors from the transmitting step in a packet group.

6. The method of claim 5, wherein the implementing step includes interleaving packets into different packet groups accordance with a first permutation rule.

7. The method of claim 6, wherein the implementing step includes randomly interleaving packets into different groups accordance with a second permutation rule, estimating whether the first or second permutation rule provides the most correctable packet errors, wherein the repair packets are permuted using the permutation rule with the most correctable packet errors.

8. The method of claim 1, wherein the implementing step includes the substeps of: sorting users per the number of reported packets received in error, each user have a list of error packets; placing sequential error packets of each user list into a respective set; eliminating duplicate error packets within the set; ensuring that not more than one of the error packets in the set is present in any error packet user list; removing duplicate error packets that exist in any next set, and repeating the eliminating, ensuring, and removing steps for each next set in turn until there are no more sets.

9. An evolved NodeB for retransmitting data in a communication network, the eNB comprising: a memory that is operable to hold established parameters for packet transmissions arranged into a transmission interval followed by a retransmission interval, the parameters including at lease one permutation rule wherein repair packets are derived from packet data to be transmitted by permuting the data packets and XOR'ing them into a permuted sequence; a transmitter operable to transmit the packet data to a plurality of UE receivers in a sequence of N packets in the transmission interval, wherein at least one of the transmitted packet is received in error by the UE receivers; a receiver operable to receive feedback from the UE receivers about the at least one transmitted packet received in error; a processor coupled to the memory, transmitter and receiver, the processor operable to implement the at least one permutation rule to provide repair packets, wherein the processor directs the transmitter to send the repair packets to the plurality of receivers in R repair packet slots in the retransmission interval after the N packets have been transmitted such that the UE receivers can receive the R repair packets, and the UE processors can recover the at least one transmitted error packet from one of the repair packets.

10. A user equipment operable to correct data using retransmitting data in a communication network, the user equipment comprising: a memory that is operable to hold established parameters for packet transmissions arranged into a transmission interval followed by a retransmission interval, the parameters including at lease one permutation rule wherein repair packets are derived from packet data to be transmitted by permuting the data packets and XOR'ing them into a permuted sequence; a receiver operable to receiver packet data from an evolved NodeB in a sequence of N packets in the transmission interval, wherein at least one of the transmitted packet is received in error by the receiver; a transmitter operable to send feedback to the evolved NodeB about the at least one transmitted packet received in error; a processor coupled to the memory, transmitter and receiver, the processor operable to direct the receiver to receive repair packets in accordance with the permutation rule from the evolved NodeB, after the N packets have been transmitted, and recover the at least one transmitted error packet from one of the repair packets.