GB2506368A

GB2506368A - Medium access with backoff timer recovery in a wireless network

Info

Publication number: GB2506368A
Application number: GB201217213A
Authority: GB
Inventors: Patrice Nezou; Pascal Viger
Original assignee: Canon Inc
Current assignee: Canon Inc
Priority date: 2012-09-26
Filing date: 2012-09-26
Publication date: 2014-04-02
Anticipated expiration: 2032-09-26
Also published as: GB201217213D0; GB2506368B

Abstract

A communication network consists of a group of nodes, each node using a medium access mechanism with collision avoidance (CSMA-CA) to control access . A computation of backoff values corresponding to a number of time-slots a node waits before accessing the medium, is shared between the group of nodes. Each node, therefore, holds a list of expected backoff values of the other nodes. Upon reception of a message from a first node, the time of reception is checked to ensure it corresponds to an expected reception time determined from the list of expected backoff values. If the reception time does not correspond to the expected reception time, a backoff recovery procedure corrects the values of the backoff list of the first node which is considered to be in a desynchronized state.

Description

TITLE OF THE INVENTION

Medium access with backoff timer recovery in a wireless network

FIELD OF THE INVENTION

The invention belongs to the field of wireless communication networks and relates to a method and device for accessing a communication medium by a group of nodes using a backoff mechanism. More specifically, the invention provides a method for accessing the medium enabling data communication within the group of nodes.

BACKGROUND OF THE INVENTION

The 802.11 MAC (Medium Access Control) standard supports shared access to a wireless medium through a technique called Carrier Sense multiple Access with Collision Avoidance (CSMNCA) which is employed in wireless networks, such as for example wireless local area networks (WLAN). The 802.11 medium access control mainly consists for a node to wait until the medium becomes idle before trying to access the medium. So, when the wireless channel is sensed to be idle, the transmitting node is permitted to transmit data to a destination node. When the channel is sensed to be busy, the transmitting node defers its transmission.

Backoff is a known method to solve contention between different nodes willing to access the medium. The method requires each station or node to choose a random number of backoff time slots. When the medium becomes idle, a node having data to transmit (transmitting node) has to wait a backoff period corresponding to the chosen random number of backoff time slots, during which the transmitting node continues to sense the medium channel. At the end of that period and if the medium still remains idle, the transmitting node begins transmitting. Using a random period reduces the risks of collision since another node(s) waiting to access the medium would likely choose a different backoff period.

The medium channel access scheme according to the 802.11

standard specification is illustrated in figure 1.

The time unit in the 802.11 standard is called aS/of Time parameter.

This parameter is specified by the PHY layer and has a value equal to 9 ps for the 802.11 n standard.

Different inter-frame space durations 10, 11, 12 are employed to provide access to the wireless medium with different priority levels.

The Short inter-frame space 10 (SIFS) is used to separate a response frame from the frame that solicited the response, for example between a data frame and the corresponding acknowledgement response (equal to 16 ps for the 802.lln).

The PIFS inter-frame space 12 provides the next highest access priority time space after the SIFS time (PIFS = SIFS + aSlotTime).

The DIFS inter-frame space 11 defines the minimum waiting time for a transmitting node before trying to transmit some data (DIES = SIES + 2 * aSlotTime).

A node must first sense the channel for a DIES interval 11 before initiating any transmission on the wireless medium. If the wireless medium remains idle for the initial DIES interval 11, the transmitting node initiates its medium access process. It first invokes a backoff procedure using a backoff counter to count down a random number of backoff time slots 15 selected in a dedicated range [0, CW] called contention window 14. When the backoff count reaches 0, the transmitting node initiates its transmission by sending of a frame 16 (referred as "Next Erame" in the figure). In total, the medium access is deferred from a time instant indicated by arrow 17. If a collision occurs during the contention window, the transmitting node invokes a new backoff procedure by increasing the range [0, CW] in which the number of backoff time slots is selected. Such procedure is referred to Exponential Backoff Algorithm in the 802.11 standard.

The contention window OW parameter varies from a minimum value CWmin to a maximum value CWmax. The CW parameter doubles on each erroneous transmission (that is on each collision on the medium) until reaching the CWmax value. It is reset to CWmin value after each successful packet transmission.

If the wireless medium becomes busy during a backoff slot of the contention window, the backoff procedure is frozen until the wireless medium goes idle again. The backoff counting is next resumed after a DIES time.

When the backoff count reaches 0, the transmitting node initiates its transmission 16. If collision occurs, the transmitting node invokes a new backoff procedure with an increased range for choosing the number of backoff time slots.

Figure 2 illustrates the 802.11 standard RTS / CTS transmission mechanism. A source node 21 willing to transmit a data packet 24, waits until the channel is sensed idle for a DIFS 11. Then, instead of sending the packet, it preliminarily transmits a special short frame called Request-To-Send (RTS) 25.

When the receiving node (destination node 22) detects the RTS frame 25, it responds, after a SIES 10, with a Clear-To-Send (CTS) frame 26. The source node is allowed to transmit its packet only if it correctly receives the CTS frame 26.

The frames RTS and CTS carry the length information of the packet to be transmitted. This information can be read by any listening node (for example Other' node 23), which is then able to update a Network Allocation Vector (NAV) 28 containing the information of the period of time during which the channel will remain busy. Therefore, when a node is hidden from the transmitting or the receiving node, by detecting just one frame among the RTS and CTS frames 25 and 26, it can suitably delay further its transmission, and thus avoid collision.

The RTS/CTS mechanism is effective in terms of system performance, especially when large packets are considered, as it reduces the length of the frames involved in the contention process. In fact, if both transmitting nodes employ the RTSICTS mechanism, collision occurs only on the RTS frames, and it is early detected by the transmitting stations by the lack of CTS responses.

To further improve the collision avoidance, it has been proposed in US20090141738 to employ a distributed collision avoidance channel access mechanism. In such a mechanism, at the end of a transmission, each node advertises its next backoff value. All other nodes store the advertised backoff value in a reservation table to avoid collision for future transmission attempts. If a node detects a risk of collision with another node having published its next backoff value, this node reschedules its next transmission attempt. By advertising the future channel access parameters in advance, nodes reduce the number of collisions. Such a mechanism requires heavy message exchanges at the end of each transmission for the distribution of backoff values, which reduce the bandwidth capacity for data transmission. It also implies that a node must first obtain the channel access before advertising its backoff value used for its next access.

A mechanism employing a distributed medium access scheme is also sensitive to transmission errors and may become critical in wireless networks because of the hidden node problem.

This problem is illustrated by an example presented in figure 3. In this example, it is assumed that a data packet transmitted by node Nc1 is correctly received by node Nc2 but not by node Nc3. To reduce the risk of collisions, 802.11 standard specifies an extended inter-frame space called EIFS (reference 33) that is to be used to defer transmission of a node (Nc3) having received an incorrectly decoded packet 34 instead of using DIFS. The EIFS value is equal to the time for a node to receive an acknowledgement at the lowest mandatory physical transmission rate (that is: SIFS + the time to receive an acknowledgement message at the lowest rate + DIFS). Consequently the EIFS 33 is intended to prevent a node (here node Nc3) from transmitting over the ACK message 32. If the ACK 32 or any valid data packet is correctly received by Nc3 during the EIFS time, then the EIFS duration is cancelled and replaced by a DIES defer time 31 following the reception of the actual ACK 32 (here transmitted by node Nc2) or of the valid data packet.

Because of the possibility of applying either one of the EIES or the DIFS durations depending on reception conditions, desynchronization (here of the node Nc3) with respect to the backoff values stored by the others nodes may occur. The desynchronization is characterized by the fact that the actual backoff of a node (as node Nc3) differs from the backoff values known by the other nodes.

The present invention has been devised to address at least the foregoing concerns. More specifically, an object of the present invention is to provide a reliable method and device for accessing a communication medium while using a distributed collision avoidance mechanism. A further object of the present invention is to address the desynchronization of the backoff values problem and to keep operational the distributed medium access scheme in the presence of transmission errors or collisions.

SUMMARY OF THE INVENTION

In accordance with one aspect of the present invention there is provided a method for accessing a communication medium in a network which can be accessed by a plurality of nodes, each node using a medium access mechanism with collision avoidance based on a computation of backoff values corresponding to a number of time-slots a node waits before accessing the medium, wherein each node of a group of nodes among the plurality of nodes holds a list of expected backoff values of the nodes of the group, the method comprising for a first node of the group having detected a badly received packet: waiting for the reception of a message from a sending node of the group, and, upon reception of the message, checking if the reception time of the message corresponds to an expected reception time determined from the expected backoff value of the sending node contained in the list of expected backoff values hold by the first node, and, if the reception time does not correspond to the expected reception time, implementing a backoff recovery procedure to correct the values of the backoff list of the first node which is considered to be a desynchronized node, and, accessing the communication medium using the corrected backoff list.

According to this method, each node of the group which badly receives a packet is early detected as being in an unsecure state and starts watching the behavior of other nodes of the group. Upon reception of a message from a sending node of the group it checks if its backoff list is desynchronized with respect to other node/s by comparing the reception time of the message to the expected reception time of the sending node. The expected reception time is obtained from the associated expected backoff value contained in the list of the first node. The expected reception time is determined from the backoff value, which is a number of time units (e-g 9 ps for the 802.lln) as explained in the following description. If checked desynchronized, the node implements a backoff recovery procedure for recovering exact backoff values from at least one other node. It follows that a recovery procedure is started only for a node which has first been checked desynchronized. This avoids unnecessary recoveries of the backoff list, like for example at each communication or on a regular time basis, which consume bandwidth capacity.

It further avoids heavy overhead. Therefore the invention allows saving bandwidth which can be used for data communication while reducing interference risks between the nodes belonging to the group.

Advantageously, the method further provides that the backoff recovery procedure comprises sending a recovery message requesting the transmission of the backoff list from other nodes of the group and after reception of the backoff list from at least one other node, correcting the backoff list of the first node. The request for obtaining the backoff list is thereby simplified.

In one embodiment, the method further comprises for a second node of the group considered to be desynchronized and having received the backoff list of at least one other node in the transmission, correcting the backoff list of the second node. The method therefore provides correction of the backoff list of other desynchronized nodes, possibly all the desynchronized nodes.

Preferably, upon receiving a recovery message, the second node implementing a backoff recovery procedure cancels the sending of its recovery message. So, only the node implementing a backoff recovery procedure which is the first to send a recovery message will initiate a backoff recovery procedure, therefore avoiding multiple and potentially conflicting backoff recovery procedures.

The method further provides that a transmission slot is granted in response to the recovery message and at least one node of the group not considered to be desynchronized transmits its backoff list during the granted transmission slot. By this way a transmission slot can quickly be obtained and employed for transmitting the backoff list. Using a granted transmission slot allows to transmit the backoff list to all the desynchronized nodes of the group and thus to resynchronize all those nodes. No specific message dedicated to the recovery needs to be exchanged. Advantageously, the recovery message is a RTS message used in the communication network.

According to a particular embodiment, the transmission slot is divided into sub-slots, each sub-slot being assigned to a node of the group not considered to be desynchronized in order to transmit its backoff list. The idea is here to share the granted transmission slot between the nodes of the group.

Then by inserting the actual backoff list in a sub-slot, the actual backoff list can be collected and reliably transferred to the desynchronized nodes, which includes the nodes having requested a backoff recovery procedure.

Preferably, a packet including the backoff list is prepared in advance of the transmission in order to be transmitted during a transmission slot granted to the group. Thus each node can transmit its backoff list in response to a recovery message as soon as a transmission slot is granted to the group, which allows an efficient and fast backoff recovery procedure.

According to another aspect, the method further comprises for any node considered to be desynchronized: obtaining the backoff list from at least one node of the group through the recovery procedure, and correcting the backoff list of the node considered to be desynchronized with the obtained backoff list.

These steps provides for a simple resynchronization of at least one node, which has previously been checked in desynchronized state. In practice a plurality, and possibly all, of nodes checked desynchronized will obtain the backoff list and correct their backoff list. These steps further reduce to the minimum the transmissions overhead for the recovery procedure.

Preferably, the backoff list is obtained only from node(s) not considered to be desynchronized. This avoids receiving erroneous information and at the same time further reduces the transmission requirements for the recovery procedure.

According to an embodiment, the method further provides that if the reception time corresponds to the expected reception time, no correction of the backoff list values of the first node is performed. In such a case the first node can be considered synchronized with its peer nodes and therefore a transmission of backoff list is avoided, which further saves bandwidth capacity that can be used for data communication.

According to a particular embodiment, the values contained in the backoff list are predictive values distributed among the nodes of the group. The use of predictive values distributed among the nodes reduces the transmission requirements since the backoff management is distributed at the collaborative nodes avoiding therefore unnecessary transmission of backoff values.

According to a particular characteristic, each node of the group generates internally the predictive values of its backoff list. This feature reduces the bandwidth requirements to the minimum because the backoff management is distributed to the collaborative nodes and no extra information exchange is required.

The invention also relates to a node for data communication with other nodes of a group of collaborative nodes of a radio network.

The invention also relates to a non-transitory computer-readable medium storing a program which, when executed by a microprocessor or computer system implements the method steps.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention will now be described, by way of example only, and with reference to the following drawings in which: Figure 1 illustrates basic channel access timings according to the 802.11 standard: Figure 2 illustrates a RTS/CTS mechanism according to the 802.11 standard; Figure 3 illustrates different ways to receive one data packet and how it influences the timing of the 802.11 standard; Figure 4 is an example of a communication system where the present invention can be implemented; Figure 5 describes an algorithm allowing a node to access the medium with correct backoff values; Figure 6 describes an algorithm allowing a node to recover the backoff values of all others nodes of the collaborative group; Figure 7 describes an example of a backoff recovery procedure; Figure 8 is a block diagram illustrating a schematic configuration of a communication node apparatus adapted to incorporate the invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS

Figure 4 shows an example of a communication network system where the present invention can be implemented. The communication network is assumed to implement a Carrier Sense Multiple Access with Collision Avoidance (CSMNCA) channel access mechanism, such as for example a wireless network implementing the 802.11 standard. Several nodes exchange data packets through a transmission channel (400) that may alter the data packets by dropping and/or corrupting data. This may be due to channel fading or interference/collision phenomenon.

The nodes can be divided into two groups. A first group of Nbi nodes (401 to 404) implementing a channel access mechanism but not the features of embodiments of the invention and a second group of Ncj nodes (405 to 408) implementing the channel access mechanism and the features of embodiments of the invention. The nodes of the first group are called legacy nodes, while the nodes of the second group are called collaborative nodes or peer nodes in the following description. The first group is called a legacy group, while the second group is called a collaborative group. Still, both legacy nodes Nbi and collaborative nodes Ncj are compliant with the channel access mechanism such as for example the standardized 802.lln MAC/PHY layers based on a random backoff mechanism (see figure 1) and thus data can be exchanged between these two types of nodes.

Figure 5 describes a basic algorithm allowing a node to access the medium according to an embodiment of the invention. This algorithm is implemented by each collaborative node.

The algorithm starts at 500. In step 501 a list of expected backoff values of the collaborative nodes is constructed.

The construction is preferably carried out by predicting the backoff values. An example of predicting the backoff value in each node may be carried out in the following way. A specific value may be associated with each collaborative node, the specific value of a node being known by the other nodes of the collaborative group of nodes. The determining of a backoff value for each node of the set of nodes is carried out in each collaborative node, the backoff value associated with a node being determined as a function of the specific value associated with that node. The function applied on the specific value associated to each collaborative node may be a pseudo-random generator taking the associated specific value as a seed. Advantageously, the specific values associated with the collaborative nodes are node identifiers. Thus, each node of the collaborative group is able to know when the backoff value of other nodes in the collaborative group goes to 0. Therefore, by not re-using a backoff value set in any of the other nodes, collisions among the collaborative nodes are prevented.

Thanks to this prediction mechanism, no backoff values are transmitted on the medium while still preventing different nodes of the collaborative group from using same backoff values.

Alternatively, the backoff list may be constructed by exchanging information over the network after each transmission or regularly after a predetermined interval corresponding to several data packets transmissions.

The list can be obtained by advertising next backoff value as disclosed in US20090141738.

Upon reception of a message from a sending node at step 502, a test 503 is performed. The test consists in checking if the reception time of the received message corresponds to the expected reception time determined from the expected backoff value contained in the list. If the response is positive, no desynchronization is detected and access to the medium is authorized at step 505. If the response is negative, then the node might be desynchronized. A backoff recovery procedure is implemented at step 504 to correct (update) the values of the backoff list. Once the list has been corrected, access to the medium is authorized at step 505. One detailed implementation of this general algorithm is described in relation with Figure 6.

Figure 6 describes an algorithm which first checks if peer nodes are desynchronized. A node which is identified to be desynchronized with respect to peer nodes will not be able to properly predict the expected backoff values of the other peer nodes as explained above. Therefore, such a node potentially risks causing collision with peer nodes when accessing the medium. When a particular first node has been detected to be desynchronized, the algorithm further allows this first node to recover the backoff values of all other nodes of the collaborative group. Once the first node has recovered those backoff values it is again able to collaborate with its peer nodes with a reduced risk of collision.

At step 610 a test is performed to determine if a bad packet is received, i.e. a packet has been transmitted but the node has not correctly decoded the received packet. A node receiving a bad packet sets its state to unsecure in step 620. Since its medium access scheme adds an EIFS time before restarting the backoff count down, the receiving node is potentially desynchronized with respect to other nodes of the collaborative group. To check whether the receiving node is indeed desynchronized, the node waits the reception of a RTS message sent by one of the peer node of the collaborative group (step 630).

If its own backoff value drops to 0 before receiving any RTS message from another node of the collaborative group (test at step 640), the node sends a 802.11 standard RTS message 641 to reserve the medium for the collaborative group (cf. Figure 7 for more details). Upon the reception of the associated CTS message from one of the node of the collaborative group (test 642 positive), a transmission slot is reserved for the group. In this case the badly received data packet by the node was also not correctly received by the other peer nodes and each node of the collaborative group added an EIFS time.

Therefore no desynchronization occurred between the peer nodes and data transmission can take place at step 643. If no CTS message is received (test 642 negative), the algorithm returns to initial step 610, the node remaining in an unsecure state.

If a RTS message from another node of the group is received (test 630 positive), the node compares in step 650 its predicted RTS reception time with the reception time of this RTS message. The predicted RTS reception time is defined as following: Predicted RTS reception = remaining predicted backoff value * aSlotTime duration + last starting time of the backoff countdown If the predicted RIS reception time is equal to the RIS reception time (test 650 positive), the node is considered to be synchronized in term of backoff values and the associated and reserved transmission slot can be performed normally without any modifications (goes to 640). If not (test 650 negative), the node confirms its desynchronization in term of backoff values.

Consequently, the node is considered to be desynchronized with respect to other nodes of the group and needs to recover the backoff values related to the other nodes of the collaborative group. Consequently when its own backoff value goes to 0 (test 652 positive), it sends a dedicated RIS message called recovery RTS message 653. The recovery RTS message is a 802.11 standard RTS message with slight modifications allowing other peer nodes to be notified that the node is desynchronized. For example, to identify a recovery RTS message, a dedicated flag known by all nodes of the collaborative group can be set. Another solution may be to set in the duration field an odd length for a recovery RTS message and an even length for a standard RTS message.

Alternatively, the RTS may embed a payload containing a management data representative of the synchronization error.

Upon reception of the associated CTS message (test 654 positive), the request for recovering the backoff values of the peer nodes is accepted. A transmission slot is reserved for the backoff recovery procedure 655. This transmission slot is divided into a number of sub-slots equal to the number of nodes belonging to the collaborative group whereby each node determines a sub-slot in an occupancy time interval corresponding to the reserved time slot.

The transmission slot is thus shared between the collaborative nodes such that each collaborative node obtains a dedicated sub-slot in the reserved transmission slot. Thanks to the sending of the recovery RTS message, all the peer nodes that are not under the unsecure state insert first in their transmission sub-slot their actual backoff lists allowing all nodes of the group to recover the backoff values (step 655). After having received backoff lists from peer nodes, the node realigns his backoff values contained in its backoff list (step 670) and sets its state to secure. If time remains left in the reserved transmission slot, the peer nodes can transmit additional data packets coming from the application layer. The nodes that are under the unsecure state do not transmit their backoff lists since they are erroneous, but only data packets.

After a negative check at step 650, if the node receives a recovery message from another node (test 651 positive) it cancels the sending of its own recovery message (step 660). In this particular case the node will correct its list of backoff values by receiving a valid list from a synchronized node. The transmission of this valid list is initiated by the said another node which has sent a recovery message during a recovery procedure.

The main advantage of this recovery process is that only one recovery RTS message needs to be sent to resynchronize all nodes that are in unsecure state. The recovery message is sent only after a node has been checked to be desynchronized therefore reducing to the minimum the number of transmissions for obtaining the backoff values and thus gaining bandwidth capacity for data transmission in comparison to the known techniques requiring heavy message exchanges.

It has to be noticed also that in order to be able to insert backoff values in the transmission sub-slots, it is important that each node of the collaborative group prepares in advance the data packet containing its current backoff values. In the 802.lle standard, the transmission order is managed through 4 different FIFOs. So the data packet containing the backoff values of the nodes of the collaborative group must preferably be inserted in the first place of the FIFO with the highest priority usually called the "voice" FIFO queue.

Figure 7 describes a timeline example of the backoff recovery procedure. The Nc2 and Nc4 nodes are in unsecure state due to a bad packet reception and have confirmed that their predicted backoff values are not aligned compared to the predicted backoff values of the other nodes of the collaborative group (at step 650 of figure 6). As illustrated in figure 7, when the own backoff of the Nc2 node goes to 0, the Nc2 node sends a recovery RIS message 710 (step 653 of figure 6) to launch a backoff recovery procedure. Upon reception of the associated CTS message 720, the granted transmission slot 740 is divided into sub-slots 741, 742, 743, 744 allowing each node of the collaborative group to exchange data packets. To do so, each node of the group has the ability to determine a sub-slot and to transmit during the determined sub-slot inside of the 802.11 timeslot originally granted to one node of the group. As a recovery RTS message was transmitted, each node that is in secure state, here nodes Nd and Nc3 (whose predicted backoff values have been checked synchronized) inserts its own actual backoff value and its predicted backoff values 731, 732 (corresponding to its view of peer backoffs) in a packet and sends them in its associated respective sub slots 741 and 743. So the transmitted packets contain the list of backoff values of nodes not being desynchronized. It has to be noticed that the nodes that are in unsecure state, here Nc2 and Nc4, sends only data packets coming from the application layer in order to avoid any misunderstanding among the collaborative nodes. The data packets are sent in respective sub slots 742 and 744. A time interval xIFS 751, smaller than the conventional SIFS, is inserted between each sub-slot 741, 742, 743 and 744 such that the transmission slot 740 and thus the recovery procedure is not interrupted by a node attempting to access the medium.

The recovery backoff procedure allows resynchronizing the predicted backoff values of all nodes that are in unsecure state even if they have not sent a recovery RTS message (like Nc4 in the present example) to warn the nodes of the collaborative group that they are in unsecure state and to correct the values of their backoff list.

Figure 8 shows a block diagram illustrating a schematic configuration of a communication apparatus (800) representing a transmitting node or a receiving node adapted to incorporate an embodiment of the invention. The PHY layer block (803) is in charge of formatting data packets and sending data packets on the wireless medium. The MAC layer block (802) is composed of a standard MAC 802.11 layer block (804) and four additional blocks implementing the invention: A Node Data transmission block (805) in charge of exchanging data packet with the physical layer and the application layer, A RTS/CTS module (806) in charge of managing Request-To-Send (RTS) and Clear-To-Send (CTS) messages according to an embodiment of the invention, A Backoff recovery module (807) in charge of detecting that a node is desynchronized and launching a backoff recovery procedure.

A Backoff management module (808) implementing the backoff management. This module is in charge of managing the backoff list maintained by each collaborative node.

The application layer block (801) implementing an application generating and receiving data packets, for example video data.

Any modification or improvement of the above-described embodiments that a person skilled in the art may easily conceive should be considered as falling within the scope of the invention.

Claims

CLAIMS1. A method for accessing a communication medium in a network which can be accessed by a plurality of nodes, each node using a medium access mechanism with collision avoidance based on a computation of backoff values corresponding to a number of time-slots a node waits before accessing the medium, wherein each node of a group of nodes among the plurality of nodes holds a list of expected backoff values of the nodes of the group, the method comprising for a first node of the group having detected a badly received packet: waiting for the reception of a message from a sending node of the group, and, upon reception of the message, checking if the reception time of the message corresponds to an expected reception time determined from the expected backoff value of the sending node contained in the list of expected backoff values hold by the first node, and, if the reception time does not correspond to the expected reception time, implementing a backoff recovery procedure to correct the values of the backoff list of the first node which is considered to be a desynchronized node, and, accessing the communication medium using the corrected backoff list.
2. The method of claim 1, wherein the backoff recovery procedure comprises sending a recovery message requesting the transmission of the backoff list from other nodes of the group and after reception of the backoff list from at least one other node, correcting the backoff list of the first node.
3. The method of claim 2 further comprising for a second node of the group considered to be desynchronized and having received the backoff list of at least one other node in said transmission, correcting the backoff list of the second node.
4. The method of claim 3, wherein the second node cancels the sending of its recovery message.
5. The method of any one of claims 2 to 4, wherein a transmission slot is granted in response to the recovery message and at least one node of the group not considered to be desynchronized transmits its backoff list during the granted transmission slot.
6. The method of claim 5, wherein the transmission slot is divided into sub-slots, each sub-slot being assigned to a node of the group not considered to be desynchronized in order to transmit its backoff list.
7. The method of claim 2, wherein a packet including the backoff list is prepared in advance of the transmission in order to be transmitted during a transmission slot granted to the group.
8. The method of claim 1, further comprising for any node considered to be desynchronized: obtaining the backoff list from at least one node of the group through the recovery procedure, and correcting the backoff list of the node considered to be desynchronized with the obtained backoff list.
9. The method ot claim 1 or 8, wherein the backoff list is obtained only from nodes not considered to be desynchronized.
10. The method of claim 1, wherein if the reception time corresponds to the expected reception time, no correction of the backoff list values of the first node is performed.
11. The method of claim 1, wherein the values contained in the backoff list are predictive values distributed among the nodes of the group.
12. The method of claim 11, wherein each node of the group generates internally the predictive values of its backoff list.
13. A node for data communication with other nodes of a group of nodes of a radio network, the node comprising: a backoff management module for implementing a medium access mechanism with collision avoidance based on a computation of backoff values corresponding to a number of time-slots a node needs to wait before accessing the medium, wherein the nodes holds a list of expected backoff values of the nodes of the group, a backoff recovery module for detecting a badly received packet, and, upon detection of a badly received packet, waiting for the reception of a message from a sending node of group, wherein: the backoff recovery module comprises means for checking if the reception time of the message corresponds to an expected reception time determined from the expected backoff value of the sending node contained in the list, and, if the reception time does not correspond to the expected backoff value, implementing a backoff recovery procedure to correct the values of its backoff list.
14. The node of claim 13 further comprising a data transmission block for sending a recovery message requesting the backoff list from other nodes of the group.
15. The node of claim 13 or 14 comprising a RTS/CTS module for granting a transmission slot in response to a recovery message and for transmitting its backoff list during the granted transmission slot.
16. A non-transitory computer-readable medium storing a program which, when executed by a microprocessor or computer system in an apparatus communicating data with nodes of a group of nodes implementing a medium access mechanism with collision avoidance based on a computation of a backoff value, wherein each node of the group holds a list of expected backoff values of the nodes of the group, causes the apparatus to perform for a first node of the group having detected a badly received packet the steps of: waiting for the reception of a message from a sending node of the group, upon reception of the message, checking if the reception time of the message corresponds to an expected reception time determined from the expected backoff value of the sending node contained in the list of the first node, and, if the reception time does not correspond to the expected reception time, considering the node as being desynchronized and implementing a backoff recovery procedure to correct the values of the backoff list of the first node, and, accessing the communication medium using the corrected backoff list.
17. A method for accessing a communication medium in a network substantially as hereinbefore described with reference to, and as shown in, Figures 5, 6 and 7 of the accompanying drawings.
18. A device for a node for accessing a communication medium in a network substantially as hereinbefore described with reference to, and as shown in, Figure 8 of the accompanying drawings.