EDCA Backoff Restart Procedures and State Switching in EMLSR or EMLMR Co-Affiliated Stations

Abstract

The non-AP MLD of the active EMLSR wants to initiate a frame exchange with the AP MLD over one of the EMLSR links. A first link is selected and the corresponding "first" Affiliated STA initiates an EDCA backoff procedure that decrements one or more backoff counters to access the first link. Alternatively, each EMLSR's co-affiliated STA may initiate an EDCA backoff procedure to access its respective link. In an embodiment, the first Affiliated STA is switched from a listening operation state to a valid frame exchange state before initiating the EDCA backoff procedure. In another embodiment, the first Affiliated STA initiates the EDCA backoff procedure first and is switched to a valid frame exchange state upon expiration of the backoff counter. The corresponding state switching procedure may be initiated to terminate with or in response to expiration of the backoff counter.

Description

[Technical Field]

[0001] The present invention relates generally to wireless communications, and more particularly to Multi-Link (ML) communications. [Background technology]

[0002] Wireless communication networks are widely deployed to provide various communication services such as voice, video, packet data, messaging, broadcast, etc. These wireless networks may be multiple-access networks capable of supporting multiple users by sharing available network resources. Examples of such multiple-access networks include code division multiple access (CDMA) networks, time division multiple access (TDMA) networks, frequency division multiple access (FDMA) networks, orthogonal FDMA (OFDMA) networks, and single-carrier FDMA (SC-FDMA) networks.

[0003] The 802.11 family of standards adopted by the Institute of Electrical and Electronics Engineers (IEEE®) provides a number of mechanisms for wireless communication between STAs.

[0004] With the development of latency-sensitive applications such as online gaming, real-time video streaming, virtual reality, and remote control of drones and robots, the requirements and issues of better throughput, lower latency, and robustness must be taken into account. These issues are currently being considered by the IEEE 802.11 Working Group as the primary objectives for issuing the next major 802.11 release, known as 802.11be or EHT (Extremely High Throughput).

[0005] The IEEE P802.11be / D1.5 version (March 2022, hereinafter referred to as the "D1.5 standard") introduces Multi-Link (ML) Operation (MLO), which improves data throughput by enabling communication between STAs over multiple parallel and discontinuous communication links.

[0006] MLO allows a non-AP (Access Point) MLD (ML Device) to register with an AP MLD, i.e., discover, authenticate, associate, and set up multiple links with the AP MLD. Each link allows channel access and frame exchange between the non-AP MLD and the AP MLD based on the supported capabilities exchanged during the association procedure.

[0007] An MLD is a logical entity with multiple stations (STAs) and a single Medium Access Control (MAC) Service Access Point (SAP) for a Logical Link Control (LLC), containing one MAC data service. Thus, an AP MLD consists of multiple affiliated APs, and a non-AP MLD consists of multiple affiliated non-AP STAs. Affiliated STAs in both AP and non-AP MLDs can use 802.11 mechanisms to communicate with affiliated STAs in another MLD via each of the multiple communication links set up.

[0008] Along with the introduction of spatial multiplexing capabilities of MLO and MLD, the D1.5 standard introduced new Operating Modes (OM) called Enhanced Multi-Link Operating Mode (EML OM), namely EMLSR (Enhanced Multi-Link Single Radio) mode and EMLMR (Enhanced Multi-Link Multi-Radio) mode.

[0009] A non-AP MLD declares its support for the EML mode of operation (known as EML Capabilities) to the AP MLD during the association phase. In this mode of operation, activation and deactivation of the EML mode of operation is initiated by the non-AP MLD sending a specific EHT action frame called "EML OM Notification". The D1.5 standard states that the two modes, EMLSR and EMLMR, are mutually exclusive.

[0010] When EMLMR mode is enabled, the non-AP MLD simultaneously listens to a set of active links (so-called EMLMR links, typically consisting of two active links) to receive the Initial frame sent by the AP MLD and initiate frame exchange, and then aggregates some physical resources of different radios used by the different links (so-called EMLMR links) to transmit and receive data of up to a predefined number of supported receive and transmit spatial streams via only one EMLMR link at a time (typically the link on which the Initial frame was received), which may be more than the number of receive and transmit spatial streams supported by each radio.

[0011] When EMLSR mode is activated, the non-AP MLD simultaneously listens to a set of valid links (the so-called EMLSR links, usually consisting of two valid links) to receive Initial Control frames (e.g., MU-RTS trigger frame, BSRP trigger frame) from the AP MLD to initiate frame exchange, and then can exchange data frames with the AP MLD via only one EMLSR link at a time (usually the link from which the Initial Control frame was received).

[0012] Additionally, a non-AP MLD also has the ability to initiate frame exchange with an AP MLD via one EMLSR or EMLMR link to transmit uplink data. In such a case, STAs affiliated with a non-AP MLD operating in EMLSR or EMLMR mode do not need to transmit an Initial Control frame or an Initial frame to initiate frame exchange with the AP MLD (non-triggered UL transmission), and access the wireless medium according to the rules defined in Section 10.3.2.4 (NAV Configuration and Reconfiguration) and Section 10.23.2 (HCF Contention-Based Channel Access (EDCA)) as specified in the IEEE 802.11-2020 standard.

[0013] However, conventional contention-based channel access procedures such as EDCA are not fully adapted to the EML mode, where the EMLSR or EMLMR links are not completely independent of each other. Therefore, it is necessary to improve the EDCA procedure for the EML mode. Summary of the Invention

[0014] It is a broad object of the present invention to provide an enhanced contention-based channel access procedure adapted to an active EML mode (EMLSR or EMLMR) that takes into account the EML Capabilities associated with the active EML mode.

[0015] The present invention provides a method for communicating in a wireless network, in a non-access point (non-AP) multilink device (MLD) operating in an active enhanced multilink (EML) mode, comprising: Initiating an Enhanced Distributed Channel Access (EDCA) backoff procedure by a first station (STA) corresponding to a first link among a set of valid links affiliated with a non-AP MLD and to which an EML mode is applied, to access the first link, the STA decrementing a backoff counter; The method may include switching the first affiliated STA from a listening operation state to an active frame exchange state to initiate a frame exchange with the AP MLD over the first link.

[0016] In an embodiment, switching the first affiliated STA to the valid frame exchange state is performed before initiating an EDCA backoff procedure by the first affiliated STA in the valid frame exchange state to decrement a backoff counter.

[0017] For example, a method of communication in a wireless network may involve a non-access point (non-AP) multilink device (MLD) operating in an active enhanced multilink (EML) mode: Initiating a frame exchange with the AP MLD via a first link by a first STA corresponding to the first link of the set of valid links to which an EML mode affiliated with the non-AP MLD applies; Initiating a frame exchange is The method includes switching the first affiliated STA from a listening operation state to an active frame exchange state before initiating an enhanced distributed channel access (EDCA) backoff procedure that decrements a backoff counter to access the first link by the first affiliated STA in the active frame exchange state.

[0018] In these embodiments, the EDCA backoff procedure is not affected by the STA's state switching, so this first affiliated STA is ready to perform a frame exchange as soon as its backoff counter expires.

[0019] In another embodiment, switching the first affiliated STA to the valid frame exchange state is performed when the backoff counter reaches a value of zero.

[0020] For example, a method of communication in a wireless network may involve a non-access point (non-AP) multilink device (MLD) operating in an active enhanced multilink (EML) mode: Initiating a frame exchange with the AP MLD via a first link by a first STA corresponding to a first link of a set of valid links affiliated with the non-AP MLD and to which the EML mode is applied; Initiating a frame exchange is Initiating an Enhanced Distributed Channel Access (EDCA) backoff procedure by a first affiliated STA in a listening operation state to access the first link, the EDCA backoff procedure decrementing a backoff counter; and switching the first affiliated STA from a listening operation state to an active frame exchange enabled state when the backoff counter reaches a value of zero.

[0021] In these embodiments, while the backoff counter is being decremented, other co-affiliated STAs in the non-AP MLD are still in a listening state, which means that the AP MLD can still initiate frame exchanges with these other co-affiliated STAs despite the counter decrement by the first affiliated STA, thereby improving network communications.

[0022] In another embodiment for efficiently driving EDCA on EMLSR or EMLMR links to mirror their dependencies, in response to initiating a frame exchange with the AP MLD on a second link of a set of valid links to which EML mode applies, the backoff counter driving EDCA on a first link of that set is paused.

[0023] For example, a method of communication in a wireless network may involve a non-access point (non-AP) multilink device (MLD) operating in an active enhanced multilink (EML) mode: In response to initiating a frame exchange with the AP MLD over a second link of a set of valid links to which the EML mode applies, pausing a back-off counter that drives enhanced distributed channel access (EDCA) on a first link of the set.

[0024] Therefore, a specific event occurring on the second link has a direct impact on the first link and other links of the same EMLSR or EMLMR. Unlike known techniques, pausing the back-off counter for accessing a medium (link) no longer depends solely on the idle / busy state of the target medium, but on activity (frame exchange) on the link of the other EMLSR / EMLMR. This avoids obtaining EDCA access to the first link while the corresponding co-affiliated STA is unavailable (because the frame exchange has allocated radio resources to the co-affiliated STA). It can be seen that this improves the efficiency of the EDCA procedure.

[0025] Although only one first link of the set is mentioned here, the invention can be applied to other links (different from the second link) of the EMLSR / EMLMR link set.

[0026] Optional features of the invention are defined below with reference to the method, but these can be substituted with apparatus features.

[0027] In some embodiments in which the decrement is performed while the first affiliated STA is in a valid frame exchange state, the method further includes, upon detecting that the first link becomes busy during the decrement of the backoff counter, suspending the decrement and applying one of a plurality of policies.

[0028] As an example, policy a) includes keeping the first affiliated STA in a valid frame exchange state and resuming decrementing when the first link becomes idle again.

[0029] As another example, policy b) includes returning the first affiliated STA to a listening operational state regardless of the time determined based on the frame in which the detection occurred.

[0030] As yet another example, policy c) includes switching the first affiliated STA to a listening operation state for a predetermined period based on a period determined based on the frame in which the detection occurred before switching back to an active frame exchange state to resume decrementing if the first link becomes idle again.

[0031] As yet another example, policy d) includes determining the identified period based on the frame in which the detection occurred, and determining to apply one or the other of policies a), b), or c) depending on the determined period.

[0032] In some embodiments in which the decrement is performed while the first affiliated STA is in a listening state, the method further includes suspending the EDCA backoff procedure (i.e., decrementing the backoff counter is stopped) upon receiving an Initial frame from the AP MLD via a second link of the set. An Initial frame is understood to be a frame that initiates a frame exchange initiated by the AP MLD. The Initial frame is known by this name in EMLMR mode and by the name "Initial Control frame" in EMLSR mode. Thus, a non-AP MLD may be involved in a new frame exchange initiated by the AP on the second link. This is possible because other co-affiliated STAs corresponding to the second link are still in a listening state while the backoff counter is being decremented.

[0033] More generally, the method may further include, upon receiving an Initial frame from the AP MLD via a second link of the set, determining whether to suspend the EDCA backoff procedure based on one of a plurality of criteria.

[0034] As a first example, the criteria may include determining whether the first affiliated STA has been allocated full radio resources.

[0035] As another example, the criteria may determine whether the Initial frame is an MU-RTS trigger frame.

[0036] As yet another example, the criteria may include determining whether uplink data is already preloaded in a transmission associated with only the first link compared to the second link.

[0037] As yet another example, the criteria may include determining whether the amount of buffered data is greater than a threshold.

[0038] These embodiments can be envisioned independently of the core definition of the invention above. As an independent concept, these embodiments relate to a method of communication in a wireless network, in a non-access point (non-AP) multilink device (MLD) operating in an active enhanced multilink (EML) mode: Initiating an Enhanced Distributed Channel Access (EDCA) backoff procedure by a first STA that is in a listening operation state and that corresponds to a first link of a set of valid links to which an EML mode is applied, the STA decrementing a backoff counter to access the first link; The decrementing of the backoff counter is suspended upon receipt of an Initial frame from the AP MLD over the second link in the set, or more generally, receipt of an Initial frame from the AP MLD over the second link in the set determines whether to pause the decrementing of the backoff counter.

[0039] In some embodiments of the present invention, switching the first affiliated STA includes initiating a state switch procedure for the first affiliated STA while the first affiliated STA is decrementing a back-off counter, such that the state switch procedure terminates when the back-off counter reaches a value of 0, thereby improving communication performance.

[0040] In another embodiment, switching the first co-affiliated STA includes initiating a state switching procedure for the first co-affiliated STA in response to the backoff counter reaching a value of zero. In this case, the first co-affiliated STA has not changed state when activity is sensed when the backoff counter is about to expire. This approach conserves opportunities for other co-affiliated STAs to access the medium, thereby improving communication performance.

[0041] In a particular embodiment, the method further includes transmitting a control frame over the first link in response to the backoff counter reaching a value of zero. The control frame may be a CTS-to-self frame or an RTS frame. This is to protect the available medium in case another MLD desires to access the same medium while the non-AP MLD is switching the state of its co-affiliated STA. This also contributes to improving communication performance of the network.

[0042] According to a particular feature, the control frame includes padding to terminate the control frame after a short interframe space (SIFS) prior to the end of the state switching procedure, thereby ensuring protection of the medium during all transitional periods during which co-affiliated STAs are switched.

[0043] According to another particular feature, a first affiliated STA is allocated light radio resources in a listening operation state, while a separate second STA corresponding to a second link of the set, affiliated with a non-AP MLD, is allocated full radio resources in a listening operation state. In fact, the above protection is worth implementing when, in EMLSR mode, the "active" EMLSR co-affiliated STA (here, the first STA) is initialized with a light radio stack and requires a longer time (during a switch) to be configured with a full radio stack.

[0044] In some embodiments, the method includes setting a network allocation vector (NAV) of a first affiliated STA in a listening state upon sensing any control frame over the first link having an MCS value up to 2. This extends processing of frames by the affiliated STA in a listening state beyond just Initial Control frames (MU-RTS and BSRP trigger frames).

[0045] In some embodiments, the method includes simultaneously initiating an EDCA backoff procedure on two or more links of a set, where a first link is a link corresponding to a backoff counter of the EDCA backoff procedure that first reaches a value of 0. In other words, a separate second affiliated STA in a listening operation state corresponding to a second link of the set initiates an EDCA backoff procedure simultaneously with the initiation of the EDCA backoff procedure by the first affiliated STA. This increases opportunities for non-AP MLDs to access the wireless network, improving communication performance.

[0046] In some embodiments, the method further includes switching a separate second STA corresponding to a second link of the set, affiliated with the non-AP MLD, from a listening operation state to an invalid frame exchange state concurrently with the switching of the first affiliated STA.

[0047] In some embodiments, the method further includes returning the first and second affiliated STAs to a listening operational state upon completion of the frame exchange over the first link.

[0048] In some embodiments, initiating the frame exchange excludes transmitting an AP MLD Initial frame on the first link, which demonstrates that the present invention is directed to non-triggered uplink transmission of non-AP MLD in EML mode.

[0049] In some embodiments, the method further includes performing a frame exchange (typically an uplink transmission at the initiative of the non-AP MLD) with the AP MLD over the first link after the backoff counter reaches a value of zero.

[0050] In particular, with regard to the issue of fairness in accessing the wireless network, a particular problem with EDCA arises after an uplink (UL) transmission of the link of the second EMLSR or EMLMR, since the systematic use of the legacy EDCA backoff resume or restart procedure for the link of the first EMLSR or EMLMR in the link set of the EMLSR or EMLMR is not adapted.

[0051] Therefore, it is also an object of embodiments of the present invention to provide an EDCA backoff resume or restart procedure adapted to EML mode that takes EML Capabilities into account. In particular, some embodiments of the present invention specify a new EDCA backoff resume or restart procedure for STAs affiliated with a non-AP MLD operating on a first EMLSR or EMLMR link after an UL transmission occurs on a second EMLSR or EMLMR link of the same EMLSR or EMLMR linkset.

[0052] In this context, the method may further include, again in non-AP MLD, restarting the backoff counter in response to completion of the frame exchange over the second link, applying a backoff restart strategy selected based on characteristics of the frame exchange.

[0053] "Restarting" the backoff counter should be understood as any technique for restarting the decrementation of the backoff counter, regardless of whether the decrement resumes from the last known / current value of the counter or restarts from a new counter value that reinitializes the backoff counter. These various options are described in more detail below.

[0054] Therefore, the strategy or policy for restarting the backoff counter of a link depends on what happens (frame exchanges) on another link, contrary to the traditional approach for EDCA procedures. This shows that applying or adjusting appropriate EDCA backoff restart procedures can mitigate the issue of network access fairness. In particular, the backoff counter associated with an AC on a link may be penalized if data from the same AC is transmitted over another link in the same EMLSR or EMLMR link set.

[0055] In some embodiments, the characteristics of the frame exchange include whether the frame exchange includes single-user uplink transmissions to the AP MLD, multi-user triggered uplink transmissions to the AP MLD, or only downlink transmissions from the AP MLD, which allows the network to adjust the penalty depending on whether non-AP MLDs get additional transmission opportunities (e.g., via triggered UL) compared to traditional EDCA (single-user UL) or downlink transmissions.

[0056] In some embodiments, the characteristics of the frame exchange include whether the frame exchange resulted in a successful or unsuccessful uplink transmission to the AP MLD, allowing the network to adjust the penalty depending on whether the non-AP MLD successfully utilized its transmission opportunity.

[0057] Of course, the above embodiments (SU or MU transmission as a characteristic and success or failure of transmission as another characteristic) may be advantageously combined, meaning that the SU UL or MU UL or DL ​​nature of the transmission and the success or failure status of the transmission are also taken into account. Of course, other characteristics of the frame exchange may also be used to select a restart strategy or procedure.

[0058] In some embodiments, the restart strategy applied includes one of the following: reinitializing the backoff counter using the current contention window before starting to decrement the backoff counter; Before starting to decrement the backoff counter, reinitialize the backoff counter using the new contention window associated with the new EDCA mode. It should be understood that the new EDCA mode is considered "new" compared to the affiliated STA's current EDCA mode. For example, a STA operating on the second link may switch from a legacy EDCA mode to a multi-user EDCA (MU EDCA) mode with its own MU EDCA parameters, including its own contention window value. Alternatively, a new EML EDCA mode (e.g., EMLSR EDCA mode and / or EMLMR EDCA mode) may be defined with its own EML EDCA parameters, including its own contention window value, that penalizes the AC more or less, in which case a STA operating on the second link may switch from a legacy EDCA mode to an EML EDCA mode. The backoff counter may be restarted from its current value, and a switch to another EDCA mode may be made to penalize the backoff counter for the next reinitialization.

[0059] Use of the new EML EDCA mode differs substantially from known 802.11 technologies, where only EDCA and MU EDCA modes are known. The EML EDCA mode corresponds to a specific set of EML EDCA parameters that are distinct from the EDCA and MU EDCA parameters and are transmitted (via beacon or probe response frames) from affiliated APs to co-affiliated STAs separately from the EDCA and MU EDCA parameters. The EML EDCA parameters are specific to each link and can therefore vary from one link to another, or can be set to the same value across the entire set of EMLSR or EMLMR links. Similar but distinct fields may be provided in the beacon / probe response frames to convey these parameters over each link.

[0060] Correspondingly, a communication method in a wireless network, in a non-access point (non-AP) multi-link device (MLD) operating in an active enhanced multi-link (EML) mode, includes: The method may include switching enhanced distributed channel access (EDCA) parameters of a backoff counter driving EDCA to a second link of the set of valid links from the current EDCA parameters to different EML EDCA parameters in response to completion of a frame exchange with the AP MLD over a first link of the set of valid links to which the EML mode applies. This defines an EML EDCA mode for the co-affiliated STA operating on the second link. An EML EDCA mode may be engaged on each link different from the first link of the set, if there are multiple links. The per-link EML EDCA modes may use link-specific EML EDCA parameters or may alternatively share the same EML EDCA parameters.

[0061] In this way, the penalty of EDCA for non-AP MLD can be adjusted for active EML mode.

[0062] In an embodiment, the frame exchange includes a (preferably successful) trigger-based uplink transmission to the AP MLD.

[0063] In some embodiments, the method further includes, in response to completion of the frame exchange with the AP MLD over the first link, switching EDCA parameters of a backoff counter driving EDCA for the first link from the current EDCA parameters to different EML EDCA parameters. In other words, the first link over which the frame exchange occurs is penalized in the same way as the other EMLSR / EMLMR links in the set. The relevant backoff counters are those corresponding to the ACs transmitted during the frame exchange.

[0064] In another embodiment, the non-AP MLD stores a set of EML single-radio (EMLSR) EDCA parameters and a set of EML multi-radio (EMLMR) EDCA parameters, and different EML EDCA parameters are selected from the set of EMLSR EDCA parameters and the set of EMLMR EDCA parameters depending on whether the non-AP MLD is in EMLSR mode or EMLMR mode, respectively.

[0065] In another embodiment, the method further includes receiving, from the AP MLD, a management frame (beacon or probe response frame) that includes different EML EDCA parameters, possibly in addition to the EDCA parameters and the MU EDCA parameters. Correspondingly, a communication method in a wireless network includes, in an Access Point (AP) Multilink Device (MLD) operable in an active Enhanced Multilink (EML) mode: transmitting a management frame to the non-AP MLD, the management frame including a set of EML EDCA parameters; The EML EDCA parameter is used to configure the Enhanced Distributed Channel Access (EDCA) of a non-AP MLD operating in EML mode to drive access to the first link of a set of links to which EML mode applies when frame exchange with the AP MLD via the second link of the set is completed.

[0066] Therefore, a new set of EDCA parameters is defined to penalize co-affiliated STAs when they obtain additional transmission opportunities (typically through trigger-based UL transmissions) while operating in EML mode.

[0067] The set of EML (or EMLSR and / or EMLMR) EDCA parameters may be added to the set of EDCA parameters in the management frame and / or may be added to the set of MU EDCA parameters in the management frame.

[0068] In certain embodiments, the restart strategy applied comprises one of the following: restarting the back-off counter from its current value in the case of a downlink transmission or a failed uplink transmission in a frame exchange; If a single user's uplink transmission in a frame exchange is successful, reinitializing the backoff counter using the current contention window before starting to decrement the backoff counter or restarting the backoff counter from its current value; In the event of a successful multi-user trigger-based uplink transmission in a frame exchange, reinitializing the backoff counter using the current contention window or a new contention window associated with the new EDCA mode before starting to decrement the backoff counter or restarting the backoff counter from its current value (possibly switching the STA to another EDCA mode - for example, MU EDCA mode or EML EDCA mode as defined above - to penalize the backoff counter for the next reinitialization).

[0069] In some embodiments, in response to initiating a frame exchange, a plurality of backoff counters driving EDCA on the first link for each of a plurality of access categories are paused; The restart strategy is applied to backoff counters of a plurality of suspended backoff counters corresponding to access categories exchanged during the frame exchange.

[0070] Therefore, all ACs that benefit from frame exchange may be penalized by the applied restart strategy.

[0071] In some embodiments, the restart strategy includes a synchronization delay from the end of the frame exchange before restarting the backoff counter, which allows co-affiliated STAs on inactive links during the frame exchange to resynchronize with the medium.

[0072] In certain embodiments, including a synchronization delay in the restart strategy is contingent on the duration of the frame exchange being greater than a predetermined threshold. Such delay and threshold may be MediumSyncDelay and aMediumSyncThreshold, respectively, as defined in the D1.5 standard. This configuration prevents co-affiliated STAs from wasting time before again competing for access to the medium if the STAs have not lost synchronization from the medium.

[0073] According to a particular feature, the timer that counts down the synchronization delay is initialized with the MediumSyncDelay value provided by the AP MLD on the first link (e.g., in the Basic Multi-Link element of the most recently received frame from the associated AP). In particular, the timer may be set to 0 if a control frame with an MCS value up to 2 is successfully received over the first link. Again, this is to avoid wasting time before the STA contends for access to the medium again as soon as it has resynchronized with the medium (thanks to a successfully received frame).

[0074] In some embodiments, initiating the frame exchange includes detecting the expiration of another backoff counter driving EDCA to the second link. Preferably, the backoff counter and the another backoff counter are associated with the same access category. In this configuration, a co-affiliated STA of the second link gains access to the link for SU transmissions of a particular AC. Thus, the restart strategy is applied to the backoff counter of the other link corresponding to the same particular AC. This ensures proper fairness between ACs.

[0075] In another embodiment, initiating the frame exchange includes receiving an Initial frame (including a known Initial Control frame in EMLSR mode) from the AP MLD over the second link. In particular, the Initial frame can be received while another backoff counter driving EDCA for the second link is being decremented.

[0076] In some embodiments, the method further includes, in response to a termination of the frame exchange over the second link, applying a different restart strategy based on characteristics of the frame exchange and restarting a different backoff counter that drives the EDCA on the second link.

[0077] In an embodiment, other restart strategies include one of the following: if the other backoff counter expires, reinitializing the other backoff counter using the updated contention window before starting to decrement the other backoff counter; restarting the other backoff counter from its current value, and otherwise, possibly, switching the backoff engine to the other EDCA mode (e.g., MU EDCA mode or any EML EDCA mode defined above) to penalize the backoff counter for the next reinitialization if the trigger-based UL transmission is successful. In some embodiments, the other EDCA mode is the same as the new EDCA mode entered by the co-affiliated STA operating on the first link, meaning that both (or more) co-affiliated STAs switch to the same other EDCA mode (MU EDCA mode or EML EDCA mode), even though the sets of MU or EML EDCA parameters they each use may be different because they are specific to each link. Alternatively, the co-affiliated STAs operating on the second link either remain in conventional EDCA mode or simply switch to MU EDCA mode (having benefited from trigger-based UL transmission), while the co-affiliated STAs operating on the first link switch to EML MU mode.

[0078] In some embodiments, the method further includes, in response to initiating the frame exchange, switching a second STA corresponding to the second link affiliated with the non-AP MLD from a listening operation state to an enabled frame exchange state and switching a first STA corresponding to the first link affiliated with the non-AP MLD from a listening operation state to an disabled frame exchange state.

[0079] Relatedly, the present invention also provides a wireless communication device including at least one microprocessor configured to perform the steps of any of the above methods, wherein the wireless communication device is a non-AP MLD.

[0080] In particular, a non-access point (non-AP) multilink device (MLD) capable of operating in active enhanced multilink (EML) mode: means for initiating, by a first station (STA) associated with a first link of a set of valid links for which the EML mode is applicable, an Enhanced Distributed Channel Access (EDCA) backoff procedure that decrements a backoff counter to access the first link; and means for switching the first affiliated STA from a listening operation state to an active frame exchange state to initiate a frame exchange with the AP MLD over the first link.

[0081] Another aspect of the present invention relates to a non-transitory computer readable medium storing a program which, when executed by a microprocessor or computer system in a wireless device, causes the wireless device to perform any of the methods as defined above.

[0082] At least part of the methods according to the present invention may be computer-implemented. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining software and hardware aspects, all of which may be referred to generally herein as a "circuit," "module," or "system." Furthermore, the present invention may take the form of a computer program product embodied in any tangible medium having computer-usable program code embodied in the medium.

[0083] Since the present invention can be implemented in software, the present invention can be embodied as computer-readable code for provision to a programmable device on any suitable carrier medium. Tangible, non-transitory carrier media can include storage media such as floppy disks, CD-ROMs, hard disk drives, magnetic tape devices, or solid-state memory devices. Transitory carrier media can include signals such as electrical, electronic, optical, acoustic, magnetic, or electromagnetic signals, e.g., microwave or RF signals. [Brief explanation of the drawings]

[0084] Embodiments of the present invention will now be described, by way of example only, with reference to the following drawings: [Figure 1] 1 illustrates a typical 802.11 network environment including ML transmission between EML-enabled MLDs in which the present invention may be implemented; [Figure 1ab] 1 shows an exemplary 802.11be multilink reference model for MLD, either AP MLD or non-AP MLD; [Figure 2] Schematic representation of the frame sequence for the EMLSR operation mode specified in the D1.5 standard; [Figure 3] 1 illustrates, with a flowchart, the steps performed to operate a first contention-based channel access procedure with EMLSR active non-AP MLD according to a first embodiment; [Figure 4] 4 schematically illustrates an exemplary timeline of a first contention-based channel access procedure as described in FIG. 3; [Figure 5a] 1 illustrates, with a flowchart, the steps performed for an EMLSR-active non-AP MLD to operate a second contention-based channel access procedure according to a second embodiment; [Figure 5b] 1 illustrates, with a flowchart, the steps performed for an EMLSR-active non-AP MLD to operate a second contention-based channel access procedure according to a second embodiment; [Figure 6a]5a and 5b show schematic diagrams of alternative timelines for the second contention-based channel access procedure described in FIG. 5a and FIG. 5b; [Figure 6b] A more detailed scenario is shown schematically in Fig. 6a, considering the transition period during which the state of the EMLSR co-affiliated STAs switches; [Figure 7] 10 illustrates, with a flowchart, the steps performed by an EMLSR active non-AP MLD to operate a third contention-based channel access procedure according to a third embodiment; [Figure 8] 8 schematically illustrates an exemplary timeline of the third contention-based channel access procedure described in FIG. 7; [Figure 9] 1 illustrates, with a flowchart, steps for handling restarting a backoff counter in EML mode according to an embodiment; [Figure 10a] 10A-10C schematically illustrate an exemplary timeline of a first EMLSR or EMLMR operation case with a back-off counter restart procedure according to an embodiment; [Figure 10b] 10A-10C schematically illustrate an exemplary timeline of a first EMLSR or EMLMR operation case with a back-off counter restart procedure according to an embodiment; [Figure 11] 10A and 10B schematically illustrate an exemplary timeline of a second EMLSR or EMLMR operation case including a back-off counter restart procedure according to another embodiment, when the triggering event for frame exchange is the reception of an Initial frame; [Figure 12] 10A and 10B schematically illustrate an exemplary timeline of a third EMLSR or EMLMR operation case including a back-off counter restart procedure according to yet another embodiment, where the triggering event for frame exchange is the reception of an Initial frame; [Figure 13] 1 shows a table collecting proposed EDCA backoff counter restart procedures or policies according to an embodiment; [Figure 14] 1 illustrates schematically an EMLSR-enabled architecture of an MLD for implementing an embodiment of the present invention; [Figure 15] 1 schematically illustrates an EMLMR-enabled architecture of an MLD for implementing an embodiment of the present invention; and [Figure 16] 1 shows a schematic diagram of a wireless communication device according to an embodiment of the present invention; DETAILED DESCRIPTION OF THE INVENTION

[0085] The techniques described herein may be used in various broadband wireless communication systems, including communication systems based on orthogonal multiplexing. Examples of such communication systems include spatial division multiple access (SDMA) systems, time division multiple access (TDMA) systems, orthogonal frequency division multiple access (OFDMA) systems, and single-carrier frequency division multiple access (SC-FDMA) systems. SDMA systems may utilize sufficiently different directions to simultaneously transmit data belonging to multiple user terminals, such as wireless devices or STAs. TDMA systems may allow multiple user terminals to share the same frequency channel by dividing the transmission signal into different time slots or resource units and assigning each time slot to a different user terminal. OFDMA systems utilize orthogonal frequency division multiplexing (OFDM), which is a modulation technique that divides the overall system bandwidth into multiple orthogonal subcarriers or resource units. These subcarriers may be referred to as tones, bins, etc. In OFDM, each subcarrier may be independently modulated with data. An SC-FDMA system may utilize Interleaved FDMA (IFDMA), which transmits on subcarriers distributed across the system bandwidth, Localized FDMA (LFDMA), which transmits on blocks of adjacent subcarriers, or Enhanced FDMA (EFDMA), which transmits on multiple blocks of adjacent subcarriers.

[0086] The teachings herein may be incorporated into (e.g., implemented in or performed by) a variety of apparatuses (e.g., STAs). In some aspects, a wireless device or STA implemented in accordance with the teachings herein may or may not include an access point (referred to as an AP) (referred to as a non-AP STA or STA).

[0087] Although the embodiments are described in the context of a WiFi network, the invention can be used in any type of wireless network, such as, for example, a mobile telephone cellular network, which implements very similar mechanisms.

[0088] An AP may include, be implemented as, or be known as a NodeB, Radio Network Controller ("RNC"), Evolved Node B (eNB), 5G Next Generation Base STA ("gNB"), Base STA Controller ("BSC"), Base Transceiver STA ("BTS"), Transceiver Function ("TF"), Wireless Router, Wireless Transceiver, Basic Service Set ("BSS"), Enhanced Service Set ("ESS"), Radio Base STA ("RBS"), or other term.

[0089] A non-AP STA may include, be implemented as, or be known as a subscriber STA, subscriber unit, mobile STA (MS), remote STA, remote terminal, user terminal (UT), user agent, user device, user equipment (UE), user STA, or other terminology. In some implementations, a STA may include a cellular telephone, a cordless telephone, a session initiation protocol ("SIP") telephone, a wireless local loop ("WLL") STA, a personal digital assistant ("PDA"), a handheld device with wireless connectivity capabilities, or other suitable processing device connected to a wireless modem. Accordingly, one or more aspects taught herein may be incorporated into a telephone (e.g., a cellular phone or smartphone), a computer (e.g., a laptop), a tablet, a portable communication device, a portable computing device (e.g., a personal data assistant), an entertainment device (e.g., a music or video device, or satellite radio), a global positioning system (GPS) device, or any other suitable device configured to communicate via a wireless or wired medium. In some aspects, a non-AP STA may be a wireless node. Such a wireless node may, for example, provide connectivity to a network (eg, a wide area network such as the Internet or a cellular network) via a wired or wireless communication link.

[0090] An AP manages a set of STAs (registered or associated with the AP) that together constitute access to the wireless medium for communication purposes. The STAs (including the APs with which they register) form a service set, hereafter referred to as a Basic Service Set (BSS) (although other terms may be used). The same physical STA acting as an access point may manage two or more BSSs (and thus corresponding WLANs), each BSS being therefore uniquely identified by a specific Basic Service Set Identifier (BSSID) and managed by a separate virtual AP implemented in the physical AP. Each STA is identified within a BSS by an identifier AID assigned by the AP upon registration.

[0091] The 802.11 family of standards defines various medium access control (MAC) mechanisms for driving access to the wireless medium.

[0092] Current discussions in the 802.11be task group, as outlined in the March 2022 draft IEEE P802.11be / D1.5, will introduce multi-link operation (MLO) to the MAC layer operation, allowing a multi-link device to establish or configure multiple links and operate them simultaneously.

[0093] A multilink device (MLD) is a logical entity that has multiple affiliated STAs (STAs), has a single medium access control (MAC) service access point (SAP) for a logical link control (LLC), and contains one MAC data service ( ). An access point multilink device (or AP MLD) corresponds to an MLD where each STA affiliated to the MLD is an AP (hence referred to as an "affiliated AP"). A non-access point multilink device (or non-AP MLD) corresponds to an MLD where each STA affiliated to the MLD is a non-AP STA (referred to as an "affiliated non-AP STA"). In some literature, the terms "multilink device," "ML device" (MLD), "multilink logical entity," "ML logical entity" (MLE), "multilink set," and "ML set" are synonyms referring to the same type of ML device. An exemplary architecture of a multilink device is described below with reference to FIG. 1ab.

[0094] Multiple affiliated non-AP STAs in a non-AP MLD can set up communication links with multiple affiliated APs in an AP MLD, forming a multilink channel.

[0095] Links established for MLD (or "enabled links") are theoretically independent, meaning that channel access procedures (to the communication medium) and communication are performed independently on each link. Thus, different links may have different data rates (e.g., due to different bandwidths, number of antennas, etc.) and may be used (over each particular link) to communicate different types of information.

[0096] Thus, a communication link or "link" corresponds to a given channel (e.g., 20 MHz, 40 MHz, etc.) in a given frequency band (e.g., 2.4 GHz, 5 GHz, 6 GHz) between an AP affiliated with an AP MLD and a non-AP STA affiliated with a non-AP MLD.

[0097] Affiliated APs and non-AP STAs operate on their respective channels according to one or more of the IEEE 802.11 standards (a / b / g / n / ac / ad / af / ah / aj / ay / ax / be) or other wireless communication standards.

[0098] Multi-link aggregation theoretically allows traffic associated with one MLD to be transmitted over multiple parallel communication links, thereby increasing network capacity and maximizing the utilization of available resources.

[0099] From an architectural perspective, an MLD typically includes several radios to implement affiliated STAs, but the number does not have to be equal to the number of affiliated STAs. In particular, a non-AP MLD may operate with a number of affiliated STAs greater than the number of radios (or even as few as one).

[0100] The D1.5 standard defines several enhanced multi-link operating modes (abbreviated EML OM) from this physical architecture: Enhanced Multi-Link Single Radio (EMLSR) and Enhanced Multi-Link Multi-Radio (EMLMR). The D1.5 standard states that the two modes EMLSR and EMLMR are mutually exclusive.

[0101] Non-AP MLD declares support for EMLSR and / or EMLMR mode to AP MLD during the association phase (in the so-called EML Capabilities). In operational mode, activation and deactivation of EMLSR or EMLMR mode is initiated by non-AP MLD, which sends a specific EHT action frame called "EML OM Notification" indicating, among other things, the set of valid links (so-called EMLSR or EMLMR links) to which the activation of EMLSR or EMLMR mode applies. Typically, an "EMLSR / EMLMR link" consists of two valid links. However, more valid links may be used.

[0102] When EMLSR mode is enabled, the non-AP MLD simultaneously listens to a set of enabled links in the "EMLSR links" to receive Initial Control frames (e.g., MU-RTS trigger frames and BSRP trigger frames) sent by the AP MLD, and then can exchange data frames with the AP MLD on only one link at a time (typically the link on which the Initial Control frame was received). Each non-AP MLD may or may not support the EMLSR mode of operation.

[0103] In EMLMR mode, non-AP MLD can aggregate some of the physical resources of multiple radios for multiple active links (so-called EMLMR links) to transmit and receive data up to a predetermined number of supported receive and transmit spatial streams. This predetermined number is greater than the number of receive and transmit spatial streams supported per radio, providing improved throughput and reduced latency. As an example, a multi-radio (MR) non-AP MLD supporting EMLMR mode with two links (each with associated radios) communicates over two links using each of the two radios when EMLMR mode is deactivated, such as in a 2x2 MIMO antenna configuration for each radio. On the other hand, an MR non-AP MLD aggregates the physical resources (usually antennas) of the two radios when EMLMR mode is activated, such as in a 4x4 MIMO antenna configuration, and uses one of the radios to communicate over one of the two links. At the same time, the other link (the link with the deprived physical antenna) is unavailable.

[0104] When EMLMR mode is enabled, the non-AP MLD simultaneously listens to a set of enabled links of "EMLMR links" to receive the Initial frame sent by the AP MLD and initiate frame exchange, and then can perform data frame exchange with the AP MLD (by aggregating radio resources) over only one EMLMR link at a time (usually the link that received the Initial frame).

[0105] In the following description, for simplicity, the EMLSR mode will be mainly described, but similar considerations can be made for the EMLMR mode.

[0106] FIG. 1 illustrates a typical 802.11 network environment including ML transmission between EML-enabled MLDs (EMLSR and EMLMR-enabled) in which the present invention may be implemented.

[0107] Wireless communication network 100 includes AP MLD 110 and two non-AP MLDs 120 and 130. In this example, the two non-AP MLDs are assumed to be EML-capable and declare their corresponding capabilities to AP MLD 110 in EMLSR-related and EMLMR-related fields of their EML Capabilities (these fields are hereinafter referred to as EMLSR Capabilities and EMLMR Capabilities, e.g., subparts of EML Capabilities). Of course, a different number of non-AP MLDs registering with and exchanging frames with AP MLD 110, as well as a different (or greater) number of EML-capable non-AP MLDs, may be envisioned.

[0108] The AP MLD 110 has multiple affiliated APs, two affiliated APs 111 and 112 (also referred to as AP1 and AP2, respectively) in the exemplary FIG. 1 , each operating as an 802.11 AP on an operating channel within a frequency band. Known 802.11 frequency bands include the 2.4 GHz band, the 5 GHz band, and the 6 GHz band. Of course, other frequency bands may be used instead of or in addition to these three frequency bands.

[0109] The non-AP MLDs 120, 130 have multiple affiliated non-AP STAs, each operating as an 802.11 non-AP STA in the BSS (managed by the affiliated AP 111 or 112) to which it registers. In the exemplary FIG. 1 , two non-AP STAs 121 and 122 (also referred to as A1 and A2, respectively) are affiliated with the non-AP MLD 120, and two non-AP STAs 131 and 132 (also referred to as B1 and B2, respectively) are affiliated with the non-AP MLD 130.

[0110] For illustrative purposes, non-AP MLDs 120 and 130 are single-radio non-AP MLDs. For example, AP 111 is configured to operate on channel 38, which corresponds to an operating 40 MHz channel in the 5 GHz frequency band, and AP 112 is configured to operate on channel 151, which corresponds to another operating 40 MHz channel, also in the 5 GHz frequency band. In another example, affiliated STAs may operate on different frequency bands.

[0111] Each affiliated AP provides a link toward the AP MLD 110 to affiliated non-AP STAs in the non-AP MLD (120 or 130). Thus, each non-AP MLD link may be identified simply by the identifier of the respective affiliated AP. In this context, each affiliated AP 111 and 112 may be identified by an identifier called a "link ID." The link ID of each affiliated AP is unique and does not change during the lifetime of the AP MLD. The AP MLD may assign link IDs to affiliated APs by incrementing the ID from 0 (for the first affiliated AP). Of course, other terms, such as "AP ID," may be used as variations.

[0112] To perform multi-link communication, each non-AP MLD 120, 130 must discover, authenticate, associate, and set up multiple links with the AP MLD 110, with each link established between an affiliated AP of the AP MLD 110 and an affiliated non-AP STA of the non-AP MLD. Each such link, called an "enabled link," allows individual channel access and frame exchange between the non-AP MLD and the AP MLD based on the supported capabilities exchanged during association.

[0113] The discovery phase is called the ML discovery procedure, and the multilink setup phase (or association phase) is called the ML setup procedure.

[0114] The ML discovery procedure enables a non-AP MLD to discover wireless communication network 100, e.g., various links to an AP MLD provided by multiple affiliated APs. Accordingly, the ML discovery procedure attempts to advertise various affiliated APs in the AP MLD along with their respective network information (e.g., including all or part of their capabilities and operating parameters). After the non-AP MLD discovers wireless communication network 100 through the ML discovery procedure and the MLD authentication procedure, the ML setup procedure can select a set of candidate link setups between affiliated non-AP STAs in the non-AP MLD and some of the discovered affiliated APs and request the AP MLD 110 to set up these links, which can be accepted or rejected by the AP MLD. If the AP MLD accepts, the non-AP MLD is provided with an Association Identifier (AID) by the AP MLD, which is used by the affiliated non-APs in the non-AP MLD to wirelessly communicate with their corresponding affiliated APs over multiple links (communication channels). During the ML setup procedure, a non-AP MLD declares some or all of its capabilities, e.g., EMLSR capabilities. To do this, appropriate fields are provided in management frames. In particular, management frames exchanged during the ML discovery and ML setup procedures contain a new information element specific to multi-link operation (MLO), called the Basic Multi-Link element. In fact, in all management frames containing the Basic Multi-Link element, except authentication frames, a non-AP or AP MLD that is EMLSR-capable (dot11EHTEMLSROptionImplemented is true) or EMLMR-capable (dot11EHTEMLMROptionImplemented is true) sets the EMLSR or EMLMR Support bit in the EML Capabilities subfield of the Common Info field to 1.

[0115] For illustrative purposes, in wireless communication network 100, during an ML setup procedure, two setup candidate links are requested by non-AP MLD 120 and accepted by AP MLD 110: a first link 151 between affiliated AP 111 (AP1) and affiliated non-AP STA 121 (A1), and a second link 152 between affiliated AP 112 (AP2) and affiliated non-AP STA 122 (A2). Similarly, two setup link candidates are requested by multi-radio non-AP MLD 130 and accepted by AP MLD 110: a first link 161 between affiliated AP 111 (AP1) and affiliated non-AP STA 131 (B1), and a second link 162 between affiliated AP 112 (AP2) and affiliated non-AP STA 132 (B2).

[0116] AP MLD 110, non-AP MLD 120, and non-AP MLD 130 are EMLSR capable (dot11EHTEMLSROptionImplemented is true) or EMLMR capable (dot11EHTEMLMROptionImplemented is true). EMLSR or EMLMR capabilities (subpart of EML Capabilities) are exchanged during the ML discovery procedure and the multilink setup phase.

[0117] The EMLSR and EMLMR capabilities currently defined in the D1.5 standard include the following subfields: - the "EMLSR Support" subfield indicates that the MLD supports EMLSR operations. The EMLSR Support subfield is set to 1 if the MLD supports EMLSR operations, otherwise it is set to 0; - The 3-bit subfield "EMLSR Padding Delay" indicates the minimum MAC padding duration of the Padding field of the Initial Control frame requested by non-AP MLD, as defined in Enhanced multi-link single radio operation (clause 35.3.17). The table converts the 3-bit value into a padding delay in μs. This delay is used to define the transition period required for MLD to switch the state of an affiliated station from the listening operation state to the enabled / disabled frame exchange state. This transition period is the duration of the Initial Control frame response or Initial frame response, as described below, plus this delay. Therefore, this transition period is called the "EMLSR active switch delay" or "EMLMR active switch delay" depending on the active EML mode, or more generally, the "EML active switch delay" in the following; - The 3-bit subfield "EMLSR Transition Delay" indicates the transition delay time required for a non-AP MLD to switch from the so-called frame exchange mode (on one of the active links) to the so-called listening operation mode on the active links. The table converts the 3-bit value into a delay in μs, e.g., 0 for 0 μs, 1 for 16 μs, 2 for 32 μs, 3 for 64 μs, 4 for 128 μs, 5 for 256 μs, and values ​​from 6 to 7 are reserved; - The "EMLMR Support" subfield indicates that the MLD supports EMLMR operation. If the MLD supports EMLMR operation, the EMLMR Support subfield is set to 1, otherwise it is set to 0; - The 3-bit subfield "EMLMR Delay" indicates the minimum padding period required for switching an EMLMR link when a non-AP MLD operates in EMLMR mode. This delay is used to define the transition period required for the MLD to switch the state of an affiliated station when starting or ending a frame exchange; - The "Transition Timeout" subfield indicates the timeout value for EML Operating Mode Notification frame exchange in the EMLSR (or EMLMR).

[0118] When an EMLSR (or EMLMR)-capable non-AP MLD wishes to operate in the corresponding mode on a set of valid links, called EMLSR (or EMLMR) links, a STA affiliated with the non-AP MLD sends an EML Operating Mode (OM) Notification frame (as specified in the D1.5 standard) with the EMLSR (or EMLMR) Mode subfield of the EML Control field set to 1 to the AP affiliated with the EMLSR (or EMLMR)-capable AP MLD (here, AP MLD 110). The EMLSR (or EMLMR) links are indicated by setting the bit position of the EMLSR (or EMLMR) Link Bitmap subfield of the EML Control field of the EML OM Notification frame to 1 for each EMLSR (or EMLMR) link. For example, in the EMLSR (or EMLMR) Link Bitmap, bit position i corresponds to the link whose link ID is equal to i and is set to 1 to indicate that the link is a member of the EMLSR (or EMLMR) link.

[0119] An AP affiliated to an AP MLD that receives an EML Operating Mode Notification frame from a STA affiliated to a non-AP MLD then transmits an EML Operating Mode Notification frame to one of the STAs affiliated to the non-AP MLD as an acknowledgment of the EML Operating Mode Notification sent by the STA affiliated to the non-AP MLD within the timeout interval indicated in the Transition Timeout subfield of the EML Capabilities subfield of the Basic Multi-Link element and beginning at the end of the PPDU sent by the AP affiliated to the AP MLD.

[0120] After successful transmission of an EML Operating Mode Notification frame by a STA affiliated to a non-AP MLD over one of the EMLSR (or EMLMR) links, the non-AP MLD operates in EMLSR (or EMLMR) mode and the EMLSR is considered active (or EMLMR is active).

[0121] When an EMLSR-enabled non-AP MLD attempts to disable EMLSR (or EMLMR) mode, the STA affiliated to the non-AP MLD sends an EML Operating Mode (OM) Notification frame (as specified in the D1.5 standard) with the EMLSR (or EMLMR) Mode subfield of the EML Control field set to 0 to the AP affiliated to the AP MLD. Again, an AP affiliated to the AP MLD that receives an EML Operating Mode (OM) Notification frame from a STA affiliated to the non-AP MLD sends an EML Operating Mode Notification frame as described above as an acknowledgment of the EML Operating Mode (OM) Notification frame. After the STA affiliated to the non-AP MLD successfully sends the EML Operating Mode Notification frame via one of the EMLSR (or EMLMR) links, the non-AP MLD disables EMLSR (or EMLMR) mode.

[0122] The set of STAs affiliated with an EMLSR (or EMLMR)-enabled non-AP MLD operating on an EMLSR (or EMLMR) link may be all or a subset of the STAs affiliated with the non-AP MLD, and this set of STAs is hereinafter referred to as the "EMLSR co-affiliated STAs" (or EMLMR co-affiliated STAs) of the non-AP MLD.

[0123] In the example of FIG. 1, the co-affiliated STAs of the EMLSRs of non-AP MLD 120 and non-AP MLD 130 operate on the same link (e.g., the same affiliated AP, AP1 and AP2), meaning they share the same EMLSR link.

[0124] FIG. 1a shows an exemplary 802.11be multilink reference model for MLD, either AP MLD or non-AP MLD.

[0125] The MLD includes a PHY layer 200, a MAC layer 220, a Logical Link Control (LLC) sublayer and upper layers.

[0126] The upper layers may include applications that generate traffic data or use received traffic data.

[0127] Transmission and reception of traffic data is handled by the MAC 220 and PHY 200 layers. Such transmission and reception of traffic data may occur over multiple links 20-x, 20-y, and 20-z, such as links 151, 152, 161, and 162 introduced with reference to FIG. 1. Three links, and therefore three affiliated stations, are shown. Of course, other configurations including two affiliated stations or more than three affiliated stations are also contemplated.

[0128] Traffic data is provided from higher layers as a series of data frames, or "traffic streams." Each traffic stream, and therefore each data frame, is associated with an Access Category (AC), as defined by the EDCA mechanism (FIG. 1b). This mapping between streams or data frames and ACs is performed by classifier 213.

[0129] Recall that 802.11 stations (AP and non-AP stations) maintain four access categories (ACs), each with one or more corresponding transmit buffers or queues. The four ACs are conventionally defined as follows: - AC1 and AC0 are reserved for best effort and background traffic, which have the second lowest and lowest priority, respectively. - AC3 and AC2 are typically reserved for real-time applications (such as voice and video transmissions). They have the highest and second highest priority, respectively.

[0130] Data frames, also known as MSDUs (MAC Service Data Units), coming from higher layers of the protocol stack are mapped by the classifier 213 to one of the four ACs and are therefore entered into the queue of the AC to which they are mapped.

[0131] Figure 1b shows an implementation model with four transmission queues, one for each access category.

[0132] The 802.11be multi-link reference model reflects the fact that an MLD may transmit and receive using multiple links, particularly at the MAC layer 220 and PHY layer 200 levels.

[0133] The MAC layer 220 includes one Unified Upper-MAC (UMAC) layer 230 and multiple Lower-MAC (LMAC) layers 220-x, 220-y, 220-z combined with respective PHY layers 200-x, 200-y, 200-z, each combination corresponding to a link 20-x, 20-y, 20-z.

[0134] The UMAC 230 performs functions common to all links, and each LMAC 220-x, 220-y, 220-z performs functions local to each link 20-x, 20-y, 20-z. The UMAC layer provides a UMAC interface to the link-specific blocks 220-x, 220-y, 220-z and a UMAC Service Access Point (SAP) to the LLC and upper layers.

[0135] The UMAC 230 is responsible for link-independent MAC procedures such as authentication, association, security association, sequence number allocation, MAC Protocol Data Unit (MPDU) encryption / decryption, aggregation / deaggregation, and acknowledgement scoreboarding procedures.

[0136] The type of traffic (each data unit (MSDU) with User Priority (UP) priority, and therefore Traffic IDentifer (TID)) arriving at the MAC layer 220 from a higher layer (e.g., link layer) is mapped to one of the ACs according to a mapping rule in the UMAC layer 230. Then, also in the UMAC layer 230, the data unit (MSDU) is provided with the next available sequence number and stored in the queue corresponding to its TID (or UP) within the mapped AC. Recall that an 802.11 station maps TIDs to ACs as follows (TIDx means TID=x): - TID1 and TID2 are mapped to AC0, which is typically used for background traffic, - TID0 and TID3 are mapped to AC1, which is typically used for best-effort traffic, - TID4 and TID5 are mapped to AC2, which is typically used for video traffic, - TID6 and TID7 are mapped to AC3, which is typically used for voice traffic.

[0137] Each LMAC 220-x, 220-y, 220-z is responsible for link-specific functions such as channel access. In particular, each MLD Lower MAC includes its own contention-based channel access procedure, e.g., EDCA 221-x, 221-y, 221-z. Some functions require joint processing by both the UMAC 230 and the LMAC 220-x, 220-y, 220-z.

[0138] As shown in Figure 1a, each EDCA 221-x, 221-y, 221-z per link performs contention for each AC queue per link. In that regard, each AC has its own set of queue contention parameters (i.e., EDCA access parameters) per link, associated with a priority value, thus defining higher or lower priority traffic for an MSDU. Thus, multiple traffic queues exist for serving data traffic with different priorities for a given link. The arbitration inter-frame space (AIFSn), contention window (CW), and backoff values, known as EDCA access parameters, are specific to each AC on each link 20-x, 20-y, 20-z.

[0139] The default EDCA access parameters for 802.11 stations are shown in the table below: TIFF0007799860000001.tif44112

[0140] Also, recall that for 802.11 stations, the backoff value is chosen randomly from the range [0, CW], with CW typically initialized to the value of CWmin. This backoff value is used to initialize the backoff counter (BC). During the backoff procedure, if the medium is sensed idle for a given slot time (typically 9 μs), the backoff counter is decremented by 1; if the medium is sensed busy for a given slot time, the backoff counter is paused.

[0141] Also, recall that IEEE Std. 802.11ax-2021 introduced the multi-user (MU) EDCA parameter set in addition to the EDCA parameter set. Because 802.11ax stations can transmit UL data using both EDCA contention-based and trigger-based transmissions, they have more medium access opportunities than legacy 802.11 stations. Therefore, to ensure fairness, IEEE Std. 802.11ax-2021 introduced the MU EDCA mechanism, which deprioritizes 802.11ax stations after trigger-based UL transmissions by using the (less favorable) MU EDCA parameter set instead of the legacy EDC parameter set for a period specified in the so-called MU EDCA Timer (one per AC). Using the MU EDCA parameter set typically results in longer contention and backoff periods. The AP signals the legacy EDCA parameter sets and the MU EDCA parameter sets in broadcast beacon frames and / or association response frames that it exchanges with non-AP STAs.

[0142] For MLD, each AC or traffic queue 210 is mapped to one EDCA engine 221 per link. Thus, each backoff entity 211 in a link-specific EDCA engine 221 uses queue contention parameters to initialize a backoff counter (BC) for each queue specific to each AC and link, associated with each AC queue 210 for subtracting a backoff value. In FIG. 1b, backoff counters BC[x0], BC[x1], BC[x2], and BC[x3] are associated with traffic queues 210 for AC0, AC1, AC2, and AC3, respectively, and are simultaneously used to compete for access to link 20-x. Similarly, backoff counters BC[y0], BC[y1], BC[y2], and BC[y3] are associated with traffic queues 210 for AC0, AC1, AC2, and AC3, respectively, and are simultaneously used to compete for access to link 20-y. Similarly, backoff counters BC[z0], BC[z1], BC[z2], and BC[z3] are associated with traffic queues 210 AC0, AC1, AC2, and AC3, respectively, and are simultaneously used to compete for access to link 20-z. Here, a numbering scheme proposed for easily identifying backoff counters is BC[Link, AC]. For example, BC[z2] identifies the backoff counter corresponding to link 20-z and AC2. Next, it can be noted that in MLD, the number of backoff counters is typically equal to the number of ACs times the number of links. In the example of FIG. 1b, the number of backoff counters is 4 × 3 = 12.

[0143] However, it should be pointed out that this number is the maximum possible number of backoff counters. This is obtained when the default TID-to-Link mapping is used for MLD. Along with multilink operation, the D1.5 standard defines a TID-to-Link mapping mechanism that allows AP MLD and non-AP MLDs that have performed or are performing multilink setup to determine how to allocate UL and DL QoS traffic corresponding to TID values ​​from 0 to 7 to the non-AP MLD's setup links. By default, all TIDs are mapped to all setup links, both DL and UL, and all setup links are enabled. As a result, when the default TID-to-Link mapping is used, the maximum number of backoff counters is actually implemented. However, if the TID-to-Link mapping is negotiated between the AP MLD and non-AP MLD, mapping some TIDs to a set of links and other TIDs to a different set of links, the number of backoff counters is automatically reduced. As an example, consider a TID-To-Link mapping negotiated to map TIDs 4 and 5 belonging to AC2 only to link 20-x. Only BC[x2] is used; BC[y2] and BC[z2] are unused in this case.

[0144] In the remainder of this description, unless explicitly stated, the use of the default TID-To-Link mapping is considered.

[0145] The backoff counters are used to compete for access to link 20-x, 20-y, or 20-z to transmit data queued in the AC. In effect, the backoff counters are decremented from their initialized value when the medium is idle, and when the backoff counter reaches zero, the corresponding affiliated STA 201-x, 201-z is allowed to transmit (is granted access). It is understood that the TID-To-Link mapping (negotiated or default) affects the number of BCs that are simultaneously decremented in the non-AP MLD to drive EDCA access to each link.

[0146] Once an AC on a given link is granted access to the wireless medium, the MSDUs stored in the traffic queue 210 corresponding to that AC are sent to the physical (PHY) layers 200-x, 200-y, 200-z for transmission over the given link.

[0147] When MLD is multi-radio and simultaneous transmission and reception (STR) is enabled, each affiliated STA operates on the link independently of other affiliated STAs operating on other links.

[0148] 2 illustrates, using a frame sequence, the EMLSR operation mode in the non-AP MLD 120 when the AP MLD 110 decides to use the EMLSR mode. Of course, the EMLSR mode is emphasized here as an example, but similar considerations can be made for the EMLMR mode.

[0149] In this sequence, the non-AP MLD operates in EMLSR mode, meaning that an EML Operating Mode Notification frame activating EMLSR mode has been successfully sent by the affiliated STA of the non-AP MLD 120. In other words, it has entered an active Enhanced Multi-Link Single Radio (EMLSR) mode that applies to a specific set of two or more active links.

[0150] Affiliated STAs 121 and 122 are co-affiliated STAs of the EMLSR in the non-AP MLD 120. Each affiliated STA can be in one of three defined states: a listening operation state, a valid frame exchange state, and an invalid frame exchange state.

[0151] The non-AP MLD 120 can simultaneously listen on the EMLSR links by placing the co-affiliated STAs of the EMLSRs corresponding to those links in an "awake" or "listening" state. For example, affiliated STAs A1 and A2 are in the listening state (references 241 and 242). The listening state includes Clear Channel Assessment (CCA) and reception of Initial Control frames of a frame exchange initiated by the AP MLD. Thus, in the non-AP MLD 120, the co-affiliated STAs of the two EMLSRs simultaneously listen to reception of Initial Control frames from the AP MLD.

[0152] When the AP MLD 110 wishes to initiate a frame exchange with one or more non-AP MLDs on one of the EMLSR links, it initiates the frame exchange by transmitting an Initial Control frame 245 that explicitly triggers the non-AP MLDs. To some extent, the Initial Control frame schedules the non-AP MLDs. The Initial Control frame for a frame exchange is transmitted in an OFDM PPDU or non-HT duplicate PPDU format using a rate of 6 Mbps, 12 Mbps, or 24 Mbps (i.e., the MCS subfield in the frame is set to a value up to 2). As defined in the D1.5 standard, the Initial Control frame must be an MU-RTS trigger frame or a BSRP trigger frame, as defined in IEEE Standard 802.11ax-2021. Considering the trigger frame format according to a frame containing one or more User Info fields, this condition means that the frame 245 contains a User Info field addressed to the non-AP MLD, e.g., the AID12 field is set to the AID of the non-AP MLD (obtained during registration).

[0153] In this example, and as indicated by the reference "IC(A)," the Initial Control frame 245 explicitly triggers non-AP MLD A 120. The Initial Control frame may explicitly trigger multiple non-AP MLDs using multiple User Info fields therein.

[0154] Upon receiving the Initial Control frame 245, the co-affiliated STA of the non-AP MLD EMLSR (e.g., affiliated STA A1 in the embodiment) that is explicitly triggered initiates a state change of the co-affiliated STA of the non-AP MLD EMLSR under consideration (e.g., a state change of affiliated STAs A1 and A2 in the embodiment) and sends an Initial Control frame response (IC resp.) 246 to the AP AP1 affiliated to the AP MLD 110.

[0155] After receiving the Initial Control frame of frame exchange 245 and transmitting immediate response frame 246 in response to the Initial Control frame, the non-AP MLD-affiliated STAs listening on the corresponding link (e.g., co-affiliated STA A1 of the receiving EMLSR in this example) are configured to transmit or receive frames on the active link on which Initial Control frame 245 was received (e.g., link 151 in this example). To that end, a state switching procedure is initiated upon reception of frame 245, resulting in the co-affiliated STAs of the receiving EMLSR switching from the listening operation state 241 to the "active frame exchange" or "valid frame exchange" state, referenced 251 in the figure, after an EMLSR active switch delay. In this new state, the co-affiliated STAs of the receiving EMLSR are able to receive PPDUs transmitted using multiple spatial streams on the link on which Initial Control frame 245 was received. The EMLSR Active Switch Delay corresponds to the delay required for a non-AP MLD to switch from EMLSR listening mode of operation to EMLSR frame exchange mode. As mentioned above, this is derived from the notification specified in the EML Capabilities (via EMLSR Padding Delay) exchanged with the AP MLD: the duration of the Initial Control frame response 246 plus the EMLSR Padding Delay.

[0156] At the same time, other co-affiliated STAs of the same non-AP MLD (e.g., STA A2 in this example) are configured not to transmit or receive on the other EMLSR link until the frame exchange is complete. To this end, a state switching procedure is also initiated for the co-affiliated STAs of the other EMLSRs, which in turn switch from the listening operation state 242 to the "blindness frame" or "invalid frame exchange" state referenced at 252 in the figure. In particular, the AP MLD prevents data from being transmitted to these co-affiliated STAs of the other EMLSRs.

[0157] The state switching of all EMLMR co-affiliated STAs in the same non-AP MLD is inseparable and occurs simultaneously, since it is a matter of allocating a complete radio resource chain (see Figure 14 below) to one of the STAs and depriving the others of that chain. For EMLMR mode, the physical resources (e.g., antennas) of one radio resource chain are allocated (aggregated) to another radio resource chain, resulting in the former being deprived of transmit and receive capabilities (see Figure 15 below).

[0158] The above indicates that when a non-AP MLD operates in EMLSR mode (or more generally, either EMLSR mode or EMLMR mode), it is either in listening operation mode (its co-affiliated STAs are in listening operation state) or in frame exchange mode (one of its co-affiliated STAs is in valid frame exchange state and the other co-affiliated STAs are in invalid frame exchange state).

[0159] In EMLSR mode, a single complete radio resource is available allocated only to the co-affiliated STAs of the receiving EMLSR, as described below with reference to FIG. 14, whereas in EMLMR mode, the antenna resources of one of the radio stacks are allocated to the other radio stack, requiring simultaneous state changes, as described below with reference to FIG. 15.

[0160] It can be seen that only one of the co-affiliated STAs of an explicitly triggered non-AP MLD EMLSR can exchange data frames with the AP MLD at a time.

[0161] An exemplary frame exchange sequence is shown in the figure, which includes the transmission (hence downlink transmission) of an A-MPDU frame 255 by affiliated AP AP1 to co-affiliated STA A1 of the explicitly triggered EMLSR of non-AP MLD A 120, followed by a corresponding block acknowledgement 256 from the latter.

[0162] This means that after the EMLSR Transition Delay specified in the EML Capabilities, in addition to the completion of the frame exchange operated by the co-affiliated STAs of the receiving EMLSR, the non-AP MLD 120 switches to the listening operational state of the EMLSR, and the co-affiliated STA A1 of the receiving EMLSR switches to the listening operational state 241, as does the co-affiliated STA A2 of the other EMLSR (listening operational state 242). Thus, a state switching procedure is initiated for each of the co-affiliated STAs of the EMLSR.

[0163] The end of the frame exchange may be detected by the non-AP MLD (here, non-AP MLD 120) if one of the following conditions is met: (1) The MAC of a STA affiliated with a non-AP MLD that received Initial Control frame 245 did not receive a PHY-RXSTART.indication primitive within a timeout interval of aSIFSTime + aSlotTime + aRxPHYStartDelay, which begins at the end of a PPDU (e.g., acknowledgment 256) sent by the non-AP MLD STA in response to a frame (e.g., A-MPDU frame 255) most recently received from an AP affiliated with the AP MLD, or begins at the end of reception of a PPDU containing a frame from an AP affiliated with the AP MLD to the STA that does not require immediate acknowledgment. This indicates that the actual exchange with the AP MLD has ended without receiving a subsequent frame from the AP MLD. (2) The MAC of a STA affiliated to a non-AP MLD that received Initial Control frame 245 receives a PHY-RXSTART.indication primitive during a timeout interval of aSIFSTime + aSlotTime + aRxPHYStartDelay, which starts at the end of a PPDU (e.g., acknowledgment 256) sent by the STA in response to a frame most recently received from an AP affiliated to the AP MLD (e.g., AP-MPDU frame 255), or starts at the end of reception of a PPDU containing a frame from an AP affiliated to the STA that does not require immediate acknowledgment, and the STA affiliated to the non-AP MLD does not detect any of the following frames in the PPDU corresponding to the PHY-RXSTART.indication: - individually addressed frames with RA equal to the MAC address of a non-AP MLD affiliated STA; - a trigger frame with one of the User Info fields addressed to a non-AP MLD affiliated STA; - a CTS-to-self frame with RA equal to the MAC address of an AP affiliated to the AP MLD, - a Multi-STA BlockAck frame with one of the Per AID TID Info fields addressed to a non-AP MLD affiliated STA; - an NDP Announcement frame with one of the STA Info fields addressed to a non-AP MLD affiliated STA; This corresponds to the case where, after the actual exchange with the AP MLD, the non-AP MLD receives another frame from the AP MLD that is not addressed to the non-AP MLD (e.g., there is no data addressed to the non-AP MLD or there are no resources allocated to the non-AP MLD). (3) A STA affiliated with a non-AP MLD that received the Initial Control frame 245 does not respond to a frame (e.g., A-MPDU frame 255) most recently received from an AP affiliated with an AP MLD that requires an immediate response after a SIFS.

[0164] Now that the non-AP MLD 120 is in EMLSR listening mode of operation, the AP MLD may initiate a new frame exchange sequence (with either the non-AP MLD 120 or 130) by sending a new Initial Control frame.

[0165] In the illustrated example, the AP MLD 110 decides to start such a new sequence again with the non-AP MLD 120 using its EMLSR's co-affiliated STA A2 122. Specifically, the AP MLD 110 uses the other affiliated AP 112 to transmit a new Initial Control frame 265 IC(A) that explicitly triggers the non-AP MLD A120, and this frame is received by the EMLSR's co-affiliated STA A2 122. The receiving EMLSR's co-affiliated STA A2 122 transmits a response frame 266 to the Initial Control frame 265. After the EMLSR active switch delay, the explicitly triggered non-AP MLD 120 switches to EMLSR frame exchange mode, the receiving EMLSR's co-affiliated STA A2 122 switches from listening operation state 242 to valid frame exchange state 272, and the other EMLSR's co-affiliated STA A1 121 simultaneously switches from listening operation state 241 to invalid frame exchange state 271. Frames 275, 276 are then exchanged during the frame exchange sequence until the end of the sequence at which point the non-AP MLD 120 returns to the EMLSR's listening operation mode.

[0166] The A-MPDU 255 / 275 is provided by way of example only. Other types of frames may be transmitted by the AP MLD, such as, for example, a basic trigger frame to trigger an UL transmission. While FIG. 2 shows a frame exchange in which the acknowledgment 256 / 276 consists of a single frame 255 / 275, a simpler frame exchange may include only a single frame transmitted by the AP MLD without an acknowledgment, and a more complex frame exchange may include multiple exchange sequences, for example, cascaded TXOPs of UL transmissions (triggered by a basic trigger frame) and / or DL ​​transmissions (via an HE MU PPDU).

[0167] This example shows the advantages of EMLSR mode in terms of throughput and latency. AP MLD can quickly switch from one link to another, improving communication performance with less added complexity and cost.

[0168] In this example, the AP MLD 110 initiates a frame exchange sequence with one or more designated non-AP MLDs. The D1.5 standard also allows non-AP MLDs to initiate a frame exchange sequence with an AP MLD. In other words, STAs affiliated with a non-AP MLD operating in EMLSR mode do not need to send an Initial Control frame to initiate a frame exchange with the AP MLD. Such affiliated STAs access the wireless medium according to the rules defined in Section 10.3.2.4 (NAV Configuration and Reset) and Section 10.23.2 (HCF Contention-Based Channel Access (EDCA)).

[0169] However, traditional contention-based channel access is not defined with respect to the particularities of an EMLSR-active MLD (in particular, the state of co-affiliated STAs). Recall that an EMLSR-enabled non-AP MLD becomes EMLSR-active after successfully exchanging an EML OM Notification frame with an EMLSR-enabled AP MLD, in which the EMLSR Mode subfield of the EML Control field is set to 1, the EMLSR link is identified, and the co-affiliated STAs of the corresponding EMLSR are identified.

[0170] As mentioned above, the preceding description also applies to EMLMR mode with the following matching, among others: EMLMR Delay applies to both EMLSR Padding Delay and EMLSR Transition Delay; Initial Frame in EMLMR mode matches Initial Control Frame in EMLSR mode, and similarly Initial Frame Response in EMLMR mode matches Initial Control Frame Response in EMLSR mode; Although not specified in the D1.5 standard, the EMLMR listening operational state / mode may be defined to match the EMLSR listening operational state / mode in which co-affiliated EMLSRs are listening to their links prior to aggregation of physical radio resources.

[0171] An embodiment of the present invention seeks to arrange the EDCA procedure to suit the particularities of the EMLSR or EMLMR of a co-affiliated STA in either EML mode.

[0172] In a first embodiment, initiation of a frame exchange by a co-affiliated STA of a "first" EMLSR or EMLMR includes: a first switching of a first co-affiliated STA from a listening operation state to an active frame exchange state; and before initiating an enhanced distributed channel access (EDCA) backoff procedure that decrements a backoff counter to access the first link by the same first co-affiliated STA that is in a valid frame exchange state.

[0173] In these embodiments, the EML-active non-AP MLD changes its operating mode to EML frame exchange mode by decrementing its backoff counter before initiating contention on the wireless medium. As a result, state switching delays (EMLSR active switch delay or EMLMR active switch delay) are unlikely to affect contention procedures with co-affiliated STAs, particularly creating the risk that after contention ends, access to the medium has been gained by the other MLD, but the STA has not yet finished switching to transmit its first frame.

[0174] These first embodiments are illustrated by Figures 3 and 4, which highlight the EMLMR mode as an example: the same mechanism applies to the EMLMR mode with the matching terms mentioned above.

[0175] Figure 3 illustrates, by means of a flow chart, the steps performed by an EMLSR-active non-AP MLD to operate a first contention-based channel access procedure according to a first embodiment. Figure 4 schematically illustrates an example timeline of the first contention-based channel access procedure described in Figure 3.

[0176] The process begins in step 310, where the non-AP MLD enters EMLSR listening mode of operation. That is, the co-affiliated STAs of that EMLSR are set to a listening operational state and simultaneously listen to their respective EMLSR links. The non-AP MLD may enter EMLSR listening mode of operation in response to receiving an EML OM Notification frame with the EMLSR Mode subfield (of the EML Control field) set to 1. As a variant, the non-AP MLD may enter EMLSR listening operation by switching back from EMLSR frame exchange mode.

[0177] As shown in FIG. 4 with the EMLSR active non-AP MLD 120 in the EMLSR listening mode of operation, EMLSR co-affiliated STAs A1 121 and A2 122 are both in listening operational states 410 and 411 .

[0178] In step 320, the non-AP MLD waits to buffer uplink data to send to the AP MLD 110. As mentioned above, such data may be provided from higher layers and stored in the buffer 210 of the non-AP MLD.

[0179] When such data is identified, the non-AP MLD attempts to access one of the EMLSR links and initiate a contention-based channel access procedure to transmit the buffered data.

[0180] To that end, one of the EMLSR's co-affiliated STAs is selected to perform the contention-based channel access procedure in step 330. The selected STA is called the sending EMLSR's co-affiliated STA, and the corresponding link is called the sending EMLSR link.

[0181] Any selection method may be used, for example, relying on a round-robin strategy, a random strategy, or a load balancing distribution strategy based on link occupancy (in which case EMLSR links with low occupancy are selected preferentially), or a radio-based strategy (in which case co-affiliated STAs of EMLSRs with current full radios are selected preferentially).

[0182] In the example of Figure 4, EMLSR's co-affiliated STA1 is selected, which may be a station with a full radio or a station with a light (reduced functionality) radio.

[0183] Next, in step 340, a state switching procedure is initiated by the non-AP MLD to switch the mode from EMLSR listening operation mode to EMLSR frame exchange mode, and the co-affiliated STAs of the sending EMLSR are in a frame exchange enabled state.

[0184] For this purpose, the co-affiliated STAs of the sending EMLSR are switched from a listening operation state to an active frame exchange state (step 341), and in parallel (synchronously or simultaneously), the co-affiliated STAs of the other EMLSRs are switched from a listening operation state to an inactive frame exchange state (step 342).

[0185] In some embodiments, step 340 is conditioned on detecting that the corresponding link is idle through CCA verification. This means that the co-affiliated STAs of the sending EMLSR perform CCA during or immediately before triggering the switchover to ensure that the link is idle, and thus are ready for the EDCA backoff procedure. In these embodiments, the co-affiliated STAs of the sending EMLSR initiate or perform the switchover only if the corresponding link is detected as idle.

[0186] In FIG. 4 , if the non-AP MLD 120 initiates a frame exchange sequence with the AP MLD 110, wishes to transmit buffered data, and selects the EMLSR's co-affiliated STA A1 121 as the sending EMLSR's co-affiliated STA, it switches (possibly after CCA verification) the latter (A1 121) from listening operation state 410 to valid frame exchange state 420, and in parallel (synchronously or simultaneously) switches the EMLSR's co-affiliated STA A2 122 from listening operation state 411 to invalid frame exchange state 421.

[0187] The simultaneous switching continues up to the EMLSR active switch delay, defined above and indicated in the figure by reference numeral 499. In practice, the switching of co-affiliated STAs of an EMLSR with a full radio (only antenna connection required) is shorter than the switching of co-affiliated STAs of other EMLSRs with light radios (due to the need for physical reconfiguration of the full radio chain).

[0188] Once the switch occurs, the co-affiliated STAs of the sending EMLSR operate a backoff procedure on the sending EMLSR link in step 350. As shown in Figure 4, the co-affiliated STA A1 121 of the sending EMLSR operates a backoff procedure 422 by decrementing the counters 211 (only one shown in the figure) of the ACs 210 that have data to send as long as the medium (sending EMLSR link 151) is sensed idle.

[0189] The backoff counter is decremented in a conventional manner.

[0190] If the medium becomes busy during the decrement, the non-AP MLD may adopt various alternative actions.

[0191] In the first operation, the co-affiliated STAs of the sending EMLSR remain in an active frame exchange state and wait for the medium to return to an idle state, while the co-affiliated STAs of the other EMLSRs also maintain their current states.

[0192] In the second operation, the co-affiliated STAs of the EMLSR return to a listening mode of operation, regardless of whether the Duration field in the frame indicates the medium is busy. The non-AP MLD waits for a new opportunity to select one link for a new EDCA backoff procedure, such as initiating a new frame exchange with the AP MLD. Therefore, the switchback is an uncontrolled period, as opposed to the third operation below.

[0193] In a third operation, the co-affiliated STAs of the EMLSR still switch to a listening mode of operation, but only for a predefined period of time corresponding to the Duration field of the frame that caused the medium to become busy. At the end of that period, the co-affiliated STAs of the EMLSR switch to their previous states (i.e., the co-affiliated STAs of the sending EMLSR to a valid frame exchange state, and the co-affiliated STAs of the other EMLSR to an invalid co-affiliated STA of the sending EMLSR). This switch to the previous state may be conditioned on detecting that the corresponding link is idle via CCA verification (as described above).

[0194] In the fourth operation, the non-AP MLD first determines the Duration field of the frame that caused the medium to become busy, and then determines whether to return to the listening operation mode or remain in the current EMLSR frame exchange mode based on the Duration value. For example, if the Duration is long, i.e., if its value is higher than the high threshold, the EMLSR co-affiliated STAs are switched to the listening operation mode, similar to the second operation (i.e., no time limit). If the Duration is average, i.e., if its value is between the low and high thresholds, the EMLSR co-affiliated STAs are switched to the listening operation mode for a predefined time, similar to the third operation. On the other hand, if the Duration is short, i.e., if its value is lower than the low threshold, the EMLSR co-affiliated STAs remain in their current state, similar to the first operation. In a variant, only one threshold is used to distinguish between the first operation (short duration) and the second or third operation (long duration).

[0195] When the backoff counter reaches zero (step 360), the co-affiliated STAs of the transmitting EMLSR perform a frame exchange with the AP MLD, and in particular transmit buffered uplink data for the non-AP MLD in step 370. In the example of Figure 4, the co-affiliated STA A1 121 of the transmitting EMLSR transmits an A-MPDU frame 424 corresponding to its buffered uplink data to the AP MLD 110 via its corresponding EMLSR link, i.e., link 151, when one backoff counter reaches zero.

[0196] Optionally, before transmitting the A-MPDU frame 424, the co-affiliated STA A1 121 of the transmitting EMLSR may transmit an RTS frame or a CTS-to-self frame to better protect the granted channel.

[0197] Once the frame exchange on link 151 is complete, non-AP MLD 120 again initiates the state switching procedure to switch EMLSR co-affiliated STAs A1 121 and A2 122 back to the listening operational states 410, 411. Thus, non-AP MLD 120 returns to the EMLSR listening operational mode. The switchback operates for the EMLSR Transition Delay (specified in EML Capabilities) after the frame exchange is complete.

[0198] The first embodiment illustrated throughout Figures 3 and 4 may not be fully satisfactory because during the EDCA procedure (backoff decrement), the EMLSR's co-affiliated STA A2 122 is unable to receive the Initial Control frame 265.

[0199] A second embodiment provides that initiation of a frame exchange by a co-affiliated STA of a "first" EMLSR or EMLMR includes: initiating an Enhanced Distributed Channel Access (EDCA) backoff procedure by a first co-affiliated STA still in a listening state to access the first link, decrementing a backoff counter; When the backoff counter reaches a value of 0, the first affiliated STA switches from the listening operation state to the active frame exchange state.

[0200] In these embodiments, the EML active non-AP MLD performs medium contention (by decrementing its backoff counter or counters) before changing its operating mode to EML frame exchange mode. It is noted that Initial Control frames received on the other link during the EDCA procedure may be taken into account by the EML active non-AP MLD.

[0201] These second embodiments are illustrated by Figures 5a, 5b, 6a, 6b, 7 and 8, which highlight the EMLMR mode as an example. The same mechanism applies to the EMLMR mode with the matching terms mentioned above.

[0202] Figures 5a and 5b illustrate, by means of a flow chart, steps performed by an EMLSR-active non-AP MLD to operate a second contention-based channel access procedure according to a second embodiment. Figure 6a schematically illustrates an exemplary timeline of the second contention-based channel access procedure described in Figures 5a and 5b. Figure 6b schematically illustrates an alternative timeline of the second contention-based channel access procedure as described in Figures 5a and 5b.

[0203] Similar to Figure 3, processing begins at step 310, where the non-AP MLD enters the EMLSR's listening mode of operation. This means that the EMLSR's co-affiliated STAs are set to a listening operational state and are simultaneously listening to their respective links. As shown in Figure 6a, where the EMLSR active non-AP MLD 120 is in the EMLSR's listening mode of operation, the EMLSR's co-affiliated STAs A1 121 and A2 122 are both in listening operational states 610 and 611.

[0204] In step 320 , the non-AP MLD waits to buffer uplink data to send to the AP MLD 110 .

[0205] Once such data is identified, the non-AP STA selects one of the EMLSR's co-affiliated STAs to perform a contention-based channel access procedure in step 330. The selected STA is referred to as the sending EMLSR's co-affiliated STA, and the corresponding link is referred to as the sending EMLSR's link. An exemplary selection procedure is provided above. In Figure 6a, EMLSR's co-affiliated STA A1 121 is selected as the sending EMLSR's co-affiliated STA.

[0206] Next, in step 535, the non-AP MLD initiates a contention-based channel access procedure to access the sending EMLSR link. The sending EMLSR's co-affiliated STAs activate a backoff procedure. As shown in FIG. 6a, the sending EMLSR's co-affiliated STA A1 121 decrements its backoff counter while in the listening operation state 610.

[0207] The backoff counter is decremented in a conventional manner.

[0208] Contrary to the D1.5 standard, which states that a co-affiliated STA of an EMLSR in a listening operation state should only react upon receiving an Initial Control frame 245 or 265 (i.e., an MU-RTS trigger frame or a BSRP trigger frame), embodiments provide that it should react and set its NAV upon receiving any type of 802.11 control frame with an MCS up to 2 (i.e., up to 24 Mbps). Such frames correspond to OFDM PPDU or non-HT duplicate PPDU formats using rates of 6 Mbps, 12 Mbps, or 24 Mbps. In fact, a co-affiliated STA of an EMLSR (even one equipped with a light radio) can decode the Duration field of such frames.

[0209] As indicated by reference numeral 613, upon receiving such a frame (with an MCS value up to 2), the co-affiliated STA A1 121 of the transmitting EMLSR, which is performing the EDCA procedure 612 in the listening mode of operation 610, sets or updates its network allocation vector (NAV) based on the Duration field of this frame. The EDCA procedure can resume (614) once the NAV reaches zero.

[0210] In operation, this means that the co-affiliated STAs of the transmitting EMLSR receive, in step 540, an OFDM PPDU or non-HT duplicate PPDU format using a rate of 6, 12, or 24 Mbps.

[0211] Upon such receipt, the co-affiliated STAs of the sending EMLSR stop or suspend the EDCA procedure in step 542, which means that the backoff counter stops decrementing. Next, in step 544, the NAV of the co-affiliated STAs of the sending EMLSR is set or updated based on the Duration field of the received frame. The co-affiliated STAs of the sending EMLSR then wait for their NAV to expire before resuming the EDCA procedure in step 546.

[0212] When the backoff counter finally reaches zero (step 550 - signifying the end of the EDCA procedure), the non-AP STA switches from EMLSR listening operation mode to EMLSR frame exchange mode in step 560, and the co-affiliated STAs of the sending EMLSR enter the active frame exchange state.

[0213] To this end, the co-affiliated STAs of the sending EMLSR are switched from a listening operation state to an active frame exchange state (step 561), and in parallel (synchronously or simultaneously), the co-affiliated STAs of the other EMLSR are switched from a listening operation state to an inactive frame exchange state (step 562).

[0214] In FIG. 6 a, the non-AP MLD 120 initiates a frame exchange sequence with the AP MLD 110, wishes to transmit buffered data, selects the EMLSR's co-affiliated STA A1 121 as the transmitting EMLSR's co-affiliated STA, and when the backoff counter reaches zero, switches the latter (A1 121) from listening operation state 610 to valid frame exchange state 620, and in parallel (synchronously or simultaneously) switches the EMLSR's co-affiliated STA A2 122 from listening operation state 611 to invalid frame exchange state 621.

[0215] The simultaneous switching continues up to the EMLSR active switch delay specified above.

[0216] Once the switch occurs, the co-affiliated STAs of the transmitting EMLSR transmit their buffered uplink data in step 370. In the example of Figure 6a, the co-affiliated STA A1 121 of the transmitting EMLSR transmits an A-MPDU frame 624 corresponding to its buffered uplink data to the AP MLD 110 via the corresponding EMLSR link, i.e., link 151, when one backoff counter reaches 0.

[0217] Optionally, before transmitting the A-MPDU frame 624, the co-affiliated STA A1 121 of the transmitting EMLSR may transmit an RTS frame or a CTS-to-self frame to better protect the granted channel.

[0218] Once the frame exchange on link 151 is complete, non-AP MLD 120 again initiates the state switching procedure to return EMLSR co-affiliated STAs A1 121 and A2 122 to the listening operational states 610, 611. Thus, non-AP MLD 120 switches back to the EMLSR listening operational mode. The switch back operates for the EMLSR Transition Delay (specified in EML Capabilities) after the frame exchange is complete.

[0219] 6a, while in the listening state, the co-affiliated STA A1 121 of the sending EMLSR may again initiate the EDCA procedure with backoff counter decrement 632. Because the co-affiliated STA A2 122 of the other EMLSR is also listening to its EMLSR link 152, it may receive an Initial Control frame 634 from the co-affiliated AP AP2 112 over that EMLSR link while the co-affiliated STA A1 121 of the sending EMLSR is decrementing the backoff counter of the other EMLSR link 151.

[0220] In response to such reception, the non-AP MLD 120 can stop or suspend the EDCA procedure (and thus the decrement of the back-off counter) in STA A1 121, while the other EMLSR's co-affiliated STA A2 122 can send an Initial Control frame response (IC resp.) 635 to trigger or initiate a state change for the EMLSR's co-affiliated STAs and thus execute a frame exchange with AP2 112 over the second link 152. Of course, other strategies may be implemented by which the non-AP MLD 120 receiving the Initial Control frame 634 decides whether it is worthwhile to stop decrementing the back-off counter and execute an AP-initiated frame exchange, or to continue decrementing the back-off counter without responding to the Initial Control frame 634 in order to gain medium access for its own device to initiate a frame exchange. The decision (to stop or continue) in these strategies may be based on various parameters and / or policies.

[0221] As an example, the amount of buffered data can be considered. If that amount is equal to or greater than a threshold, a strategy of continuing decrementing is preferred to ensure medium access for non-AP MLD transmissions. The Initial Control frame 634 is discarded and no response is sent. On the other hand, if the amount is small, AP-initiated frame exchange is preferred (decrementing is stopped and a response 635 is sent).

[0222] As another example, if the co-affiliated STA A1 121 of the sending EMLSR has been assigned a full wireless stack, it may be worth prioritizing access by this STA to mitigate the time required to switch the full wireless stack to the co-affiliated STA of the other EMLSR. Therefore, the co-affiliated STA A1 121 of the sending EMLSR continues to decrement its backoff counter. The Initial Control frame 634 is discarded and no response is sent. On the other hand, if the co-affiliated STA A2 122 of the EMLSR has been assigned a full wireless stack, it may be worth prioritizing this STA. Therefore, the decrement is stopped and a response 635 is sent to perform an AP-initiated frame exchange.

[0223] As yet another example, if the Initial Control frame 634 is an MU-RTS trigger frame, indicating that the AP-initiated frame exchange is likely to be a downlink transmission, priority is given to the co-affiliated STAs of the transmitting EMLSR for medium access. Therefore, the co-affiliated STA A1 121 of the transmitting EMLSR continues to decrement its backoff counter. On the other hand, if the Initial Control frame 634 is a BSRP trigger frame, indicating that the AP-initiated frame exchange is likely to be an uplink transmission, priority is given to such AP-initiated frame exchange, in which a non-AP MLD may have resource units to transmit. Therefore, the decrement is stopped, and a response 635 is sent to perform the AP-initiated frame exchange.

[0224] As yet another example, the decision may be based on the previous uplink data load (or buffering) in a buffer or queue for a particular one of the EMLSR links. Indeed, some implementations of AC in the EDCA mechanism require preloading of uplink data into a particular queue so that it can be transmitted when the associated backoff counter expires. Of course, this preloading may only be triggered near the expiration of such a counter. The decision may be based on whether such preloading has occurred: if uplink data has already been preloaded into the buffer for the first link on which the EDCA procedure is currently being performed, the co-affiliated STA A1 121 of the transmitting EMLSR continues to decrement the backoff counter to gain medium access for itself. On the other hand, if uplink data has not yet been preloaded into the buffer for the first link, an AP-initiated frame exchange takes priority. Therefore, the decrementation is stopped, and a response 635 is sent to perform the AP-initiated frame exchange.

[0225] If response 635 is sent, then after the EMLSR active switch delay, the non-AP MLD 120 switches to EMLSR frame exchange mode, EMLSR co-affiliated STA A2 122 switches from listening operation state 611 to valid frame exchange state 641, and EMLSR co-affiliated STA A1 121 simultaneously switches from listening operation state 610 to invalid frame exchange state 640. Frames 644, 645 are then exchanged during the frame exchange sequence until the end of the sequence at which time the non-AP MLD 120 switches to EMLSR listening operation mode (not shown).

[0226] Typically, affiliated AP2 112 may transmit a basic trigger frame 644 to co-affiliated STA A2 122 of the EMLSR to allocate uplink resource units for the non-AP MLD 120, as specified in IEEE Standard 802.11ax®-2021. In such a case, via co-affiliated STA A2 122 of the EMLSR, the non-AP MLD 120 transmits a High-Efficiency Trigger-Based (HE TB) PPDU 645 in the allocated resource units.

[0227] The final flowcharts of Figures 5a and 5b illustrate the process of causing the non-AP MLD to pause or stop or abort the EDCA backoff procedure being performed on the first link of the EMLSR link upon receiving an Initial Control frame from the AP MLD over the second link of the EMLSR link (e.g., the backoff counter or counter decrement is stopped).

[0228] The co-affiliated STA of the other EMLSR (the other of the co-affiliated STAs of the sending EMLSR that decrements its backoff counter) receives an Initial Control frame from the AP MLD in step 570. This means that the AP MLD requests a new EMLSR sequence exchange to be initiated on the EMLSR link corresponding to the co-affiliated STA of the other EMLSR.

[0229] In step 575, the co-affiliated STA of the sending EMLSR pauses the backoff procedure, which means it stops decrementing the backoff counter.

[0230] Next, in step 580, the co-affiliated STA of the other EMLSR transmits a response to the Initial Control frame over its EMLSR link.

[0231] Then, according to the mechanism of the EMLSR, after the EMLSR active switch delay, in step 590, the non-AP STA switches from the EMLSR listening operation mode to the EMLSR frame exchange mode to operate the co-affiliated STAs of the other EMLSR for frame exchange.

[0232] For this purpose, the co-affiliated STAs of the other / requested EMLSR are switched from a listening operation state to an active frame exchange state (step 591), and in parallel (synchronously or simultaneously), the co-affiliated STAs of the sending EMLSR are switched from a listening operation state to an inactive frame exchange state (step 592).

[0233] This new mechanism can also be considered separately from the core aspect of the second embodiment, in this context, regarding a communication method in a wireless network in non-AP MLD operating in active EMLSR mode, comprising: Initiating an EDCA backoff procedure by a first STA that is in a listening operation state, is affiliated with the non-AP MLD, and corresponds to a first link of a set of valid links of links to which the EMLSR mode applies, to access the first link, decrementing a backoff counter; Here, the decrementing of the backoff counter is paused upon receiving an Initial Control frame from the AP MLD over the second link in the set.

[0234] In an embodiment, the decrement is resumed after the frame exchange initiated by the Initial Control frame is completed.

[0235] 6a may be particularly applicable when the co-affiliated STA of the sending EMLSR that decrements the backoff counter is a co-affiliated STA with a full radio. Indeed, in this case, the switch to the valid frame exchange state 620 is substantially immediate.

[0236] If the co-affiliated STA of the sending EMLSR is a co-affiliated STA with a light radio, the situation is slightly different due to the longer transition between the listening operation state 611 and the active frame exchange state 641. During this transition period, non-AP MLD functionality may be different and care must be taken to ensure access to the wireless medium.

[0237] Figure 6b shows a more detailed scenario, taking into account the transition period of Figure 6a. The length of the transition period corresponds to the active switch delay of the EMLSR mentioned above. The flowcharts of Figures 5a and 5b still apply.

[0238] Similar to FIG. 6a, the EMLSR's co-affiliated STAs A1 121 and A2 122 are both in listening operational states 610 and 611. The EMLSR's co-affiliated STA A1 121 is selected as the co-affiliated STA of the sending EMLSR. While in listening operational state 610, the sending EMLSR's co-affiliated STA A1 121 decrements its backoff counter (612). When the sending EMLSR's co-affiliated STA A1 121 receives a control frame (with an MCS value up to 2), it sets or updates its network allocation vector (NAV) based on the frame's Duration field (613). When the NAV reaches zero, the EDCA procedure may be restarted (614).

[0239] When the backoff counter of the sending EMLSR's co-affiliated STA A1 121 reaches zero, the non-AP MLD 120 switches STA A1 121 from the listening operation state 610 to the valid frame exchange state 620, while in parallel (synchronously or simultaneously) switching the EMLSR's co-affiliated STA A2 122 from the listening operation state 611 to the invalid frame exchange state 621.

[0240] In a first implementation, the non-AP MLD operates CCA during a transition period, e.g., during a switchover. In other words, to switch a co-affiliated STA of a sending EMLSR, the non-AP MLD initiates a state switch procedure for this STA while the STA is decrementing its back-off counter, and ends the state switch procedure as soon as the back-off counter reaches 0 (e.g., back-off counter expiration). Of course, the switchover of the co-affiliated STA of the other EMLSR occurs synchronously / simultaneously.

[0241] The switching of the co-affiliated STAs of the sending EMLSR is indicated in the figure by reference numeral 622a, and the switching of the co-affiliated STAs of the other EMLSR is indicated in the figure by reference numeral 623a.

[0242] In such a case, the EDCA procedure (decrementing the backoff counter) is not affected by the switching mechanism, and the two operations (decrement and switch) can be performed simultaneously. The switching mechanism 622a (resp. 623a) is pre-activated corresponding to the EMLSR active switch delay so that when the backoff counter reaches zero, the transmitting co-affiliated EMLSR STA A1 121 will be in a valid frame exchange state (and the other co-affiliated EMLSR STA will be in an invalid frame exchange state).

[0243] Similarly, the listening operational state is restored and becomes operational at the end of the transition period 625 corresponding to the switchback (the period defined by the EMLSR Transition Delay set in the EML Capabilities).

[0244] In a second embodiment, the non-AP MLD cannot operate CCA during the transition period, i.e., during the switchover. In such a case, the switchover mechanism 622b (resp. 623b in the figure) is activated after the backoff counter reaches zero. In other words, to switch the co-affiliated STAs of the sending EMLSR, the non-AP MLD initiates a state switch procedure for this STA in response to the backoff counter reaching the value 0. Of course, the switchover of the co-affiliated STAs of the other EMLSR is performed in a synchronous / simultaneous manner.

[0245] Reference numbers 622b and 623b indicate this implementation.

[0246] Since there is a risk that the medium will be claimed by another MLD, it is advantageous to protect the medium during the transition period that begins upon expiration of the back-off counter. In that regard, a protection frame 629 as a Request To Send (RTS) or Clear To Send (CTS-to-self) frame is transmitted by the co-affiliated STA A1 121 of the transmitting EMLSR during the switching procedure 622b before the transmission of the buffered uplink data 624 is actually performed. In other words, it is provided that in response to the back-off counter reaching the value 0, the non-AP MLD transmits a control frame on the link corresponding to the back-off counter.

[0247] To make the protected frame 629 readable by any legacy station on the medium, the frame preferably follows the format of an OFDM PPDU or non-HT duplicate PPDU using a rate of 6 Mbps, 12 Mbps, or 24 Mbps.

[0248] Because the length of the transition period is MLD dependent, the protection frame 629 is preferably sized to protect the medium until transmission 624. This may require that the protection frame 629 include padding to terminate the control frame after a point that precedes the end of the state switching procedure by a short interframe space (SIFS). Because the non-AP MLD knows the active switch delay of the EMLSR and also knows the conventional length of the control frame used, it has no problem determining the amount of padding required. Therefore, the co-affiliated STA A1 121 of the transmitting EMLSR can begin transmission 624 immediately after frame 629, given a legal SIFS period.

[0249] The above embodiment provides that a co-affiliated STA (e.g., a sender) of a single EMLSR performs the EDCA procedure and thus decrements its backoff counter. However, to increase the chance of gaining access to the wireless medium, the embodiment may trigger the EDCA procedure in two or more or all co-affiliated STAs of the EMLSR in the same non-AP MLD. This is illustrated in Figures 7 and 8. The EDCA backoff procedure is initiated simultaneously on two or more links of the EMLSR link. Of course, the link on which the co-affiliated station of the corresponding EMLSR switches to the valid frame exchange state is the link corresponding to the link on which the backoff counter of the EDCA backoff procedure first reaches 0. As shown in the above figure, the EMLMR mode is highlighted as an example. The same mechanism also applies to the EMLMR mode with the above-mentioned matching terminology.

[0250] 7 is a flowchart illustrating steps performed by an EMLSR-active non-AP MLD to operate a third contention-based channel access procedure according to a third embodiment. FIG. 8 is a schematic diagram illustrating an example timeline of the third contention-based channel access procedure described in FIG. 7.

[0251] As with the previous embodiment, processing begins at step 310, where the non-AP MLD enters the EMLSR listening mode of operation. That is, the EMLSR's co-affiliated STAs are set to listening operational states and are therefore simultaneously listening to their respective links. As shown in FIG. 8, where the EMLSR's active non-AP MLD 120 is in the EMLSR listening mode of operation, the EMLSR's co-affiliated STAs A1 121 and A2 122 are both in listening operational states 810 and 811.

[0252] In step 320 , the non-AP MLD waits to buffer uplink data to send to the AP MLD 110 .

[0253] If such data is identified, the non-AP STA no longer selects a co-affiliated STA of a single EMLSR and, in step 720, operates a random backoff procedure simultaneously and independently for each EMLSR's co-affiliated STA A1 121 (step 813) and A2 122 (step 814).

[0254] Again, as described above, the decrementing of the backoff counter may be paused (and thus the NAV set) upon receiving the control frame.

[0255] When the backoff counter of either EDCA procedure 813, 814 reaches zero (step 730), the non-AP STA switches from the EMLSR listening mode of operation to the EMLSR frame exchange mode in step 740, where the EMLSR co-affiliated STA corresponding to the expired backoff counter enters the active frame exchange state. This EMLSR co-affiliated STA is called a "ready EMLSR co-affiliated STA."

[0256] For this purpose, the co-affiliated STA of the ready EMLSR is switched from a listening operation state to an active frame exchange state (step 741), and in parallel (synchronously or simultaneously), the co-affiliated STA of the other EMLSR is switched from a listening operation state to an inactive frame exchange state (step 742).

[0257] 8, both EMLSR co-affiliated STAs A1 121 and A2 121 are decrementing their backoff counters. STA A1 121 is the first to have its backoff counter reach zero and is therefore the ready STA. In response to the counter expiring, ready STA A1 121 switches from a listening operation state 810 to a valid frame exchange state 820, while in parallel (synchronously or simultaneously), EMLSR co-affiliated STA A2 122 switches from a listening operation state 811 to an invalid frame exchange state 821.

[0258] The simultaneous switching continues up to the active switch delay of the EMLSR mentioned above.

[0259] Once the switchover occurs, the co-affiliated STAs of the ready EMLSR transmit their buffered uplink data in step 370. In the example of Figure 8, the co-affiliated STA A1 121 of the ready EMLSR transmits A-MPDU frames 825 corresponding to its buffered uplink data to the AP MLD 110 via the corresponding EMLSR link, i.e., link 151 to the co-affiliated AP AP1 111 of the AP MLD 110.

[0260] Optionally, before transmitting the A-MPDU frame 825, the co-affiliated STA A1 121 of the ready EMLSR may transmit an RTS frame or a CTS-to-self frame to better protect the granted channel.

[0261] Upon completion of the frame exchange initiated by the ready EMLSR co-affiliated STA A1 121, as well as the EMLSR Transition Delay specified in the EML Capabilities, the non-AP MLD 120 switches back to the EMLSR listening operation mode, i.e., the ready EMLSR co-affiliated STA A1 121 switches back to the listening operation state 810 (listening operation state 811), as does the other EMLSR co-affiliated STA A2 122.

[0262] As previously described with reference to FIG. 1ab, in MLD, each AC is mapped to one EDCA engine per link.

[0263] In the case of multi-radio MLD with simultaneous transmission and reception (STR), each affiliated STA operates on a link independently of other affiliated STAs operating on other links. In other words, the EDCA engine for one link can operate in a legacy manner because it is independent of other EDCA engines for other links. In the example of FIG. 1b, this means that EDCA engines 221-x, 221-y, and 221-z and their respective backoff entities 211 can operate independently of one another in a legacy manner.

[0264] However, in the case of an MLD operating in EML mode, EMLSR mode or EMLMR mode, an affiliated STA operating on an EMLSR or EMLMR link depends on another affiliated STA operating on another EMLSR or EMLMR link.

[0265] For example, when a co-affiliated STA of one EMLSR switches to an active frame exchange state on its EMLSR link, the co-affiliated STA of the other EMLSR automatically switches to an inactive frame exchange state on its other EMLSR link. This means that an EDCA engine processing an EML link has some dependency on another EDCA engine processing another EML link of the same EMLSR linkset or EMLMR linkset. This is because in an MLD operating in EML mode, data frame exchange effectively occurs on only one EML link of the set at a time. In the example of Figure 1b, EDCA engines 221-x, 221-y, 221-z and their respective backoff entities 211 cannot operate independently of each other in a legacy manner.

[0266] Therefore, the embodiment of the present invention seeks to arrange the EDCA procedure to suit the particularities of the EMLSR or EMLMR of its cooperating STAs.

[0267] In a non-AP MLD operating in an initially active EML mode, a backoff counter driving or managing EDCA access to a second link of a set of EMLSRs or EMLMRs may be paused in response to initiating a frame exchange with the AP MLD over a first link of that set, thereby avoiding obtaining EDCA access to the second link while the corresponding co-affiliated STA is unavailable (because the frame exchange allocated radio resources to the co-affiliated STA). It can be seen that this improves the efficiency of the EDCA procedure.

[0268] As an example of a specific EDCA-related issue, if an MLD is operating in EMLSR or EMLMR mode on links 20-x and 20-z, and BC[x3] becomes 0, which allows the MLD access to link 20-x to transmit the MSDU stored in AC3, the following dependency emerges: - the other three backoff counters BC[x0], BC[x1], and BC[x2] belonging to the same EDCA engine 221-x are suspended for legacy 802.11 operation; Since link 20-z is no longer operational due to MLD, the four backoff counters BC[z0], BC[z1], BC[z2], and BC[z3] belonging to EDCA engine 221-z are also suspended. This dependency is specific to MLD operating in EML mode on links 20-x and 20-z. In fact, when affiliated STA 201-x switches to an active frame exchange state on EMLSR / EMLMR link 20-x for transmission, the other co-affiliated STA 201-z automatically switches to an inactive frame exchange state on the other EMLSR / EMLMR link 20-z. In the inactive frame exchange state, medium sensing is not possible, and therefore the co-affiliated STA cannot process the EDCA backoff procedure. - Once the transmission of the MSDU stored in AC3 is completed on link 20-x, the MLD operating in EML mode switches back to the listening operation state on both links 20-x and 20-z. At this stage, another dependency emerges between EDCA engine 221-x and EDCA engine 221-z, specifically between their respective backoff counters BC[x3] and BC[z3]: In the EDCA engine 221-x, restarting or restarting the backoff counters BC[x0], BC[x1], BC[x2], and BC[x3] can be handled in the legacy manner: BC[x0], BC[x1], and BC[x2] are restarted, and BC[x3] is reinitialized. In EDCA engine 221-z, the restart or resumption of backoff counters BC[z0], BC[z1], BC[z2], and BC[z3] is problematic. While BC[z0], BC[z1], and BC[z2] may simply be restarted because no data from the corresponding ACs has been transmitted, this is not so trivial for BC[z3], which corresponds to the AC for which transmission occurred. In fact, the mere restart of BC[z3] may cause fairness issues among the ACs in the MLD, since in the next round, BC[z3] may gain access to link 20-z to transmit the MSDU stored in AC3 again.

[0269] Therefore, a new strategy for managing EDCA backoff restarts for STAs affiliated with a non-AP MLD operating in EML mode is proposed. In particular, embodiments of the present invention define a new EDCA backoff restart procedure for STAs affiliated with a non-AP MLD operating on a second EMLSR or EMLMR link after an UL transmission has occurred on a first EMLSR or EMLMR link of the same set.

[0270] In this regard, in response to the completion of the frame exchange on the first link, a restart strategy selected based on the characteristics of the frame exchange is applied to restart the paused backoff counter (driving access to the second link).

[0271] Therefore, the strategy or policy for restarting the backoff counter of a link depends on what happened (frame exchange) on another link. This differs from the traditional EDCA approach. By applying or adjusting appropriate EDCA backoff restart procedures, we can see that the network access fairness issue can be mitigated. In particular, the backoff counter associated with an AC on a link can be penalized when data from the same AC is transmitted over other links in the same set of EMLSR or EMLMR links.

[0272] More specifically, a procedure is proposed for restarting the back-off counter 211BC[n,m] running in the EDCA engine 221-n of link n of the second EMLSR or EMLMR after an uplink (UL) transmission of an MSDU of ACm is made on the link of the first EMLSR or EMLMR of the link set of the same EMLSR or EMLMR.

[0273] In practice, in response to the initiation of a frame exchange, multiple backoff counters driving EDCA on the second link for each of multiple access categories (e.g., one for each of the four ACs) are paused. In that case, a restart strategy may be applied to the paused backoff counters corresponding to the access categories exchanged during the frame exchange. In other words, if data from AC2 and AC3 are transmitted uplink over link 1 during the frame exchange, the paused backoff counters corresponding to AC2 and AC3 on link 2 may be restarted using the adapted restart strategy, and the paused backoff counters corresponding to the other ACs (AC0 and AC1) may be restarted in a conventional EDCA manner.

[0274] 9 illustrates, in a flow chart, the steps for handling restarting the backoff counter in EML mode according to an embodiment of the present invention. For ease of explanation, we will mainly refer to EMLSR mode, but the same applies to EMLMR mode.

[0275] The process begins in step 900, where the non-AP MLD enters a listening mode of operation (EMLSR or EMLMR), which means that its co-affiliated STAs are set to a listening state and are therefore simultaneously listening to their respective links. The non-AP MLD may enter a listening mode of operation in response to receiving an EML OM Notification frame with the corresponding Mode subfield (EMLSR or EMLMR Mode subfield of the EML Control field) set to 1.

[0276] A non-AP MLD may have all or some of its ACs in EDCA mode (e.g., using EDCA parameters for medium contention) and may already be in MU EDCA mode (i.e., using MU EDCA parameters for medium contention) if it has recently benefited from a trigger-based MU UL transmission for the corresponding AC. As is known, an MLD may switch one AC from MU EDCA mode back to EDCA mode upon expiration of the respective MU EDCA Timer.

[0277] In step 905, the non-AP MLD waits to buffer uplink data to transmit to the AP MLD 110. As described above, such data may be provided by higher layers and stored in the non-AP MLD's buffer 210. The non-AP MLD identifies all ACs that have buffered data to be transmitted to the AP MLD.

[0278] Next, for step 910, the non-AP MLD determines the back-off counter to be activated for contending for access to the EMLSR or EMLMR link (link 1 and link 2 in the figure).

[0279] In some embodiments, such as those based on the default TID-To-Link mapping, all backoff counters corresponding to the ACs identified for that link must be used for medium contention.

[0280] In some embodiments where TID-to-Link is negotiated, some TIDs (and therefore corresponding ACs) may be barred on one of the links, in which case non-AP MLD considers the negotiated TID-to-Link mapping and lists all backoff counters corresponding to the identified ACs for the links in the EMLMR or EMLSR's linkset that are not barred.

[0281] In other embodiments, the non-AP MLD may decide to contend for access to only one of the set of links, rather than contending for both links simultaneously. To do so, the non-AP MLD selects one of its co-affiliated STAs to operate a contention-based channel access procedure. The selected STA is referred to as the transmitting co-affiliated STA, and the corresponding link is referred to as the transmitting link. In the embodiment shown in FIG. 12, described below, the non-AP MLD decides to contend for access to only link 152. The non-AP MLD may also decide to contend for access to different links for different ACs.

[0282] Any selection procedure can be used, for example a round-robin strategy, a random strategy, or a load-balancing strategy based on link occupancy (in which case the least occupied links in the set are selected in preference).

[0283] In yet another embodiment, the non-AP MLD considers the AIFSN value applicable to each backoff counter to determine when it can begin decrementing each backoff counter.

[0284] Once the active backoff counters are known, they are started in step 915: they are decremented each time slot as long as the corresponding link is sensed as idle. This continues until a frame exchange begins over one of the links in the set (test 920).

[0285] Various events allow non-AP MLD to detect the start of a frame exchange.

[0286] 10a and 10b, the event is the expiration of one of the decremented backoff counters, in which case the non-AP MLD actually gains access to the link corresponding to the expiring backoff counter in order to transmit data belonging to the AC corresponding to the expiring backoff counter.

[0287] In another embodiment specific to EML mode, e.g., corresponding to Figures 11 and 12 below, the event is the receipt of an Initial frame (known as an Initial Control frame 245 in EMLSR mode) from the AP MLD over one of the links. Indeed, in this case, the non-AP MLD is involved in the frame exchange over that link.

[0288] Upon detecting the start of a frame exchange over one of the EMLSR or EMLMR's links (say link 1), the non-AP MLD suspends all other active backoff counters competing for access to the same link in step 925. This is the legacy behavior of 802.11 stations.

[0289] Due to the dependencies between the links in the set, the non-AP MLD must also suspend all active backoff counters competing for access to the other link in step 930 (because due to the switchover described below, the non-AP MLD will no longer be operating on this other link).

[0290] Next, in step 935, the non-AP MLD initiates a state switching procedure to switch its mode from listening operation mode to frame exchange mode, and the co-affiliated STA corresponding to link 1 (where the frame exchange is initiated) is placed in an active frame exchange state.

[0291] For this purpose, this co-affiliated STA is switched from a listening operation state to an active frame exchange state in order to perform frame exchange, and in parallel (synchronously or simultaneously), other co-affiliated STAs of the link set (of link 2) of the considered EMLSR or EMLMR are switched from a listening operation state to an inactive frame exchange state.

[0292] The non-AP MLD is now ready to exchange frames with the AP MLD over link 1 (step 940).

[0293] When the frame exchange over link 1 is completed (test 945), the non-AP MLD again initiates the state switching procedure (step 950) to return its co-affiliated STAs to a listening operational state. Thus, the non-AP MLD switches back to a listening mode of operation. The switch back occurs the EMLSR Transition Delay or EMLMR Delay (as specified in EML Capabilities) after the frame exchange is completed.

[0294] Next, the backoff counters for links 1 and 2 (which were paused in steps 925 and 930) are prepared for subsequent contention (eg, for their respective next decrements) in step 955.

[0295] The backoff counter for link 1 (ie, the link 1 over which the frame exchange occurred) may be restarted in a conventional manner.

[0296] When one backoff counter expires (i.e., when a non-AP MLD gains access to the medium for the corresponding AC, meaning a SU transmission), the backoff counter is reinitialized using the updated contention window (CW=CWmin for successful uplink transmission, CW=min(2.CW, CWmax) for unsuccessful uplink transmission). Other backoff counters not related to transmissions are simply restarted from their last value.

[0297] On the other hand, if the backoff counters have not expired (eg, if the frame exchange was initiated by AP MLD), the backoff counters of the ACs involved in the frame exchange are simply restarted.

[0298] These ACs may remain in their current EDCA mode for link 1 (legacy EDCA mode using the set of EDCA parameters of the corresponding BC of link 1, or MU EDCA mode using the set of MU EDCA parameters of the corresponding BC of link 1), in which case the contention window of the next re-initialization of the BC is updated to CWmin. An AC being in a given EDCA mode for a link means that the use of the backoff counter corresponding to this AC to access the link is driven by the parameter set applicable to the given EDCA mode.

[0299] The non-AP MLD may also switch these ACs on Link 1 from legacy EDCA mode to MU EDCA mode (e.g., using the set of MU EDCA parameters for medium contention) if they are currently in legacy EDCA mode and MU EDCA mode is activated on Link 1. It may then reinitialize the MU EDCA Timer if a frame exchange includes a successful trigger-based MU UL transmission. The contention window for the next reinitialization of that BC may be CWmin, as defined in the applicable parameter set. Other backoff counters not related to the frame exchange are simply restarted from their last value.

[0300] Alternatively, the non-AP MLD may switch these ACs of link 1 from their current EDCA mode (legacy EDCA mode or MU EDCA mode) to a so-called EML EDCA mode. Indeed, a new EML EDCA mode (e.g., EMLSR EDCA mode and EMLMR EDCA mode, if a distinction is provided between the management of the two EML modes) may be defined to provide a separate set of EML EDCA parameters (e.g., in beacon frames) for configuring the back-off counter of this AC for accessing link 1 and a separate EMLEDCATimer for driving the switch back to the legacy EDCA mode of this AC.

[0301] As will be described below, switching of the AC of link 1 to another EDCA mode may be independent of switching of the AC of link 2 to another EDCA mode.

[0302] However, preferably, the switching of the ACs of link 1 and link 2 is interdependent. For example, the ACs of link 1 and link 2 can be switched to the same other EDCA mode (MU EDCA mode, EML EDCA mode, EMLSR EDCA mode, EMLMR EDCA mode) even if the applied parameter sets have different values. In another example, when an AC of one link in the EMLSR / EMLMR set is switched to another EDCA mode, the same AC of the other link is also switched to another, possibly different, EDCA mode. For example, the AC of link 1, where frame exchange has occurred, is switched to MU EDCA mode, and the same AC of the other link of the set (link 2 in the example) is switched to any EML EDCA mode, or vice versa.

[0303] This is summarized in column 1330 of FIG. 13 below.

[0304] The backoff counters of link 2 (the other link in the EMLSR or EMLMR's set of links) are also prepared for the next medium contention. Recall that these suspended backoff counters do not expire; the backoff counters of ACs not involved in the frame exchange over link 1 are simply restarted from their last values.

[0305] The restart strategy used for the backoff counters of the ACs involved in the frame exchange over link 1 is based on the characteristics of the frame exchange.

[0306] In an embodiment, the frame exchange is based on the success or failure of the uplink transmission to the AP MLD (e.g., based on the status of the frame exchange). For example, the backoff counter is simply restarted from its last value if the uplink transmission in the frame exchange fails. The same simple restart applies to the case of downlink transmission. In fact, in all cases, the non-AP MLD does not utilize the transmission opportunity on link 1. On the other hand, the backoff counter may be reinitialized if the uplink transmission in the frame exchange is successful, even if it has not expired. This is to guarantee the transmission opportunity on link 1 with regard to fairness among the MLDs in the network.

[0307] In some embodiments, which may be combined with the previous ones, the strategy is based on the nature of the frame exchange, for example, whether the frame exchange includes a single-user uplink transmission to the AP MLD, a multi-user trigger-based uplink transmission to the AP MLD, or only downlink transmissions from the AP MLD.

[0308] For example, the backoff counters of the ACs involved in the frame exchange could be: - in the case of a frame exchange, downlink transmission or if uplink transmission fails, is resumed, - On the other hand, if the SU UL transmission in a frame exchange is successful, the backoff counter is reinitialized using the current contention window or restarted from its current value before starting to decrement it. Alternatively, if a multi-user trigger-based uplink transmission in a frame exchange is successful, the backoff counter may be reinitialized using the current contention window or a new contention window associated with the new EDCA mode before starting to decrement, or may be restarted from its current value. Similar to Link 1 above, the AC may remain in its current EDCA mode (legacy EDCA mode or MU EDCA mode) for Link 2, or may be switched from legacy EDCA mode to MU EDCA mode or EML EDCA mode, after which the MUEDCATimer or the applicable EMLEDCATimer may be reinitialized.

[0309] The switching of the AC of link 2 to a new EDCA mode may be independent of the switching of the AC of link 1 to another EDCA mode, as described above, but preferably the switching of the AC of link 1 and the AC of link 2 are interdependent, as described above.

[0310] The above-mentioned "new EDCA mode" is one of the MU EDCA mode, the EML EDCA mode, the EMLSR EDCA mode, and the EMLMR EDCA mode.

[0311] This is summarized in column 1340 of FIG.

[0312] If a frame exchange involves cascaded successive transmissions (for example, if the frame exchange is initiated by AP MLD and cascades multiple TXOPs), then successful transmissions are prioritized over failed uplink transmissions (for a given AC) in selecting a restart strategy. In other words, as soon as one uplink transmission is successful for a given AC, the above strategy used for successful transmissions is applied to the backoff counter of the AC, without considering the strategy related to the failed transmissions.

[0313] Next at step 955, the backoff counter is prepared for the next medium contention if there is more buffered data to be transmitted (looping back to step 905).

[0314] 10a and 10b schematically illustrate an exemplary timeline for a first EMLSR or EMLMR operation case including a backoff counter restart procedure according to an embodiment of the present invention, where the triggering event for the frame exchange is the expiration of the backoff counter driving the EDCA of link 1 (link 151). This means that the frame exchange that takes place is a single-user (SU) uplink transmission.

[0315] In Figure 10a, both co-affiliated STAs A1 121 and A2 122 decrement their backoff counters (step 915 above), but for purposes of illustration, few backoff counters are shown. Only one backoff counter decrement 1013 is shown for co-affiliated STA A1 121 on link 151, and one backoff counter decrement 1014 is shown for co-affiliated STA A2 122 on link 152. As an example, backoff counter decrements 1013 and 1014 correspond to the backoff counters of the same AC (here, AC3) where UL transmissions occur during the frame exchange. These counters are BC[151,3] and BC[152,3], respectively.

[0316] For both links 151 and 152, the backoff counters of the other ACs BC[151,0], BC[151,1], BC[151,2] and BC[152,0], BC[152,1], BC[152,2] are also decremented simultaneously with BC[151,3] and BC[152,3].

[0317] STA A1 121 is the first to have its backoff counter expire (step 920), where BC[151,3] reaches 0. STA A1 121 is therefore the STA ready to transmit. In response to the expiration of counter BC[151,3], ready STA A1 121 switches from listening operation state 1010 to valid frame exchange state 1020 (step 935), and in parallel (synchronously or simultaneously), co-affiliated STA A2 122 switches from listening operation state 1011 to invalid frame exchange state 1021. Also, in response to the expiration of counter BC[151,3], backoff counter decrement 1014 of backoff counter BC[152,3] is paused (step 930). The other backoff counters for links 151, 152 are also paused (steps 925, 930).

[0318] The simultaneous switching continues up to the EMLSR active switch delay or the EMLMR active switch delay as described above.

[0319] Once the switch occurs, the ready STA A1 121 transmits its buffered uplink data (step 940) in step 1025. In the example of Figure 10a, the ready co-affiliated STA A1 121 transmits an A-MPDU frame 1025 to the AP MLD 110 corresponding to its buffered uplink data for AC3 via the EMLSR or EMLMR link (e.g., link 151) to the AP MLD 110 corresponding to the co-affiliated AP1 111.

[0320] Optionally, before transmitting the A-MPDU frame 1025, the ready co-affiliated STA A1 121 may transmit an RTS frame or a CTS-to-self frame for better protection on the granted channel.

[0321] In addition to the completion of the frame exchange performed by the ready co-affiliated STA A1 121, after the EMLSR Transition Delay or EMLMR Delay specified in the EML Capabilities, the non-AP MLD 120 switches back to listening operation mode (step 950), which means that the ready co-affiliated STA A1 121, as well as the other co-affiliated STA A2 122, switches back to listening operation state 1010 (listening operation state 1011).

[0322] At this stage, both co-affiliated STAs A1 121 and A2 122 may process restarts of their backoff counters. On Figure 10a, backoff counter restart 1032 is represented for co-affiliated STA A1 121 of link 151, and backoff counter restart 1033 is represented for co-affiliated STA A2 122 of link 152. Backoff counter restart 1032 and 1033 correspond to backoff counters BC[151,3] and BC[152,3], respectively.

[0323] Referring to the table shown in Figure 13, the backoff counter restarts 1032 and 1033 depend on the status of the previous UL transmission of the A-MPDU frame 1025 (e.g., whether the SU UL transmission was successful or unsuccessful). The case of Figure 10a is shown in the first row 1301.

[0324] The backoff counter restart 1032 strategy for BC[151,3] is shown in column 1330 .

[0325] If the previous SU UL transmission 1025 was successful, the backoff counter BC[151,3] is processed in the legacy manner: it is simply reinitialized using the EDCA parameters from the range [0,CW=CWmin] and processed using the conventional EDCA backoff procedure.

[0326] If the previous SU UL transmission 1025 was not successful, the backoff counter BC[151,3] is processed in the legacy manner: it is simply reinitialized using the EDCA parameters from the range [0, CW=min(2xCW, CWmax)] and processed using the conventional EDCA backoff procedure.

[0327] The backoff counter restart 1033 strategy for BC[152,3] is shown in column 1340 .

[0328] If the previous SU UL transmission 1025 was successful, the backoff counter BC[152,3] may be reinitialized from the range [0,CW=CW] using EDCA parameters and processed using conventional EDCA backoff procedures. In this case, BC[152,3] is reinitialized to account for its dependency on BC[151,3] and the fact that the non-AP MLD benefited from the transmission opportunity on link 151. This restores some fairness to compensate for the additional transmission opportunity for the MLD. However, because no UL transmissions from STA A2 122 have occurred on link 152, the contention window may maintain the same (current) value to reflect the current state of this link.

[0329] Alternatively, it may be restarted from the current value and processed through conventional EDCA backoff procedures, then reinitialized from the range [0, CW=CW] and processed through conventional EDCA backoff procedures.

[0330] If the previous SU UL transmission 1025 was unsuccessful, the backoff counter BC[152,3] may be restarted from its current value and processed using conventional EDCA backoff procedures. In this case, because the SU UL transmission 1025 of AC3 on link 151 was unsuccessful, the priority of AC3 on link 152 remains intact by simply restarting the corresponding backoff counter BC[152,3] from its current value.

[0331] FIG. 10 b is similar to FIG. 10 a , except that an optional synchronization delay 1040 is added before the backoff counter restart 1033 for link 152 .

[0332] This optional synchronization delay reflects the fact that during the frame exchange, co-affiliated STA A2 122 is in invalid frame exchange state 1021 and may lose synchronization with the medium. In fact, STA A2 122 was unable to sense link 152 during STA A1 121's UL transmission 1025 on link 151. This optional synchronization delay therefore allows co-affiliated STA A2 122, which has switched from invalid frame exchange state 1021 back to listening operation state 1011, to resynchronize on the medium before restarting any new EDCA procedures to contend for access to link 152. In other words, in these embodiments, the restart strategy includes a synchronization delay from the end of the frame exchange before restarting the backoff counter.

[0333] Synchronization is lost if the invalid frame exchange state 1021 persists for more than a threshold, called aMediumSyncThreshold in the IEEE P802.11be / D1.5 draft. Therefore, the application of a synchronization delay in the restart strategy is contingent on a frame exchange duration greater than such a predetermined threshold.

[0334] The length of the synchronization delay is defined by the value contained in the Medium Synchronization Information field (if any) of the Basic Multi-Link element of the most recent frame received from the corresponding affiliated AP (AP2 112 in the figure).

[0335] In practice, if a loss of synchronization occurs, the restart of the backoff counter of Link 152 is frozen: a timer called MediumSyncDelay in the IEEE P802.11be / D1.5 draft and initialized with the value of the Medium Synchronization Information field is started by STA A2 122 immediately after switching back to the listening operating state 1011 (step 950). The timer counts down, and upon expiration, the backoff counter of Link 152 is started (decremented).

[0336] During the timer countdown, co-affiliated STA A2 122 may recover from the out-of-sync state, for example, by receiving any 802.11 control frame. In effect, the co-affiliated STA A2 122's radio stack is able to decode the Duration field of such a frame and thus realign with the timing of link 152. In effect, the MediumSyncDelay timer may be reset to zero upon successful reception of any type of 802.11 control frame with an MCS set up to 2 (i.e., up to 24 Mbps). Such reception reduces the synchronization delay 1040 and updates STA A2 122's NAV.

[0337] In the scenario of Figure 10b, co-affiliated STA A2 122 is considered to have resynchronized to the medium either upon successful reception of an 802.11 control frame or upon expiration of the MediumSyncDelay timer. The next decrement of the backoff counter may then be initiated.

[0338] 11 schematically illustrates an exemplary timeline for a second EMLSR or EMLMR operation case including a backoff counter restart procedure according to an embodiment of the present invention, where the triggering event for frame exchange is the receipt of an Initial frame (in EMLMR mode) or an Initial Control frame (in EMLSR mode) from the AP MLD over link 151. In this scenario, the Initial frame is received while the backoff counters driving the EDCAs of the same AC for link 151 and link 152, respectively, are being decremented.

[0339] In Figure 11, both co-affiliated STAs A1 121 and A2 122 decrement their backoff counters (step 915 above). As with Figure 10a, for purposes of illustration, few backoff counters are shown. In this example, it is assumed that traffic from AC3 is exchanged during the frame exchange, so only BC[151,3] and BC[152,3] are shown. The same teachings below apply to any AC (whose backoff counters are simultaneously decremented), known as the "transmitted AC," through which data is exchanged during the frame exchange.

[0340] While in the listening operation state 1010 and while the transmitted backoff counter of its transmitted AC is decremented 1013, co-affiliated STA A1 121 receives an Initial Control frame (IC) or Initial frame 1134 from co-affiliated AP AP1 111 via link 151. Co-affiliated STA A2 122 also decrements its backoff counter 1014.

[0341] In response to such reception, the non-AP MLD 120 pauses the decrements 1013 and 1014 of the transmitted backoff counters and the other backoff counters of the two links (steps 925, 930), initiates a state change for the co-affiliated STAs (step 935), and then transmits an Initial Control frame response (IC resp.) or Initial frame response 1135 to the co-affiliated AP AP1 111.

[0342] After the EMLSR active switch delay or the EMLMR active switch delay, co-affiliated STA A1 121 switches from the listening operation state 1010 to the valid frame exchange state 1020, and co-affiliated STA A2 122 switches from the listening operation state 1011 to the invalid frame exchange state 1021 simultaneously.

[0343] Once the switch occurs, co-affiliated STA A1 121 exchanges AC frames transmitted with co-affiliated AP AP1 111 via link 151 .

[0344] Typically, affiliated AP1 111 may transmit a basic trigger frame 1144 to affiliated STA A1 121 to allocate uplink (UL) resource units for non-AP MLD 120, as specified in IEEE Standard 802.11ax®-2021. The basic trigger frame 1144 may identify the transmission ACs on which data is permitted to be transmitted by the triggered non-AP MLD during the frame exchange. Alternatively, the triggered non-AP MLD may locally determine the ACs to be transmitted. In either case, non-AP MLD 120 transmits a High-Efficiency Trigger-Based (HE TB) PPDU 1124 in the allocated resource units.

[0345] Alternatively, affiliated AP1 111 may perform direct DL transmission (without a trigger frame) through AP1 111 via AP MLD 110 transmitting a High-Efficiency Multi-User (HE MU) PPDU 1124′ on link 151. The HE MU PPDU 1124′ may allocate resource units of link 151 to various affiliated STAs and provide a PPDU on each allocated resource unit.

[0346] In addition to completing this frame exchange, after the EMLSR Transition Delay or EMLMR Delay specified in the EML Capabilities, the non-AP MLD 120 switches back to listening operation mode (step 950), which means that co-affiliated STA A1 121 switches back to listening operation state 1010, as does the other co-affiliated STA A2 122 (listening operation state 1011).

[0347] At this stage, both co-affiliated STAs A1 121 and A2 122 may process the restart of their backoff counters. In Figure 11, backoff counter restart 1132 of the backoff counter corresponding to the transmitted AC is shown for co-affiliated STA A1 121 of link 151, and backoff counter restart 1133 of the same AC is shown for co-affiliated STA A2 122 of link 152. Backoff counter restart 1132 and 1133 correspond to backoff counters BC[151,3] and BC[152,3], respectively, in this example. Other "non-transmitted" backoff counters (i.e., not involved in the frame exchange) are restarted from their current values ​​in the conventional manner.

[0348] Referring to the table shown in Figure 13, the backoff counter restarts 1132 and 1133 for transmitted AC3 depend on the type of previous transmission (UL or DL) and, in the case of UL transmission, on the status of the upcoming UL transmission of the HE TB PPDU frame 1124. In the case of Figure 11, this is shown in the second and third rows 1302 (for UL transmission) and 1303 (for DL ​​transmission only).

[0349] The backoff counter restart 1132 strategy for BC[151,3] is shown in column 1330 .

[0350] If an UL transmission of AC3 data occurs during a frame exchange, and the previous MU UL transmission 1124 was successful, the backoff counter BC[151,3] is handled in a legacy manner: it is simply restarted from its current value, handled using the traditional EDCA backoff procedure, and may then be reinitialized (upon the first subsequent expiration) using the MU EDCA parameters from the range [0,CW=CWmin]. This is to penalize non-AP MLDs 120 that benefited from the TB uplink transmission opportunity.

[0351] If the previous MU UL transmission 1124 was not successful, the backoff counter BC[151,3] is processed in the legacy manner: it is simply restarted from its current value and processed using the conventional EDCA backoff procedure.

[0352] If only MU DL transmissions of AC3 data occur during a frame exchange, then whatever the status of the previous MU DL transmission 1124', the backoff counter BC[151,3] is handled in the conventional manner: it is simply restarted from its current value and processed using conventional EDCA backoff procedures.

[0353] The strategy for the backoff counter restart 1133 of BC[152,3] is shown in column 1340.

[0354] When an UL transmission of AC3 data occurs during a frame exchange, if the previous MU UL transmission 1124 was successful, several options are available to restart the backoff counter BC[152,3], which may be used without discrimination.

[0355] The first option involves reinitializing BC[152,3] from the range [0,CW=CW] using the current EDCA parameters and proceeding using the traditional EDCA backoff procedure. In this case, BC[152,3] is reinitialized (not simply restarted) to account for the dependency from BC[151,3] and restore some fairness. However, because no UL transmissions from STA A2 122 are occurring on link 152, the contention window CW remains the same to reflect the current state of this link.

[0356] The second option, if not already in the corresponding MU EDCA or EML EDCA mode (and therefore switching to this mode occurs), consists in reinitializing BC[152,3] from the range [0, CW = CWmin] using the MU EDCA or EML EDCA parameters (defined above) and proceeding using the conventional EDCA backoff procedure. MUEDCATimer or EMLEDCATimer can be set accordingly. In this case, BC[152,3] is reinitialized to take into account its dependency on BC[151,3] and uses the MU or EML EDCA parameters to take into account AC3 trigger-based UL transmission 1124.

[0357] The third option consists of restarting BC[152,3] from its current value, proceeding using the conventional EDCA backoff procedure, and (upon the first subsequent expiration) reinitializing using either the first option (from the range [0,CW]) or the second option (from the range [0,CWmin] using MU or EML EDCA parameters) if not already in MU or EML EDCA mode.

[0358] If the previous MU UL transmission 1124 was unsuccessful, the backoff counter BC[152,3] is restarted from its current value and processed using conventional EDCA backoff procedures. In this case, because the MU UL transmission 1124 of AC3 on link 151 failed, the priority of AC3 on link 152 is maintained by simply restarting the corresponding backoff counter BC[152,3] from its current value.

[0359] If the only MU DL transmission of AC3 data occurs during a frame exchange, regardless of the status of the previous MU DL transmission 1124', the backoff counter BC[152,3] is restarted from its current value and processed using conventional EDCA backoff procedures. In this case, because the previous MU DL transmission 1124' was a downlink transmission, the backoff counter BC[152,3] is mapped to AC3, which is associated with uplink traffic, and is restarted from its current value.

[0360] The optional synchronization delay 1040 of FIG. 10b may also be implemented in the scenario of FIG. 11 to compensate for possible loss of synchronization of co-affiliated STA A2 122 in invalid frame exchange state 1021.

[0361] 12 schematically illustrates an exemplary timeline for a third EMLSR or EMLMR operation case including a backoff restart procedure according to an embodiment of the present invention, where the triggering event for the frame exchange is the reception of an Initial frame (in EMLMR mode) or an Initial Control frame (in EMLSR mode) from the AP MLD on link 151. In this scenario, a frame exchange related to a transmitting AC occurs over link 151 after the Initial frame is received, but the backoff counter for that transmitting AC on that link is not decremented, while the backoff counter driving the EDCA for the same transmitting AC on the other link 152 is decremented.

[0362] As discussed above (step 910), the determination of which backoff counters should be decremented can be based on a variety of criteria.

[0363] The scenario of Figure 12 applies specifically when the TID-to-Link mapping prohibits an AC (AC3 in the example) on link 151, but the trigger frame sent by the AP MLD 110 in the frame exchange allows this AC in TB MU UL transmissions.

[0364] The scenario in Figure 12 also applies when the backoff counter cannot start decrementing before the Initial frame 1134 is received. This can occur when the AIFSN value governing the backoff counter operation has a high value. Incidentally, this also applies to the case of AIFSN = 0, which defines a specific MU EDCA mode in which EDCA is disabled for the AC on link 151 (AC3 in the example). Indeed, according to 802.11ax-2021, the MU AC Parameter Record field of the MU EDCA Parameter Set element may contain a value of 0 in the ACI / AIFSN field, thus indicating that EDCA is disabled for the period specified in the MU EDCA Timer for the corresponding AC.

[0365] Therefore, the backoff counter of one AC in link 2 may be decremented while the corresponding backoff counter of the same AC in link 1 is not decremented.

[0366] 12 shows that only one backoff counter (corresponding to the transmitted AC over which data is to be transmitted during a frame exchange) is decremented by co-affiliated STA A2 122 operating on link 152 (step 915 above). However, multiple transmitted backoff counters may be simultaneously decremented by co-affiliated STA A2 122 while corresponding backoff counters of the same transmitted AC by co-affiliated STA A1 121 operating on link 151 are not decremented. In this example, BC[152,3] is decremented. Similar to FIG. 10a, other backoff counters BC[152,0], BC[152,1], BC[152,2] of other ACs are also decremented simultaneously with BC[152,3], and in some cases, other backoff counters BC[151,0], BC[151,1], BC[151,2] of co-affiliated STA A1 121 (not corresponding to AC3) are also decremented.

[0367] While in the listening operating state 1010, co-affiliated STA A1 121 receives an Initial Control frame (IC) or Initial frame 1134 from co-affiliated AP AP1 111 via link 151. In parallel, co-affiliated STA A2 122 decrements its backoff counter 1014.

[0368] In response to such reception, the non-AP MLD 120 pauses the backoff counter decrement 1014 (steps 925, 930), initiates a state change for the co-affiliated STA (step 935), and then sends an Initial Control frame response (IC resp.) or Initial frame response 1135 to the co-affiliated AP AP1 111.

[0369] After the EMLSR active switch delay or the EMLMR active switch delay, co-affiliated STA A1 121 switches from the listening operation state 1010 to the valid frame exchange state 1020, and co-affiliated STA A2 122 switches from the listening operation state 1011 to the invalid frame exchange state 1021 simultaneously.

[0370] Once the switch occurs, co-affiliated STA A1 121 exchanges frames with co-affiliated AP AP1 111 via Link 151.

[0371] Typically, affiliated AP1 111 may transmit a basic trigger frame 1144 to affiliated STA A1 121 to allocate uplink (UL) resource units for the non-AP MLD 120, as specified in IEEE Standard 802.11ax®-2021. The basic trigger frame 1144 may specify the AC to be transmitted (here, AC3). Alternatively, the triggered non-AP MLD may locally determine the AC to be transmitted. In either case, the non-AP MLD 120 transmits a High-Efficiency Trigger-Based (HE TB) PPDU 1124 in the allocated resource units.

[0372] Alternatively, affiliated AP1 111 may perform direct DL transmission (without a trigger frame) through AP1 111 via AP MLD 110 transmitting a High-Efficiency Multi-User (HE MU) PPDU 1124′ on link 151. The HE MU PPDU 1124′ may allocate resource units of link 151 to various affiliated STAs and provide a PPDU on each allocated resource unit.

[0373] In addition to completing this frame exchange, after the EMLSR Transition Delay or EMLMR Delay specified in the EML Capabilities, the non-AP MLD 120 switches back to listening operation mode (step 950), which means that co-affiliated STA A1 121 switches back to listening operation state 1010, as does the other co-affiliated STA A2 122 (listening operation state 1011).

[0374] At this stage, co-affiliated STA A2 122 may process its transmitted backoff counter restart. On Figure 12, backoff counter restart 1233 corresponding to backoff counter BC[152,3] is shown for co-affiliated STA A2 122 on link 152.

[0375] As in Figure 11, the backoff counter restart 1233 for AC3 depends on the type of previous transmission (UL or DL) and, in the case of a UL transmission, on the status of the upcoming UL transmission of the HE TB PPDU frame 1124. Again, the restart strategies available in this case are shown in Figure 13 in the box whose last column 1340 spans the second and third rows 1302 (for UL transmissions) and 1303 (for DL ​​transmissions only). The restart strategies will not be repeated here for reasons of brevity.

[0376] FIG. 13 shows a table 1300 that collects proposed EDCA backoff counter restart procedures or policies according to an embodiment of the present invention.

[0377] Column 1310 collects exemplary trigger events that may suspend the decrementing of the back-off counter. Column 1320 indicates possible uplink transmission statuses for each event. Column 1330 collects, for each event and each transmission status, one or more back-off counter restart strategies to apply to the transmitted AC (e.g., BC[151,3] in the above example). Column 1340 collects, for each event and each transmission status, one or more back-off counter restart strategies to apply to the transmitted AC (e.g., BC[152,3] in the above example).

[0378] Row 1301 collects the backoff counter restart strategy to apply to the transmitted AC in the case of a SU UL transmission (e.g., when the trigger event for pausing the backoff counter decrement is the expiration of one backoff counter contention access to link 151).

[0379] Row 1302 collects the backoff counter restart strategy to apply to the transmitted AC in case of TB MU UL transmission (e.g., when the trigger event for pausing the backoff counter decrement is the reception of an Initial frame or Initial Control frame on link 151 followed by trigger-based uplink traffic).

[0380] Row 1303 collects the backoff counter restart strategy to apply to the transmitted AC in the case of MU DL transmission (e.g., when the trigger event for pausing the backoff counter decrement is the reception of an Initial frame or Initial Control frame on link 151 followed by downlink traffic).

[0381] Figure 14 shows a schematic diagram of an EMLSR-enabled architecture for an MLD. This diagram illustrates the case where two affiliated non-AP STAs share the hardware resources of a non-AP MLD when EMLSR mode is enabled. The EMLSR-enabled architecture for an MLD shown in this diagram is for illustrative purposes only, and other alternative architectures are also possible.

[0382] The architecture includes two wireless stacks: a light wireless stack and a full wireless stack.

[0383] The complete wireless stack includes a complete 802.11be MAC module 1400a (which exchanges data with upper layers), a complete 802.11be PHY module 1405a connected to the complete MAC module, a complete radio frequency chain 1415a connected to the complete PHY module, and an antenna 1420a connected to the complete RF chain via the EMLSR switch 1410.

[0384] The light wireless stack includes a light 802.11be MAC module 1400b (which exchanges data with upper layers), a light 802.11be PHY module 1405b connected to the light MAC module, a light radio frequency chain 1415b connected to the light PHY module, and an antenna 1420b connected to the light RF chain via an EMLSR switch 1410.

[0385] The EMLSR switch 1410 is shared by the two wireless stacks and is configured to switch the EMLSR co-affiliated STAs from / to a listening operation state to / from a frame exchange enabled or disabled state when the EMLSR mode is activated.

[0386] Radio chains 1400a, 1405a, and 1415a are complete radio resources that allow the reception and transmission of any IEEE 802.11 frame. Specifically, they include encoding and decoding modules for encoding and decoding any IEEE 802.11 frame. Radio chains 1400b, 1405b, and 1415b are reduced-function (or "lite") radio resources that allow the reception and transmission of only specific IEEE 802.11 frames. Specifically, they include only encoding and decoding modules for encoding and decoding specific frames using rates of 6 Mbps, 12 Mbps, or 24 Mbps.

[0387] The diagram on the bottom left shows the functionality of the non-AP MLD when it is in EMLMR listening mode of operation: A common EMLMR switch 1410 connects each radio chain 1400a / 1405a / 1415a and 1400b / 1405b / 1415b to antennas 1420a and 1420b, respectively. Thus, each radio stack can be used to listen to each link simultaneously. As shown in the diagram, two links are available. The complete radio chain 1400a / 1405a / 1415a and antenna 1420a are configured to operate on link 1, and the light radio chain 1400b / 1405b / 1415b and antenna 1420b are configured to operate on link 2.

[0388] The bottom-center diagram illustrates the functionality of the non-AP MLD when switching to the first EMLSR frame exchange mode. The EMLSR's co-affiliated STAs corresponding to link 1 are in the active frame exchange state, while the other EMLSR's co-affiliated STAs corresponding to link 2 are in the inactive frame exchange state. In this case, the EMLSR shared switch 1410 connects the complete radio chain 1400a / 1405a / 1415a to both antennas 1420a / 1420b, and the complete radio chain 1400a / 1405a / 1415a and antennas 1420a / 1420b are configured to operate on Link 1. Here, because the complete radio chain remains configured to operate on Link 1, the switching time from the EMLSR's listening operation state to the active frame exchange state can be considered short. In fact, in this case, the only switching is an antenna switch. Meanwhile, the common EMLSR switch 1410 disconnects the light radio chain 1400b / 1405b / 1415b from the antenna 1420b. In this configuration, the light radio chain 1400b / 1405b / 1415b cannot receive or transmit frames on Link 2, and only Link 1 is available.

[0389] The bottom right diagram illustrates the functionality of the non-AP MLD when switching in a second EMLSR frame exchange mode. The co-affiliated STAs of the EMLSR corresponding to link 2 are in an active frame exchange state, while the co-affiliated STAs of the other EMLSR corresponding to link 1 are in an inactive frame exchange state. In this case, the shared switch 1410 of the EMLSR connects the complete radio chain 1400a / 1405a / 1415a to both antennas 1420a / 1420b, and the complete radio chain 1400a / 1405a / 1415a and antennas 1420a / 1420b are configured to operate on Link 2. Here, the switching time from the EMLSR's listening operation state to the active frame exchange state is expected to be long because the complete radio chain is switched to operate on Link 2. In fact, in this case, the switching includes both an antenna switch and a complete radio chain configuration switch. Meanwhile, the common EMLSR switch 1410 disconnects the light wireless chain 1400b / 1405b / 1415b from the antenna 1420b. In this configuration, the light wireless chain 1400b / 1405b / 1415b cannot receive or transmit frames on Link 1, and only Link 2 is available.

[0390] The functionality of the common EMLSR switch 1410 clearly indicates that state changes of co-affiliated STAs of two EMLSRs in the same MLD are necessarily simultaneous, as antenna resources are either connected to one STA or the other, but are never available to both STAs at the same time.

[0391] Figure 15 shows a schematic diagram of the EMLMR-enabled architecture of MLD, taking the example of two affiliated non-AP STAs sharing antenna resources when EMLMR mode is enabled.

[0392] This architecture includes two wireless stacks, one for each non-AP STA.

[0393] The radio stack includes a complete 802.11be MAC module 1500a or 1500b (which exchanges data with upper layers), a complete 802.11be PHY module 1505a or 1505b connected to the MAC module, a radio frequency chain 1515a or 1515b connected to the PHY module, an EMLMR switch 1510 shared by the two radio stacks and configured to perform antenna resource aggregation when EMLMR mode is activated, and an antenna array 1520a or 1520b.

[0394] The diagram on the bottom left shows the functionality when non-AP MLD is listening for an Initial frame: A common EMLMR switch 1510 connects each antenna array to an RF chain. Each radio stack is therefore complete and may serve its respective link using, for example, a 2x2 MIMO antenna configuration. As shown in the diagram, two links are available.

[0395] The bottom-center diagram illustrates the functionality of non-AP MLD when switching in the first EMLMR frame exchange mode. The EMLMR co-affiliated STAs corresponding to link 2 are in an enabled frame exchange state, while the other EMLMR co-affiliated STAs corresponding to link 1 are in an disabled frame exchange state. The common EMLMR switch 1510 aggregates antenna resources to link 2 by connecting the antenna array 1520a of the second wireless stack to the RF chain 1515b of the first wireless stack. Thus, the first wireless stack operates in a 4x4 MIMO antenna configuration, improving throughput for link 2. However, link 1 becomes unavailable because its antenna array 1520a is no longer available to the second wireless stack.

[0396] The bottom right diagram illustrates the functionality of non-AP MLD when switching to a second EMLMR frame exchange mode. The EMLMR co-affiliated STAs corresponding to link 1 are in an enabled frame exchange state, while the other EMLMR co-affiliated STAs corresponding to link 2 are in an disabled frame exchange state. The common EMLMR switch 1510 aggregates antenna resources to link 1 by connecting the antenna array 1520b of the first wireless stack to the RF chain 1515a of the second wireless stack. Therefore, the second wireless stack operates in a 4x4 MIMO antenna configuration, improving throughput for link 1. Meanwhile, link 2 becomes unavailable because its antenna array 1520b is no longer available to the first wireless stack.

[0397] The functionality of the common EMLMR switch 1510 clearly shows that state changes of two EMLMR co-affiliated STAs in the same MLD are necessarily simultaneous, since antenna resources are either connected to one STA or the other, but not both STAs, at the same time.

[0398] 16 shows a schematic diagram of a communication device 1600, typically one of the MLDs described above, of a wireless network, configured to implement at least one embodiment of the present invention. The communication device 1600 may preferably be a device such as a microcomputer, a workstation or a lightweight handheld device. The communication device 1600 preferably includes a communication bus 1613 connected to: a central processing unit 1601 such as a processor, denoted as CPU; a memory 1603 for storing executable code of a method or method steps according to an embodiment of the present invention and registers adapted to record variables and parameters necessary for the implementation of the method; and At least two communication interfaces 1602 and 1602' connected to a wireless communication network (eg, a communication network according to one of the IEEE 802.11 family of standards) via transmit and receive antennas 1604 and 1604', respectively.

[0399] Preferably, a communications bus 1613 provides for communication and interoperability between various elements included in or connected to communications device 1600. The representation of a bus is not limiting, and in particular a central processing unit is operable to communicate instructions to any element of communications device 1600 directly or by way of another element of communications device 1600.

[0400] The executable code may be stored in a memory, either read-only, on a hard disk, or on a removable digital medium, such as a disk. According to an optional variant, the executable code of the program may be received by the communication network, via the interface 1602 or 1602', so as to be stored in the memory of the communication device 1600 before being executed.

[0401] In one embodiment, the device is a programmable apparatus that uses software to implement embodiments of the invention, however, embodiments of the invention may alternatively be implemented in whole or in part in hardware (e.g., in the form of an Application Specific Integrated Circuit or ASIC).

[0402] Although the present invention has been described with reference to particular embodiments, it is not limited to those embodiments, and modifications within the scope of the invention will be apparent to those skilled in the art.

[0403] Many further modifications and variations will be suggested to those skilled in the art upon reference to the exemplary embodiments described above, but these embodiments are given by way of example only and are not intended to limit the scope of the invention, which is determined solely by the appended claims. In particular, different features from different embodiments may be interchanged where appropriate.

[0404] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite articles "a" or "an" do not exclude a plurality. The mere fact that different features are recited in mutually different dependent claims does not indicate that a combination of these features cannot be used to advantage.

Claims

1. 1. A method of communication in a wireless network, comprising: a non-access point (non-AP) multilink device (MLD) operating in an active enhanced multilink (EML) mode; Initiating an Enhanced Distributed Channel Access (EDCA) backoff procedure by a first affiliated station (STA) that is affiliated to the non-AP MLD and corresponds to a first link of a set of valid links to which the EML mode is applied, the STA decrementing a backoff counter to access the first link; switching the first affiliated STA from a listening operation state to an active frame exchange state to initiate a frame exchange with an AP MLD over the first link; and pausing the EDCA backoff procedure in the first affiliated STA upon receiving an Initial frame from the AP MLD via a second link of the set of valid links to which the EML mode is applied. method.

2. switching the first affiliated STA to the valid frame exchange state is performed before the first affiliated STA in the valid frame exchange state initiates the EDCA backoff procedure to decrement the backoff counter. The method of claim 1.

3. pausing the decrementation upon detecting that the first link becomes busy during the decrementation of the backoff counter; a) maintaining the first affiliated STA in the active frame exchange state and resuming the decrement when the first link becomes idle again; b) switching the first affiliated STA back to a listening state regardless of the period specified in the frame in which the detection occurred; c) switching the first affiliated STA back to the listening operation state for a predetermined period based on the period identified in the frame in which the detection occurred, and then switching back to an active frame exchange state to resume the decrement when the first link becomes idle again; d) determining a period of time identified in the frame in which the detection occurred, and deciding to apply one of policies a), b), or c) depending on the determined period of time; applying a policy from The method of claim 2.

4. When the backoff counter reaches a value of 0, the first affiliated STA switches to the active frame exchange state. The method of claim 1.

5. switching the first affiliated STA includes initiating a state switch procedure for the first affiliated STA while the first affiliated STA decrements the back-off counter, and terminating the state switch procedure when the back-off counter reaches a value of 0. The method of claim 1.

6. Switching the first affiliated STA includes initiating a state switching procedure for the first affiliated STA in response to the backoff counter reaching a value of zero. The method of claim 1.

7. and transmitting a control frame over the first link in response to the backoff counter reaching a value of zero. The method of claim 6.

8. The control frame is a CTS-to-self frame or an RTS frame. The method of claim 7.

9. The control frame includes padding to terminate the control frame after a short interframe space (SIFS) preceding the end of the state switch procedure. The method of claim 7.

10. and setting a network allocation vector (NAV) of the first affiliated STA in the listening state upon sensing a control frame having an MCS value up to 2 on the first link. The method of claim 1.

11. and simultaneously initiating an EDCA backoff procedure on two or more links of the set, the first link being the link corresponding to the link on which a backoff counter of the EDCA backoff procedure first reaches a value of zero. The method of claim 1.

12. and switching a separate second affiliated STA, affiliated with the non-AP MLD and corresponding to a second link of the set, from a listening operation state to an invalid frame exchange state concurrently with the switching of the first affiliated STA. The method of claim 1.

13. and switching the first affiliated STA and the second affiliated STA to the listening operation state upon completion of the frame exchange over the first link. The method of claim 12.

14. Initiating the frame exchange excludes sending an Initial frame to the AP MLD via the first link. The method according to claim 1 or 3.

15. and exchanging frames with the AP MLD via the first link after the backoff counter reaches a value of 0. The method of claim 1.

16. and suspending the back-off counter driving the EDCA of the first link of the set in response to initiating a frame exchange with the AP MLD via a second link of the set of valid links to which the EML mode is applied. The method of claim 1.

17. and, in response to a completion of the frame exchange over the second link, restarting the back-off counter using a restart strategy selected based on characteristics of the frame exchange.

17. The method of claim 16.

18. The characteristics of the frame exchange include whether the frame exchange includes a single-user uplink transmission to the AP MLD, a multi-user trigger-based uplink transmission to the AP MLD, or only a downlink transmission from the AP MLD.

18. The method of claim 17.

19. The characteristics of the frame exchange include whether the frame exchange is a successful or unsuccessful uplink transmission to the AP MLD.

18. The method of claim 17.

20. The restart strategy applied is: reinitializing the backoff counter using the current contention window before beginning to decrement the backoff counter; reinitializing the backoff counter using a new contention window associated with a new EDCA mode before beginning to decrement the backoff counter; restarting the backoff counter from its current value; Contains one from 18. The method of claim 17.

21. The restart strategy applied is: restarting the back-off counter from its current value in the case of a downlink transmission or a failed uplink transmission in the frame exchange; reinitializing a back-off counter using a current contention window before restarting the back-off counter from its current value if a single-user uplink transmission is successful in the frame exchange before starting to decrement the back-off counter or before restarting the back-off counter from its current value; if a multi-user trigger-based uplink transmission in the frame exchange is successful, reinitializing the back-off counter using the current contention window or a new contention window associated with a new EDCA mode before starting to decrement the back-off counter or restarting the back-off counter from its current value; Contains one from 18. The method of claim 17.

22. A non-access point (non-AP) multi-link device (MLD) capable of operating in an active enhanced multi-link (EML) mode, comprising: initiating means for initiating an Enhanced Distributed Channel Access (EDCA) backoff procedure by a first affiliated station (STA) that is affiliated to the non-AP MLD and corresponds to a first link of a set of valid links for which the EML mode is applicable, the STA decrementing a backoff counter to access the first link; switching means for switching the first affiliated STA from a listening operation state to an active frame exchange state to initiate a frame exchange with an AP MLD over the first link; and stopping means for suspending the EDCA backoff procedure in the first affiliated STA when an Initial frame is received from the AP MLD via a second link of the set of valid links to which the EML mode is applied. Non-AP MLD.

Images