Impactless hand-over in a wireless time-sensitive network

Abstract

Example embodiments relate to a method for performing a hand-over of a travelling end node associated to an initial access point in a wireless network. The wireless network comprises access points that operate according to a wireless time-sensitive networking mechanism by having allocated schedules with time slots for exchanging data with respective end nodes. The wireless network is managed by a network controller configured to allocate the schedules, and to perform the steps of determining one or more candidate access points; selecting a target access point; allocating a target time slot for exchanging data with the travelling end node within the schedule of the target access point avoiding time slots of the target access point that are allocated to other end nodes; determining a hand-over time; and providing to the travelling end node the hand-over time, an identifier of the target access point, and the target time slot.

Description

IMPACTLESS HAND-OVER IN A WIRELESS TIME-SENSITIVE NETWORKField of the Invention

[0001] The present invention generally relates to performing a hand-over of an end node in a wireless time-sensitive network.Background of the Invention

[0002] Time-Sensitive Networking, TSN, refers to a group of networking protocols and standards developed by the TSN task group of the IEEE 802.1 working group to provide accurate time synchronisation, data delivery with bounded latency, and zero congestion loss in Local Area Networks, LANs. In TSN, the nodes in a network are time synchronised and are assigned a respective time slot for exchanging data. As such, TSN can enable applications which require reliable and accurately timed data delivery such as, for example, devices in smart factories, devices connected to the Industrial Internet of Things, HoT, mobile and collaborative robots, autonomous systems, audio and video distribution systems, and virtual or mixed reality applications. Wireless Time-sensitive Networking, W-TSN, typically refers to a wireless network that extends TSN capabilities over wireless media, e.g. Wi-Fi. Such a network usually comprises a wired portion and a wireless portion. In the wireless portion, end nodes, sometimes also referred to as clients or end stations, which are wirelessly serviced by access points that are connected to the wired portion of the network via switches. As such, the access points form the interface between the wired and wireless portions of the network.

[0003] A network controller, e.g. comprising a Centralized Network Controller, CNC, and a Centralized User Configuration, CUC, controls the network by managing schedules, also referred to as time schedules, of the access points, end devices and switches in the wired portion. Each access point operates according to a respective schedule, which is a collection of successive time slots that are allocated by the network controller. Each time slot is reserved for a specific task, e.g. exchanging data with a specific end node. A schedule is repeated in time by thecorresponding access point until the network controller adjusts it. By assigning time slots to respective end nodes in the network, time sharing of access point resources is realized. By scheduling the time slots of the access points within the network, the network controller ensures that the network complies with the TSN requirements.

[0004] A wireless network may have mobile end nodes that can pass through regions serviced by different access points. During travelling, a mobile end node may thus disconnect from a first access point servicing a first region, and connect to a second access point servicing a second region. As such, the travelling end node is handed over from the first access point to the second access point. Such a handover is sometimes referred to as a hand-over protocol, or a hand-over process. The related first access point, i.e. the access point to which the end node is associated before the hand-over, is also referred to as the initial access point. The second access point, i.e. the access point to which the end node is associated after the hand-over, is also referred to as the target access point. The initial and target access points typically operate on different communication channels. A hand-over is initiated upon identifying a need for an end node to be serviced by another access point, for example based on degrading signal quality of the communication link with the initial access point. The end node then identifies a target access point by actively scanning the environment, e.g. by sending out probe requests and receiving responses by other access points, or by listening to beacons transmitted by other access points. Based on the scanning, the end node can select an access point. The end node can then associate with the selected target access point, thereby completing the hand-over.

[0005] The active scanning can introduce interference in the communication or data exchange between other access points and their respectively associated end nodes, which should be avoided in W-TSN as it can affect the reliability and timing of data delivery. Additionally, when end nodes are only provided with one radio for communicating with the W-TSN network, the data exchange between the end node and its initial access point is interrupted during scanning. The communication can suffer an undesirable delay due to this interruption, further violating the restrictions of the W-TSN network.

[0006] It is thus a problem to hand over an end node within a W-TSN network without impacting the data exchange within the network, i.e. without traffic collisions between the travelling end node and other end nodes within the network, and without causing delays that obstruct the time synchronisation within the network.Summary of the Invention

[0007] It is an object of the present invention, amongst others, to solve or alleviate the above identified challenges and problems by providing impactless handing-over of an end node from an initial access point to a target access point in a wireless time-sensitive network.

[0008] According to a first aspect, this object is achieved by a method for performing a hand-over of a travelling end node associated to an initial access point in a wireless network. The wireless network comprises access points that operate according to a wireless time-sensitive networking, W-TSN, mechanism by having allocated schedules with time slots for exchanging data with respective end nodes. The wireless network is managed by a network controller configured to allocate the schedules. Further, the network controller is configured to perform a step of determining one or more candidate access points that are within range for exchanging data with the travelling end node. The network controller is further configured to perform a step of selecting a target access point from the one or more candidate access points. The network controller is further configured to perform a step of allocating a target time slot for exchanging data with the travelling end node within the schedule of the target access point avoiding time slots of the target access point that are allocated to other end nodes. The network controller is further configured to perform a step of determining a hand-over time for initiating the handover from the initial access point to the target access point. The network controller is further configured to perform a step of providing to the travelling end node: the hand-over time, an identifier of the target access point, and the target time slot, for performing the hand-over.

[0009] When a travelling end node travels outside of a region serviced by the initial access point, there is a need to hand over the travelling end node to another accesspoint in order to avoid the travelling end node losing connection to the network. In order to perform a hand-over, it needs to be decided to which access point the travelling end node will be handed over. To that end, the network controller determines one or more candidate access points, i.e. access points that are within the possibility to service the travelling end node. The one or more candidate access points form a subset of all the access points within the network. Whether an access point can be considered within the possibility to service the travelling end node, may depend on a proximity of the travelling end node to the access point. It may also depend on a physical environment, e.g. a surroundings or objects obstructing wireless communication, of the region serviced by the access point. Thus, determining the one or more candidate access points is based on whether or not an access point is within range of the end node, i.e. whether a wireless communication link of sufficient quality can be established between the end node and the access point. A target access point is selected from the subset of one or more candidate access points by the network controller. The network controller enables data exchange without interference between the travelling end node and the target access point after the hand-over by allocating a target time slot within the schedule of the target access point for exchanging data with the travelling end node. The hand-over time is the moment when the communication with the target access point can be established by the end node. In other words, from that moment on the end node can switch to the channel of the target access point and start communicating with the target access point.

[0010] The network controller prepares the hand-over from the initial access point to the target access point by allocating the target time slot without violating the schedules of the access points within the network. This allows performing the handover of a travelling node without interfering with data exchanged between other end nodes and access points in the network. The data exchange between the target access point and other end nodes remains unaffected by avoiding time slots in the schedule of the target access point allocated to those other end nodes. It is a further advantage that the data exchange between the initial access point and other end nodes also remains unaffected. Therefore, the network still complies with the W- TSN requirements and no data is lost in the network. Further, as the determining of the candidate access points and the selecting of the target access point is performedby the network controller, the end node is relieved from these tasks and can thus keep exchanging data with the initial access point. This further allows maintaining an unaffected data exchange between the initial access point and the travelling end node while preparing the hand-over. The method for performing a hand-over results in a faster hand-over since the end node or the initial access point do not need to spend time on deciding on a target access point. This efficient implementation fits the W-TSN requirements seamlessly since data loss due to tardiness can be avoided as a result of the end node or initial access point having to choose a target access point. It is an advantage that the method contributes to optimally deploying the resources of the network as the method enables performing a hand-over without violating W-TSN requirements. The method allows preparing the hand-over within the same channel as used for data exchange, i.e. by time-interleaving the hand-over preparation with the data exchange. This is especially advantageous when end nodes are only configured for communicating on a single channel or frequency at a given moment in time. When end nodes are only equipped with a single radio, comprising a single receiver antenna and a single transmitter antenna, the end node is restricted to use the radio for data exchanging as well as for performing a handover. It is a further advantage that the method is compatible with existing Wi-Fi association procedures.

[0011] According to an embodiment, the access points of the network operate in different channels, wherein the identifier specifies a channel of the target access point, and wherein the performing of the hand-over includes, by the travelling end node, switching to the channel of the target access point.

[0012] To avoid interference between data exchange of neighbouring access points, the access points in a wireless network typically operate at different frequencies, i.e. on different channels. When the travelling end node is handed over from the initial access point to the target access point, it needs to know what frequency to switch to in order to communicate on the channel that corresponds to the target access point. Therefore, the provided identifier may comprise information that enables the travelling end node to determine the channel to switch to for communicating with the target access point.

[0013] By providing the identifier to the travelling end node, a switching, by the travelling end node, to the correct channel can be ensured for performing the hand over.

[0014] According to a further embodiment, determining the candidate access points comprises allocating, in schedules of one or more access points, a scan time slot for scanning, by the travelling end node, for beacons transmitted by one or more access points.

[0015] Beacons are messages, broadcasted by access points, that contain information enabling a prospective end node to associate to a respective access point. During the scan time slot, the travelling end node can switch to another channel and listen for beacons of access points operating on that channel. The scan time slot may be allocated in the schedules of all the access points of the network, or may be allocated in the schedules of only a subset of the access points of the network. The scan time slot may be common for the one or more access points, or, in other words, the transmitting of beacons may be done by the one or more access points during a coinciding time slot, i.e. the scan time slot.

[0016] In classical approaches, when a hand-over is imminent, the travelling end node scans for other access points to determine a target access point. When the travelling end node can only make use of one radio, this approach requires the travelling end node to interrupt its communication of exchanging data with the initial access point to perform the scanning. As a result, the time-sensitivity of the network may be violated. By scheduling the scanning process into the schedule of the initial access point, the initial access point will not send data to the travelling end node when the travelling end node is scanning for other access points. Hence, no data will be lost. By allocating a scan time slot in the schedules of the access points, the access points can broadcast beacons for the travelling end node, at a time that the travelling end node is sure to be scanning. As a result, the scanning is performed very efficiently. In addition, during the scan time slot, other travelling end nodes in the network can also scan for access points, further adding to the efficiency of the scanning process.

[0017] According to a further embodiment, determining the candidate access points comprises allocating, in schedules of one or more access points, a common probe time slot shared among the one or more access points for sending, by the travelling end node, probe requests.

[0018] A probe request is a message sent out by a prospective end node that has the purpose of identifying access points, e.g.. during active scanning. During the probe time slot, the travelling end node sends a probe request on the channels of access points other than the initial access point. The probe time slot may be allocated in the schedules of all the access points of the network, or may be allocated in the schedules of only a subset of the access points of the network. The transmitting of the probe requests is done by the travelling end node during a time slot that is coinciding in the schedules of the one or more access points.

[0019] During active scanning, a probe request may interfere with data exchange between other access points and end nodes they are servicing. By allocating a common probe time slot, probe requests cannot interference with ongoing data exchange between access points and end nodes they are servicing. As a result, no data is lost and the time-sensitivity of the network is preserved.

[0020] According to a further embodiment, the method further comprises, by the one or more access points, upon receiving a probe request from the travelling end node, sending a probe response to the travelling end node, either directly or via the initial access point, or to the network controller.

[0021] The sending of the probe response may be done directly, i.e. wirelessly over the channel the respective access point is operating on. Alternatively, the sending of the probe response may be done via the initial access point. In other words, the respective access point will send the probe response to the initial access point via the wired portion of the network. Upon receiving the probe response via the wired portion of the network, the initial access point will then wirelessly send the probe response to the travelling end node over the channel the initial access point operates on. Alternative, the respective access point may send the probe response directly to the central network controller.

[0022] By sending a probe response in response to a probe request, the identifying of candidate access points is expedited significantly, as the travelling end node can expect a reply shortly after sending the probe request. The initial access point can disregard a probe request from the travelling end node, but may send a probe response upon receiving a probe request received from a prospective end node that is associated to another access point. When sending the probe response to the end node via the initial access point or directly to the central network controller, the travelling end node can continue to communicate with the initial access point, instead of listening on other channels to receive probe responses from other access points.

[0023] According to further example embodiments, determining the candidate access points comprises intercepting data exchanged between access points and other end nodes in the same channel in which the travelling end node exchanges data with the initial access point.

[0024] In wireless networks, access points may be divided in groups that operate on the same channel, i.e. frequency reuse may be applied. In such a set-up, the travelling end node can listen to communication, e.g. data exchange, of access points that operate on the same channel as the initial access point. From the intercepted data, the travelling end node can derive useful information, e.g. signal quality, for determining the relative position of the travelling end node with respect to multiple access points operating on the same channel.

[0025] By intercepting data exchanged on the same channel, the travelling end node can gather information considering candidate access points without having to switch to another channel. The data exchange between the initial access point and other end nodes can continue as before without being interrupted during the intercepting. In addition, the data exchange between the initial access point and the travelling end node can also continue as before, because the travelling end node can perform the intercepting during idle times of the data exchange with the initial access point. As such, this technique further adds to efficiently performing the step of determining the candidate access points, without compromising the constraints imposed by the W-TSN mechanism.

[0026] According to further embodiments, determining the one or more candidate access points comprises determining a location of the travelling end node.

[0027] Based on the location of the travelling end node, the network controller can determine how near access points of the network are to the travelling end node. The proximity of an access point to the travelling end node influences the quality of the data exchange between the travelling end node and the access point. Access points that are close to the travelling end node may be considered as candidate access points. The determined location may be an absolute location, e.g. a coordinate, or may be a relative location, e.g. a distance between the travelling end node and an access point.

[0028] According to further example embodiments, determining the one or more candidate access points further comprises determining non-overlapping regions within an environment covered by the wireless network, each of the regions respectively associated with one of the access points. Such regions may, for example, have polygon shapes, e.g. centred around respective access points. Selecting the target access point may comprise determining in which region the travelling end node resides based on the location of the travelling end node. The target access point may then be determined as the access point associated with that region. By determining non-overlapping regions, it is guaranteed that a single target access point can directly be selected. Further, a union of the regions may comprise the entire environment. Alternatively, according to further example embodiments, a buffer zone may be provided at a boundary between nonoverlapping regions, wherein the buffer zone is a region disassociated from all of the access points of the wireless network. The wireless network may, in such a case, be configured to only perform a hand-over of the travelling end node upon entering of a new region. As such, ping-ponging between access nodes by end nodes travelling along borders between the regions is avoided. This may reduce a number of hand-overs performed and may therefore increase the efficiency of the wireless network.

[0029] Taking into account the location of the travelling end node simplifies the determining of suitable candidate access points by the network controller. This limitsoverhead resulting from scanning, probing, intercepting communication or other techniques. However, a combination of localisation and other approaches may further contribute to determining candidate access points. Without fully relying on these other approaches, using the location of the travelling end node still significantly reduces the overhead.

[0030] According to further example embodiments, determining the location of the travelling end node comprises receiving the location from a localisation network distinct from the wireless network.

[0031] The localisation network is a separate network, distinct from the wireless network, and may for example be an Ultra-Wideband, UWB, network. The localisation network has its own resources, different from the resources of the wireless network. The location is received by a connection with the localisation network via the travelling end node. The end node then forwards the location to the network controller via the initial access point. Alternatively, the network controller directly communicates with the localisation network to retrieve the location of the travelling end node.

[0032] By receiving the location from a localisation network distinct from the wireless network, no resources of the wireless network are preserved for determining the location of the travelling end node. The overhead is limited to communicating with the localisation network to receive the location. The network is thus minimally affected by the use of the localisation network and is still able to comply with the required time-sensitivity.

[0033] According to further example embodiments, determining the location of the travelling end node comprises allocating, in schedules of one or more access points, a common localisation time slot shared among access points for localizing the travelling end node with respect to one or more access points.

[0034] During the localisation time slot, the travelling end node communicates with the one or more access points. Based on that communication, e.g. based on time of flight, the one or more access points can determine the location of the travelling end node, or can determine how far away the respective access point is locatedfrom the travelling end node. The network controller, knowing the location of each access point in the network, can determine the candidate access points based on the information regarding the location of the travelling end node, as provided by the one or more access points.

[0035] By allocating a localisation time slot without interfering with other time slots, the network controller ensures maintaining of the time-sensitivity of the network. By making the localisation time slot common for the one or more access points, more than one access point can make use of the localisation time slot, which expedites the determining of the candidate access points.

[0036] According to further embodiments, determining the one or more candidate access points is further based on historical telemetry information between end nodes and access points.

[0037] Telemetry refers to the recording of information indicative for the quality of the communication between access points in the network and end nodes they are servicing. Advantageously, this includes localisation information of the end nodes. At certain locations of end nodes, a recording of the communication quality is stored. As such, the signal quality at a number of locations within the area serviced by the network can be determined for different access points of the network. A map of signal quality may be kept for every access point. At the position of the travelling end node, a measure for signal quality based on the historical telemetry information can then be compared for different access points. Such telemetry information may, for example, comprise Received Signal Strength Indicator, RRSI, values.

[0038] In general, the quality of communication depends on a distance between an end node and the access point servicing it. However, it may happen that a zone exists where inadequate signal quality is prevalent, even though the zone is very near an access point. This may, for example, happen due to the landscape or obstacles between the end node and the access point. Because obstacles can be present temporarily, it is relevant to include recent historical telemetry information. By examining the historical telemetry information, access points that have a zone of bad signal quality where the travelling end node is located, can be disregarded when determining the candidate access points.

[0039] Thus, proximity to an access point may give an initial indication of signal strength. In addition, the signal strength with respect to a particular access point is also influenced by the physical environment. Combining localisation information with telemetry information may further increase the efficiency of the hand-over method. The thereby reduced hand-over overhead may result in an increased throughput and a reduction of jitter, i.e. time delay variation.

[0040] This may be achieved by obtaining, for each of the access points, a model indicating the signal strength, e.g. RSSI value, across various locations within the environment. This may, for example, be done by measuring RSSI values with respect to different access points at different locations of the environment. For example, end nodes may be configured to periodically capture such location-RRSI information pairs. The signal strength models may be adaptive, i.e. updated as end nodes keep capturing such information pairs. Such a signal strength model may, for example, be a machine learning, ML, model, e.g. an artificial neural network, ANN, trained to predict RSSI with respect to a particular access point based on location. Upon obtaining a signal strength model for multiple access points, they may be used to obtain a single environment map. Each location of the environment map can then be associated to one of the access points, being the access point for which the corresponding signal strength model gives the best RSSI. As such, the environment map may determine non-overlapping regions, each of the regions respectively associated with one of the access points.

[0041] Based on signal strength information linked to location, not only an optimal hand-over time but also an optimal hand-over location may as such be determined.

[0042] According to example embodiments, the method further comprises associating the travelling end node to the target access point by the network controller.

[0043] The network controller pre-associates the travelling end node to the target access point by performing the associating on behalf of the end node. The network controller provides information to the target access point for establishing a communication link with the travelling end node. This is done via the wired portion of the network.

[0044] By conducting a pre-association, the network controller reduces the hand-over overhead between the travelling end node and the target access point.

[0045] According to example embodiments, the method further comprises authenticating the travelling end node to the target access point by the network controller.

[0046] The network controller pre-authenticates the travelling end node to the target access point by performing the authenticating on behalf of the end node. The network controller provides information to the target access point for securing the communication link with the travelling end node. This is done via the wired portion of the network.

[0047] By conducting a pre-authentication, the network controller reduces the handover overhead between the travelling end node and the target access point.

[0048] According to example embodiments, the method further comprises determining, by the travelling end node and / or by the network controller, if the handover was successful. Further, the method comprises, upon determining that the hand-over was successful: updating the schedule of the initial access point by releasing the time slot allocated to the end node; and disassociating and deauthenticating the travelling end node from the initial access point. The updating is performed by the network controller. The de-authenticating and disassociating are performed by the travelling end node and / or by the network controller.

[0049] When a communication link is established between the travelling end node and the target access point, the hand-over was successful and the travelling end node can use the target time slot to exchange data with the target access point. After the hand-over, there remains a time slot reserved in the schedule of the initial access point, i.e. the time slot for exchanging data between the travelling end node and the initial access point. This time slot can now be freed for servicing of other end nodes by the initial access point.

[0050] By releasing the time slot reserved for exchanging data with the end node, and by disassociating and de-authenticating the travelling end node, the initial access point can dedicate vacant resources to other end nodes.

[0051] According to example embodiments, the method further comprises determining, by the travelling end node and / or by the network controller, if the handover was unsuccessful. Further, the method comprises, upon determining that the hand-over was unsuccessful: reverting, by the travelling end node, to the initial access point; updating the schedule of the target access point by releasing the time slot allocated to the travelling end node; and de-authenticating and disassociating the travelling end node from the target access point. The updating is performed by the network controller. The de-authenticating and disassociating are performed by the travelling end node and / or by the network controller.

[0052] Establishing a communication link between the travelling end node and the target access point may fail. To avoid the travelling end node becoming disconnected from the network, the travelling end node reverts back to the initial access point. The failing may be determined when after a predetermined time, no communication link between the travelling end node and the target access point is detected.

[0053] By reverting back to the initial access point, the travelling end node remains connected to the network. By releasing the target time slot, and by disassociating and de-authenticating the travelling end node, the target access point can dedicate vacant resources to other end nodes.

[0054] According to further example embodiments, the access points have respective boundary clocks, and the method further comprises, by the travelling end node, synchronising with a boundary clock of the target access point.

[0055] To ensure the time-sensitivity of the network, end nodes serviced by an access point synchronise with the network in order to operate according to the schedule of the respective access point. This may be done by use of a transparent clock or boundary clocks. When boundary clocks are used, each access point may have its own boundary clock. In that case, the end nodes serviced by an access point needto synchronise with the corresponding boundary clock. Thus, the travelling end node needs to synchronise with the boundary clock of the target access point to correctly use the target time slot for exchanging data.

[0056] By synchronising with the boundary clock of the target access point, the travelling end node operates in accordance with the time-sensitivity of the network.Brief Description of the Drawings

[0057] Fig. 1 shows an example embodiment of a travelling end node in a wireless network that operates according to a wireless time-sensitive networking, W-TSN, mechanism;

[0058] Fig. 2 shows an example embodiment of performing a hand-over of a travelling end node in a wireless network that operates according to a W-TSN mechanism;

[0059] Fig. 3 illustrates a common time slot in schedules of access points in a wireless network that operates according to a W-TSN mechanism.

[0060] Fig. 4 shows steps according to a method for performing a hand-over of a travelling end node in a wireless network that operates according to a W-TSN mechanism, according to an embodiment.

[0061] Fig. 5 shows steps according to a method for performing a hand-over of a travelling end node in a wireless network that operates according to a W-TSN mechanism, according to an embodiment.

[0062] Fig. 6 illustrates obtaining three signal strength maps corresponding to three access points respectively according to example embodiments.

[0063] Fig. 7 illustrates location-signal strength information pairs according to example embodiments.

[0064] Fig. 8 shows non-overlapping regions within an environment covered by a wireless network according to example embodiments.

[0065] Fig. 9 shows steps to select a target access point according to example embodiments.Detailed Description of Embodiment(s)

[0066] Figure 1 shows an example embodiment of a wireless network 100. The wireless network 100 may, for example, operate according to a Wi-Fi standard. The example network 100 comprises associated end nodes 150, 160, 161 , 162, 163, 171 and access points 140, 141 , 142, 143 that operate according to a W-TSN mechanism. The network 100 may comprise additional access points which are not shown. End nodes 150, 160 are associated to access point 140. End nodes 161 , 171 are associated to access point 141 . End node 162 is associated to access point 142 and end node 163 is associated to access point 143. The end nodes are associated to the respective access points with the purpose of exchanging data. The access points 140, 141 , 142, 143 are able to service end nodes within respective service areas 130, 131 , 132, 133. The service areas 130, 131 , 132, 133 are shown in Figure 1 as oval-shaped regions, however, these shapes are only depicted for illustrative purposes. In practice, a service area of an access point is more complex and may even cover multiple stand-alone or detached regions. The end nodes 150, 160, 161 , 162, 163, 171 may be mobile end nodes or may be stationary end nodes. In this example, end node 150 is a mobile end node travelling away from the access point 140 towards a border of the service area 130, inciting a need for a hand-over. The travelling end node 150 moves in a direction indicated by the arrow. At a first moment in time, the travelling end node 150 is at a location denoted by reference numeral 150. At a second, later, moment in time, the travelling end node will be at a location denoted by reference numeral 151 , illustrated by a dotted contour. At location 151 , the travelling end node will be outside of the service area 130 of access point 140. Therefore, a hand-over should preferably be performed before that happens, to avoid that end node 150 loses connection to the network 100. Further, the access points 140, 141 are connected to a network controller 110 via respective switches 120, 121. All the access points of the network,including access points 142, 143, are connected to the network controller 110 via switches, even though this is not explicitly shown in Figure 1 for the sake of clarity. The network controller 110 manages the network by interacting with the access points 140, 141 , 142, 143 via the switches 120, 121 of the network.

[0067] Referring to Figure 2, performing a hand-over of the travelling end node 150 within the example network 100 of Figure 1 is illustrated. In Figure 2, only access points 140, 141 are shown for a compact illustration. Further, the same reference numerals denote the same functional elements as in Figure 1. In this example, the travelling end node 150 travels from the area 130 serviced by initial access point 140 towards the area 131 serviced by target access point 141. The end node 150 will be handed over from initial access point 140 to target access point 141. The network 100 is managed by the network controller 110. To that end, the network controller 110 allocates time slots 250, 260, 261 , 271 in respective schedules 180, 181 of access points 140, 141. Schedule 180 of initial access point 140 comprises a time slot 250, reserved for exchanging data with end node 150, and a time slot 260, reserved for exchanging data with end node 160. By having allocated the time slots 250, 260, the network controller 110 ensures that access point 140 services end nodes 150, 160. Analogously, the target access point 141 has a schedule 181 comprising time slots 261 , 271 for respectively exchanging data with end nodes 161 , 171. The time schedules 180, 181 are repeated until the network controller 110 adjusts them. Subplot 101 shows how time schedule 180 of initial access point 140 varies over time. Similarly, subplot 102 shows how time schedule 181 of target access point 141 varies over time. Upon detecting a need for handing over the travelling end node 150 from initial access point 140 to target access point 141 , the network controller 110 will amend the schedules 180, 181 to enable that the travelling end node 150 can be serviced by the target access point 141. The network controller 110 allocates a target time slot 251 for exchanging data with the travelling end node 150 within the schedule 181 of the target access point 141 , while avoiding time slots 261 , 271 of the target access point 141 that are allocated to end nodes 161 , 171. Further, the network controller 110 determines a hand-over time 400 for initiating the hand-over from the initial access point 140 to the target access point 141. From that moment on, the end node 150 can communicate with the target access point 141 without interfering with other data exchange. In this example, acontrol mechanism is provided to verify that the end node 150 was successfully handed over to target access point 141. For example, this may be done by, at a predetermined time 401 , e.g. a fixed amount of time after the hand-over time 400, determining whether or not the hand-over was successful. For example, the target access point 141 may be configured to report back to the network controller 110 when a communication link is established with the travelling end node 150. If by time 401 , a confirmation has been received, by the network controller 110, that the communication link has been established, the hand-over is considered to be successful. As illustrated in Figure 2, upon determining that the hand-over is successful, the time slot 250 in time schedule 180 of the initial access point 140 allocated for exchanging data with the travelling end node 150 may be removed. By doing so, resources of initial access point 140 are freed for exchanging data with other end nodes. Contrarily, if it is determined that the hand-over was unsuccessful, the time slot 251 in time schedule 181 of the target access point 141 allocated for exchanging data with the travelling end node 150 may be removed. In that case, a new attempt may be made, or, the travelling end node 150 may be handed over to another access point.

[0068] Before performing the hand-over as described above, the network controller 110 performs a step of determining one or more candidate access points 141 , 142 that are within range for exchanging data with the travelling end node 150. Access point 143 may for example not be considered within range of the travelling end node 150. This may be based on a distance between the access point 143 and the travelling end node 150. Alternatively, or additionally, excluding access point 143 from the set of candidate access points may be decided by the network controller based on a measure for signal quality. For example, the determining may be based on a signal strength between the end node 150 and access point 143. Scanning may, for example, be performed by the travelling end node 150, i.e. the travelling end node 150 may scan for beacons transmitted by one or more access points. In order to comply with W-TSN requirements, the network controller may allocate a scan time slot, during which access points can broadcast beacons to be received by travelling end nodes, as such preventing interference with time slots allocated for data exchange. The network controller 110 can allocate the scan time slot in the schedule of the access points 140, 141 , 142, 143 of the network, enabling travellingend nodes to scan for beacons. The scan time slot may be allocated to coincide in all access points, or, the scan time slot may be allocated interleaved for the different access points of the network, optionally interleaved with data exchange time slots, or access points may be grouped together to have a coinciding scan time slot that does not overlap with scan time slots of the remaining access points. In Figure 3, a common scan time slot 240 is illustrated within the time schedules 180, 181 , 182, 183 of respective access points 140, 141 , 142, 143, as respectively shown in subplots 201 , 202, 203, 204. Schedules 180, 181 , 182, 183 comprise respective time slots 250, 261 , 262, 263 for exchanging data with end nodes 150, 161 , 162, 163 respectively. It is noted that the travelling end node 150 also needs to know when the scan time slot 240 is allocated, in order to allow the travelling end node 150 to scan during the correct time slot 240. This may be achieved by sending, by the initial access point 140, the scan time slot 240 to the travelling end node 150.

[0069] It is noted that the schedules 180, 181 , 182 and 183 shown in subplots 101 , 102, 201 , 202, 203 and 204 are compactly represented in Figure 2 and Figure 3. In practice, schedules may have a duration longer than e.g. only two allocated time slots. When only few time slots are allocated in a schedule, the allocated time slots may be followed by one or more empty time slots. In that case, the schedule comprises a series of allocated time slots, followed by a series of empty time slots. Access points may be configured to address tasks other than data exchange during the empty time slots. In Figure 2 and Figure 3, empty time slots at the end of schedules are omitted for a concise representation.

[0070] Alternatively, probing may be performed, i.e. the travelling end node 150 may broadcast probe requests to be received by access points. Analogously to the scan time slot, a probe time slot may be allocated by the network controller 110 for this purpose. Advantageously, the probe time slot may also be common for all or for one or more access points. As such, the common time slot 240 illustrated in Figure 3 could also be a common probe time slot. The access points that receive the probe request, may respond with a probe response to the travelling end node 150.

[0071] Another way to determine signal quality may be to intercept data from overlapping basic service sets. For example, referring to Figure 1 , when frequencyre-use is applied, access point 140 and access point 142 may operate on a same channel, e.g. at the same frequency. In that case, the travelling end node 150 may, for example, overhear communication, i.e. intercept data, between the access point 142 and end node 162 to determine a signal quality to the access point 142. The determined signal quality may further be used for determining a location of the travelling end node 150.

[0072] Another way of determining the candidate access points based on signal quality may be by exploiting historical telemetry information collected by other end nodes 161 , 171 , 162, 163 respectively serviced by the other access points 141 , 142, 143. This is especially useful if there are other mobile end nodes in the network, that are able to store a measure for signal quality at various locations in the network.

[0073] Alternatively or additionally, determining the candidate access points may be based on localisation approaches. This may include determining a location of the travelling end node 150, or may comprise determining a relative distance measure, since knowing which access point is nearest to the travelling end node 150 may suffice. For example, the travelling end node 150 may be connected to a localisation network distinct from the wireless network 100, for example Ultra-Wide-Band. Alternatively, localisation may be performed by allocating, by the controller 110, a localisation time slot in the schedules of access points 141 , 142, 143, during which the travelling end node 150 can communicate with the access points 141 , 142, 143 for determining a relative location, e.g. applying the time of flight principle. The localisation time slot may be common for the access points, and 240 in Figure 3 therefore can also be interpreted as representing a common localisation time slot. Additionally, a direction of motion of the travelling end node 150 may advantageously be determined as well. A location of the travelling end node 150 can also be determined based on signal quality of intercepted data from access points that operate on the same channel as the initial access point. Determining the candidate access points 141 , 142 may be performed by taking into account more than one of the approaches discussed above.

[0074] Further, the network controller 110 selects a target access point 141 from the one or more candidate access points 141 , 142. Selecting the target access point141 from the candidate access points 141 , 142 may be done based on the same criteria as for determining the candidate access points 141 , 142, or a combination of approaches. For example, access point 143 might have been excluded from the set of candidate access points based on a distance to the travelling end node 150. Next, access point 141 may be selected as the target access point, also based on a distance to the travelling end node 150. Alternatively, selecting the target access point from the one or more candidate access points may be done randomly. Alternatively, the selecting of the target access point may be done based on the number of end nodes already serviced by the candidate access points. For example, if there is a candidate access points with very few serviced end nodes, the network controller can decide to allocate the end node to that candidate access point, which then becomes the target access point. Alternatively, the network controller may by default select the candidate access point closest to the initial access point. Advantageously, more than one criterium may be taken into account.

[0075] Figure 6 illustrates training of signal strength models to determine one or more candidate access points and to select a target access point according to example embodiments. Selection of target access point 141 may, for example, be done based upon such a scheme.

[0076] In Figure 6, subplot 601 shows an xy-plane representing the environment covered by a wireless network, e.g. network 100. The xy-plane can also show a subsection of such an environment pertaining to a subset of the access points, e.g. access points 140, 141 , 142. The xy-plane has an x-axis 630 and a y-axis 620. In subplot 601 , circles are shown at different locations in the xy-plane 620, 630. At these locations, corresponnding RSSI values are obtained with respect to one of the access points, e.g. access point 140. The circles have varying greyscale values indicating corresponding RSSI values according to a scale 641. Subplot 601 thus shows location-RSSI information pairs with respect to access point 140. Such data may, for example, be collected by different end nodes 150, 160, 161 , 162, 163, 171 over time. For this purpose, travelling nodes 150 may be configured to wander around to cover the environment, e.g. in a predetermined pattern. This may be done during one or more time slots allocated for this purpose. Alternatively, or additionally,travelling nodes may be configured to automatically collect location-RSSI information pairs during normal operation.

[0077] Similarly, subplot 602 shows measured RSSI values with respect to another access point, e.g. access point 141 , according to scale 642 in the same xy-plane 620, 630. Subplot 603 shows RSSI values with respect to yet another access point, e.g. access point 142. Subplot 603 also comprises a corresponding scale 643 to indicate the measured RSSI values.

[0078] Figure 7 shows a 3-dimensional, 3D, representation of captured location- RSSI data 700 in an alternative way to subplots 601 , 602, 603. The locations of data 700 are plotted in an xy-plane having x-axis 730 and y-axis 720, similar to subplots 601 , 602, 603. Further, instead of using a colour scale 641 , 642, 643, the RSSI value of each data point is illustrated as a height with respect to a z-axis 740.

[0079] Returning to Figure 6, the subplots 601 , 602, 603 are respectively converted into signal strength maps 611 , 612, 613 for access points 140, 141 , 142 respectively. The signal strength maps 611 , 612, 613 show a predicted RSSI value for each location in the xy-plane 620, 630. Corresponding scales 651 , 652, 653 indicate the greyscale values respectively. According to example embodiments, this is done by training a signal strength model, e.g. a machine learning, ML, model using the collected location-RSSI information pairs as labelled data in a supervised learning scheme. Such a machine learning model may, for example be an artificial neural network, ANN. The locations are provided as input and the corresponding RSSI values are used as ground truth outputs. Thus, such an ANN is trained to predict RSSI values based on location for a particular access point. A different ANN is used for each access point and thus for each signal strength map 611 , 612, 613. After training such an ANN, it can be inferred to each location of the xy-plane to obtain the corresponding signal strength map.

[0080] By combining the signal strength maps 611 , 612, 613, non-overlapping regions can be determined within an environment covered by access points 140, 141 , 142. This is illustrated by the environment map 820 shown in Figure 8. In subplot 820, each region 821 , 822, 823 is respectively associated with one of the access points 140, 141 , 142 of the network. Within a region 821 , 822, 823, thecorresponding access point 140, 141 , 142 provides the best RSSI. This allows determining an optimal hand-over location in conjunction with an optimal hand-over time. The optimal hand-over location may, for example, be a location just before, or just after crossing-over into another region. It is noted that map 820 may be explicitly determined based on the individual maps 611 , 612, 613, or may be used implicitly for handling of hand-overs.

[0081] Subplot 810 shows an alternative partition of the environment covered by a wireless network 100 according to example embodiments. The environment map 810 comprises the access points 140, 141 , 142. Corresponding regions 811 , 812, 813 are respectively associated with the access points 140, 141 , 142. Each region has a polygon shape and comprises a location of the respective access point. The edges of the polygon regions are indicated with corresponding coordinates in an xy- plane. The regions 811 , 812, 813 may be determined based on location, without taking into account any telemetry information.

[0082] In subplot 810, a buffer zone 815 is included at the boundaries between the regions 811 , 812, 813. According to example embodiments, a hand-over time is determined as the moment when the travelling end node 150 enters another region, i.e. upon having crossed the buffer zone. A corresponding hand-over location may also be determined. Such a buffer zone may, for example, also be introduced in the environment map 820 along the boundaries between the regions 821 , 822, 823.

[0083] It is noted that subplots 601 , 602, 603 and consequently subplots 61 1 , 612, 613 and subplot 820 show 3-dimensional data for a single access point. The three dimensions represent an x-coordinate, a y-coordinate and an RSSI value. It is noted that the corresponding principles can also be applied to 4-dimensional, 4D data, i.e. by modelling a 3D environment instead of a 2D environment. The four dimensions then represent an x-coordinate, a y-coordinate, a z-coordinate and an RSSI value.

[0084] Figure 9 shows steps to determine a target access point using ANNs trained on RSSI data 601 , 602, 603 in accordance with Figure 6 according to example embodiments.

[0085] In a step preceding the steps shown in Figure 9, the separate signal strength models for each access point may be trained in an offline stage, e.g. as explained above with respect to Figure 6. In such an offline stage, an initial environment map 820 may also be obtained. A subsequent online stage according to example embodiments is described in control flow diagram 900. The corresponding steps may be performed by a centralised entity of the wireless network 100, e.g. by network controller 110.

[0086] At some moment in time, the network controller 110 receives 902 a new data sample from an end node, e.g. node 150. The data sample comprises an RSSI value measured at a certain location (x;y) 901 with respect to a certain access point. The data sample comprises an x-coordinate, a y-coordinate, the corresponding RSSI value and information indicating to which access point this data pertains, e.g. access point 140.

[0087] In step 903, the network controller 110 determines whether or not to update the signal strength model, i.e. ANN, of the corresponding access point 140 based on this new sample. This may be done by comparing the measured RSSI value with an RSSI value predicted by the ANN of the access point 140. The controller 110 may, for example, move on to update 905 the ANN based on a difference between the measured and predicted RSSI values. For example, if the difference exceeds a predetermined threshold, the updating 905 will be performed. Otherwise, step 904 is taken.

[0088] The updating 905 may be done immediately, i.e. solely based on the new sample. Alternatively, updating may be performed in batches, i.e. upon having collected a certain amount of samples to collectively update the model. Batch updating provides the advantage of minimising updating overhead and resulting delays. The updating 905 may be done by applying an online gradient descent method, e.g. with a constant learning rate.

[0089] In step 904, a check is performed to determine if the node 150 is located in a buffer zone. If so, further hand-over-related operations may be skipped. If the node 150 is not located in a buffer zone, the optimal access point is determined for the node 150 to be associated to. This may be done by inferring each of the ANNscorresponding to the access points 140, 141 , 142 on the location of the new sample. Resulting RSSI values are then compared 909 to determine the access point providing the highest predicted RSSI. If this access point is different 910 from the initial access point 140, the new access point can be selected as target access point. The controller 110 may then initiate 911 a hand-over to this target access point.

[0090] In Figure 4 and Figure 5, steps of a method for performing a hand-over are shown. The network controller 110 is depicted alongside the target access point 141 , the initial access point 140, and the travelling end node 150. The steps are illustrated chronologically in time, represented by the vertical time axis, and will be discussed in the order presented in Figure 4 and Figure 5. However, these steps do not necessarily need to be performed in the order shown. It is noted that before performing the steps of Figure 4 and Figure 5, access points 141 , 142 were determined as candidate access points by the network controller 110. Thereafter, access point 141 was selected as target access point. Referring to Figure 4 and Figure 5, a first step 301 comprises allocating the target time slot 251 to the time schedule 181 of the target access point 141. Next, in step 302, information necessary for the hand-over is provided to the travelling end node 150. The information comprises the target time slot 251 , an identifier of the target access point 141 and the hand-over time 400. Providing the target time slot 251 may comprise providing a start time and an end time of the target time slot 251 inside a communication cycle, or may comprise providing either the start time or the end time of the target time slot 251 along with a duration of the target time slot inside a communication cycle. The identifier of the target access point 141 may comprise a frequency of the channel on which the target access point 141 operates. Optionally, in step 303, the network controller 110 may already associate the travelling end node 150 to the target access point 141 , i.e. perform the associating in name of the travelling end node 150 as such performing a so-called pre-association. Alternatively, the network controller 110 may aid the associating process in order to speed it up. Analogously, in step 304, the network controller 110 may preauthenticate the travelling end node 150 to the target access point 141. After step 304, the hand-over time 400 may occur.

[0091] After the hand-over time, the network controller 110 may determine 305, 306 whether or not the hand-over was successful, for example at a predetermined time 401 . The determining 305, 306 may initially be performed by the travelling end node 150. The travelling end node 150 communicates the result of the determining 305, 306 to the network controller 110, i.e. whether the hand-over was successful or unsuccessful.

[0092] Figure 4 illustrates what may happen in case the hand-over was successful. When a communication link 351 is successfully established, the travelling end node 150 may communicate the result of the determining 305 via the target access point 141 . Then, the travelling end node 150 may keep communicating 352 with the target access point 141 for exchanging data. Upon determining that the hand-over was successful, the network controller 110 may release the time slot 250 initially reserved for the travelling end node 150 within the schedule 180 of the initial access point 140, in step 308. The method for performing the hand-over may be concluded by step 309, wherein the network controller 110 undoes the authentication and association of the travelling end node 150 to the initial access point 140.

[0093] Figure 5 illustrates what may happen when the communication link 361 is not successfully established, which may result in the travelling end node 150 communicating the result of the determining 306 via the initial access point 140. Then, the travelling end node 150 may by itself revert back 307 to the initial access point 141 for exchanging data. Upon determining that the hand-over was unsuccessful, the network controller 110 may release the target time slot 251 reserved for the travelling end node 150 within the schedule 181 of the target access point 141 , in step 308. The method for performing the hand-over may be concluded by step 309, wherein the network controller 110 undoes the authentication and association of the travelling end node 150 to the target access point 141.

[0094] Although the present invention has been illustrated by reference to specific embodiments, it will be apparent to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied with various changes and modifications without departing from the scope thereof. The present embodiments are therefore to beconsidered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. In other words, it is contemplated to cover any and all modifications, variations or equivalents that fall within the scope of the basic underlying principles and whose essential attributes are claimed in this patent application. It will furthermore be understood by the reader of this patent application that the words "comprising" or "comprise" do not exclude other elements or steps, that the words "a" or "an" do not exclude a plurality, and that a single element, such as a computer system, a processor, or another integrated unit may fulfil the functions of several means recited in the claims. Any reference signs in the claims shall not be construed as limiting the respective claims concerned. The terms "first", "second", third", "a", "b", "c", and the like, when used in the description or in the claims are introduced to distinguish between similar elements or steps and are not necessarily describing a sequential or chronological order. Similarly, the terms "top", "bottom", "over", "under", and the like are introduced for descriptive purposes and not necessarily to denote relative positions. It is to be understood that the terms so used are interchangeable under appropriate circumstances and embodiments of the invention are capable of operating according to the present invention in other sequences, or in orientations different from the one(s) described or illustrated above.

Claims

CLAIMS1 . A method for performing a hand-over of a travelling end node (150) associated to an initial access point (140) in a wireless network (100); wherein the wireless network (100) comprises access points (140, 141 , 142, 143) that operate according to a wireless time-sensitive networking, W-TSN, mechanism by having allocated schedules (180, 181 ) with time slots (250, 260, 261 , 271 ) for exchanging data with respective end nodes (150, 160, 161 , 171 ); and wherein the wireless network (100) is managed by a network controller (110) configured to allocate the schedules (180, 181 ), and to perform the steps of:- determining one or more candidate access points (141 , 142) that are within range for exchanging data with the travelling end node (150, 151 );- selecting a target access point (141 ) from the one or more candidate access points (141 , 142);- allocating (301 ) a target time slot (251 ) for exchanging data with the travelling end node (150, 151 ) within the schedule (181 ) of the target access point (141 ) avoiding time slots (261 , 271 ) of the target access point (141 ) that are allocated to other end nodes (161 , 171 );- determining a hand-over time (400) for initiating the hand-over from the initial access point (140) to the target access point (141 ); and- providing (302) to the travelling end node (150, 151 ) the hand-over time (400), an identifier of the target access point (141 ), and the target time slot (251 ) for performing the hand-over.

2. The method according to claim 1 , wherein the access points (140, 141 , 142, 143) of the network operate in different channels, wherein the identifier specifies a channel of the target access point (141 ), and wherein the performing of the handover includes, by the travelling end node (150), switching to the channel of the target access point (141 ).

3. The method according to claim 1 or 2, wherein determining the candidate access points (141 , 142) comprises allocating, in schedules (201 , 202, 203, 204) of one or more access points (240, 241 , 242, 243), a scan time slot (240) for scanning, by thetravelling end node (150), for beacons transmitted by one or more access points (141 , 142, 143).

4. The method according to claims 1 or 2, wherein determining the candidate access points (141 , 142) comprises allocating, in schedules (201 , 202, 203, 204) of one or more access points (140, 141 , 142, 143), a common probe time slot (240) shared among the one or more access points (140, 141 , 142, 143) for sending, by the travelling end node (150), probe requests.

5. The method according to claim 4, further comprising, by the one or more access points (140, 141 ,142,143), upon receiving a probe request from the travelling end node (150), sending a probe response to the travelling end node (150), either directly or via the initial access point (140), or to the network controller (110).

6. The method according to any one of the preceding claims, wherein determining the candidate access points (141 , 142) comprises intercepting data exchanged between access points (142) and other end nodes (162) in the same channel in which the travelling end node (150) exchanges data with the initial access point (140).

7. The method according to any one of the preceding claims, wherein determining the one or more candidate access points (141 , 142) comprises determining a location of the travelling end node (150).

8. The method according to claim 7, wherein determining the location of the travelling end node (150) comprises receiving the location from a localisation network distinct from the wireless network (100).

9. The method according to claim 7, wherein determining the location of the travelling end node (150) comprises allocating, in schedules (201 , 202, 203, 204) of one or more access points (140, 141 , 142, 143), a common localisation time slot (240) shared among access points (140, 141 , 142, 143) for localizing the travelling end node (150) with respect to one or more access points (141 , 142, 143).

10. The method according to any of claims 7-9, wherein determining the one or more candidate access points (141 , 142) is further based on historical telemetry information between end nodes (150, 160, 161 , 171 , 162, 163) and access points (140, 141 , 142, 143).11 . The method according to any one of the preceding claims, further comprising associating (303) the travelling end node (150) to the target access point (141 ) by the network controller (110).

12. The method according to claim 11 , further comprising authenticating (304) the travelling end node (150) to the target access point (141 ) by the network controller (110).

13. The method according to any one of the preceding claims, further comprising, performing the steps of:- determining (305), by the travelling end node (150) and / or by the network controller (110), if the hand-over was successful; and- upon determining that the hand-over was successful (305):- updating (308), by the network controller (110), the schedule (101 ) of the initial access point (140) by releasing the time slot (250) allocated to the end node (150); and- de-authenticating and disassociating (309), by the travelling end node (150) and / or by the network controller (110), the travelling end node (150) from the initial access point (140).

14. The method according to any one of the claims 1 -12, further comprising, performing the steps of:- determining (305), by the travelling end node (150) and / or by the network controller (110), if the hand-over was unsuccessful; and- upon determining that the hand-over was unsuccessful (306):- reverting (307), by the travelling end node (150), to the initial access point (140); and- updating (308), by the network controller (110), the schedule (102) of the target access point (141 ) by releasing the time slot (251 ) allocated to the travelling end node (150); and- de-authenticating and disassociating (309), by the travelling end node (150) and / or by the network controller (110), the travelling end node (150) from the target access point (141 ).

15. The method according to any one of the preceding claims, wherein the access points (140, 141 , 142, 143) have respective boundary clocks, and wherein the method further comprises, by the travelling end node (150), synchronising with a boundary clock of the target access point (141 ).

Images