Technical Field of the Invention

The disclosure concerns a method of scheduling a transmission in a

telecommunications network between a user device and a base station from a plurality of base stations in the telecommunications network and an associated network device.

Background to the Invention

The Third Generation Partnership Project (3GPP) has been developing

specifications for Fifth Generation (5G) telecommunication standards, in particular including a new radio protocol stack. Providing low latency services is considered a highly desirable use of 5G technology. This new radio protocol stack is expected to operate largely on unpaired spectrum, which typically uses a Time Division Duplex (TDD) arrangement, in which the user device and base station can either transmit or receive at any given time, but not both.

A drawback of TDD operation is that, because simultaneous transmission and reception is not possible, significant round trip delays are likely, since the time duration between a downlink assignment instance of the radio frequency bandwidth and an uplink assignment instance of the radio frequency bandwidth may be much longer than for Frequency Division Duplex (FDD) operation. This round trip delay may affect error control mechanisms using acknowledgment (ACK) and non-acknowledgment (NACK) messages, such as Automatic Repeat Request (ARQ) and Hybrid ARQ (HARQ). Services requiring low latency and using such error control mechanisms could therefore have difficulty operating on these systems.

One way to mitigate the latency problem is using self-contained frames. In this structure, a fast ACK/NACK message may be transmitted within the same sub-frame as the data to which it refers. This is feasible in small isolated cells, such as cells operating above 6GHz frequency bands. However, it is impractical for use with frequency bands below 6GHz, for which there is the high possibility that (in order to coexist with other operators), the uplink-downlink configurations may be restricted. This may prevent dynamic switching to allow transmission of ACK/NACK messages in response to data. In addition, regulators in some markets have indicated specific ratios of uplink-downlink configuration that must be used. This may further prevent such implementations. Restrictions on the uplink-downlink configuration may further mean restrictions on when HARQ ACK messages can be sent in response to data received on the TDD carrier for both 'standalone' and 'dual connectivity' deployment scenarios. For data transmitted on the TDD carrier requiring low latency, delays in transmitting ACK/NACK messages would be a problem.

Existing Fourth Generation (4G) telecommunication standards, such as those specifying a radio protocol stack for Long Term Evolution (LTE) systems, have already tried to address a similar problem, using a radio frame split into 10 sub-frames. As an example, Table 1 (below) is taken from 3GPP TR 36.21 1 , "Evolved Universal Terrestrial Radio Access (E-UTRA); Physical channels and modulation" and demonstrates some selectable structures for transmission time frames (radio frames). These are temporally sub-divided into sub-frames. For each sub-frame in the transmission time frame, "D" denotes a downlink sub-frame reserved for downlink transmissions, "U" denotes an uplink sub-frame reserved for uplink transmissions and "S" denotes a special sub-frame.

Table 1 : Uplink-downlink configurations

Taking into account uplink-downlink configuration 2 (the best in terms of waiting time for switching between uplink and downlink), it can be seen that a delay of 2 to 3 ms will be experienced, depending on the sub-frame number before a ACK/NACK feedback message can be sent. This limited number of opportunities for switching from uplink to downlink or vice versa can cause a problem with sending fast feedback in the same TDD band; a wait for a two transmissions is required before the ACK/NACK message can be sent.

The use of a fixed switch point periodicity may further prevent any reduction in this delay by shortening the Transmission Time Interval (TTI). Some uplink reference signals, such as Sounding Reference Signals (SRS), are also desirably transmitted periodically for accurate channel estimation. This is intended for beamforming use in 5G systems and their transmission could be made significantly more difficult by existing structures. The 3GPP Plenary has also agreed to support 4G LTE and 5G New Radio (NR) uplink sharing. Referring to Figure 1 , there is shown a schematic diagram illustrating such LTE and NR uplink sharing. First User Equipment (UE) 1 can communicate with: a first LTE system 10 using a LTE uplink and downlink 1 1 ; a second LTE system 20 using a LTE downlink and NR uplink 21 ; and a NR system 30 using a NR uplink and a NR downlink 31 . Another, second UE 2 is shown for illustration and can communicate with the second LTE system 20 using a LTE uplink and a LTE downlink 22.

Therefore, NR UE devices are allowed to use the LTE band (preferably with FDD) to send their uplink ACK/NACK message quickly. However using the LTE band for ACK/NACK message may not be the best option for low latency scenarios in all the uplink- downlink configurations shown in the table above. For example if the UE is receiving on subframe 5 or 6, it will typically have an uplink opportunity in next subframe, so the UE can transmit its ACK/NACK in the NR TDD band. To fully utilise this feature for low latency use cases a scheduling approach would be desirable. In addition, some message exchange should be specified between the NR and LTE base stations. Each or both of these may allow the NR user device to achieve the lowest latency possible.

Summary of the Invention

Against this background, the present invention provides a method of scheduling a transmission in a telecommunications network between a user device and a base station from a plurality of base stations in the telecommunications network according to claim 1 and a network device in line with claim 15.

A transmission (preferably uplink, but optionally downlink, and more preferably comprising data acknowledgement information, such as ARQ or ACK/NACK messages) is to be scheduled in a telecommunications network between a user device (UE) and one of the network's base stations, a first of which uses a Time Division Duplex (TDD) Radio Access Technology (RAT) (such as a 5G/NR RAT) and a second of which uses a

Frequency Division Duplex (FDD) RAT (in particular, LTE). The UE (which is

advantageously NR-compatible) may already be serviced by the first base station. A first scheduling delay for transmissions between the UE and the TDD base station and a second scheduling delay for transmissions between the UE and the FDD base station are determined. The decision as to how to schedule the transmission (in particular, scheduling it with the first base station or second base station) is based on the determined first and second scheduling delays. A network device of a telecommunications network that comprises a first base station using a TDD RAT and a second base station using a FDD RAT, the network device being configured to operate in accordance with this method (or any method herein described) is also provided. The network device may be a user device, base station, a controller configured to provide scheduling information for a user device and/or base station or a combination of these. The method and/or device may be implemented in software, hardware, firmware, programmable logic, some other type of programmable or configurable hardware or a combination of these technologies, for example.

The first scheduling delay is beneficially a time until the transmission can be made using the TDD RAT (for example, based on the current sub-frame allocation and the next sub-frame allocation until the transmission can be scheduled, in accordance with the required direction of the transmission).

The second scheduling delay may comprise one or both of: (i) a time taken to schedule transmissions between the user device and the second base station; and (ii) a latency for communication between the first and second base stations. The time taken to schedule transmissions between the user device and the second base station may be determined by communicating a message from the first base station to the second base station to request an indication of the time taken to schedule transmissions between the user device and the second base station; and/or communicating an estimate of the time taken to schedule transmissions between the user device and the second base station from the second base station to the first base station (in particular at periodic intervals), for example in response to the request from the first base station. If the second base station has reserved resources for the user device, the time taken to schedule transmissions between the user device and the second base station may be determined as zero. If the first and second base stations are co-located, the latency for communication between the first and second base stations may be determined as zero. Alternatively, the latency for communication between the first and second base stations may be measured (for instance, using a ping communication).

Identifying whether to schedule the transmission using the first or second base stations may be by comparing the first scheduling delay with the second scheduling delay. Optionally, this comparison further accounts for additional delays in processing at the first and second base stations (by adding a value to the first scheduling delay).

If the first and second base stations are not co-located, the second scheduling delay may consist of (that is, comprises only) a time taken to schedule transmissions between the user device and the second base station. Then, if the comparison of the first and second scheduling delays indicates that the second base station should be used for the transmission, the first scheduling delay may additionally be compared with the sum of the second scheduling delay and a latency for communication between the first and second base stations. Optionally, this further comparison may further account for additional delays in processing at the first and second base stations (by adding a value to the first or second scheduling delay).

The transmission may then be scheduled in accordance with the identification made by the method. For example, the step of identifying may cause the transmission to be scheduled using the second base station. Then, the first base station may communicate to the second base station an instruction to schedule the transmission. In particular, the instruction may specify the user device. Additionally or alternatively, the first base station may communicate to the second base station a timing offset between transmissions in respect of the first base station in a first direction (for example, downlink) and transmissions in respect of the second base station in a second direction, opposite to the first direction (for instance, uplink). These communications may be done in the same or separate messages.

Also, combinations of any specific features shown with reference to one

embodiment or with reference to multiple embodiments are also provided, even if that combination has not been explicitly detailed herein.

Brief Description of the Drawings

The invention may be put into practice in various ways, some of which will now be described by way of example only and with reference to the accompanying drawing in which:

Figure 1 shows a schematic diagram illustrating such LTE and NR uplink sharing;

Figure 2 schematically depicts options for accessing co-located LTE and NR base stations;

Figure 3 illustrates a flowchart for deciding the base station for making a

transmission and associated steps, when the FDD and TDD base stations are co-located and no reserved resources for the UE are provided at the FDD base station;

Figure 4 shows a flowchart for deciding the base station for making a transmission and associated steps, when the FDD and TDD base stations are co-located and reserved resources for the UE are provided at the FDD base station;

Figure 5 schematically illustrates options for accessing non-co-located LTE and NR base stations; and Figure 6 depicts a flowchart for deciding the base station for making a transmission and associated steps, when the FDD and TDD base stations are not co-located and no reserved resources for the UE are provided at the FDD base station. Detailed Description of Preferred Embodiments

In general, the invention concerns the identification of whether to use an FDD link or TDD link (provided by a corresponding base station) for minimum latency of a transmission between a UE (user device) and a network, especially a ACK/NACK message. In practice, the transmission would typically be for uplink (relating to acknowledgment of a downlink message), for a UE serviced by the TDD base station, the TDD link would use a 5G RAT and the FDD link would use a 4G RAT (LTE). All of these specific configurations will be assumed in the examples below, which may be considered particularly advantageous use cases. However, one, some or all of these assumptions need not be the case and advantages of the approach may still be obtained. The 5G base station is termed a gNB and a 4G base station is termed a eNB below. Where a gNB is referenced, this may be understood as using a TDD RAT for both uplink and downlink. Where a eNB is referenced, this may be understood as using a FDD RAT for both uplink and downlink.

The basic decision between the TDD system and the FDD system for the transmission is made based on determined respective scheduling delays for the two systems. These can be compared with each other, optionally subject to other factors being taken into account. The system that will provide the lowest latency (that is, the lowest scheduling delay) is typically selected.

In practice, the way in which the choice between the systems is made may depend on a number of factors. One factor is whether the gNB and eNB are co-located (at the same site). Another factor is whether the eNB (which is not currently servicing the UE) has reserved resources for the UE. The specific implementations will be discussed below, in accordance with these factors.

Case 1 : NR gNB and LTE eNB at same site (collocated)

Referring to Figure 2, there are schematically depicted options for accessing co- located LTE and NR base stations. This drawing is similar to Figure 1 and similar (and in some cases, the same) reference numerals have been used. First User Equipment (UE) 1 can communicate with: a first LTE system 1 10 using a LTE uplink and downlink 1 1 1 using FDD RAT and frequency fi ; a second LTE system 120 using a LTE downlink with FDD RAT and NR uplink 121 using frequency f ₂ ; and a NR system 130 using a NR uplink and a NR downlink 131 using TDD RAT and frequency f ₃ . Frequency fi and frequency f ₂ will, in most circumstances, be the same frequency (so fi = f ₂ ). Another, second UE 2 is shown for illustration and can communicate with the second LTE system 120 using a LTE uplink and a LTE downlink 122. The base stations communicate with each other using an Xn interface (not shown), but since they are co-located, this interface has little or no latency. For say, first UE 1 , a decision to transmit an acknowledgement (ACK/NACK) message via either the second LTE system 120 or the NR system 130 is to be made.

Within this case there are 2 options: (1 ) no reserved resources at eNB for NR UL transmission for the UE; and (2) reserved resources at eNB for NR UL transmission for UE. These options will be discussed below.

Option 1

Referring to Figure 3, there is illustrated a flowchart for deciding the base station for making a transmission and associated steps, when the FDD and TDD base stations are co- located and no reserved resources for the UE are provided at the FDD base station. The steps will be discussed below.

201 . The 5G gNB 130 is transmitting low latency data to first UE 1 .

202. gNB 130 informs the eNB 120 to periodically send it the estimated time required to schedule UE 1 on the uplink resources of frequency f ₂ .

203. eNB 120 informs the gNB 130 of the delay (T2) taken to schedule a UE using

frequency f ₂ . It is anticipated that 3GPP would specify this message to be sent, for example through the Xn interface. An example would be a value of the lowest waiting time needed with decimal accuracy. T2 includes the Xn latency, which in the case of co-located eNB 120 and gNB 130 is considered negligible (close to zero), but is mentioned here for completeness.

204. gNB 130 checks which sub-frame in the TDD uplink/downlink configuration is due next, which depends on the chosen configuration as shown in Table 1 . Also, it calculates the delay (T1 ) until the next uplink scheduling opportunity using frequency f ₃ .

205. If T2 is larger or equal to (T1 +c), the gNB 130 schedules UE 1 on the NR uplink resources using frequency f ₃ . The value of c accounts for further delays in the UE or base station processing, c could be a positive or negative value, as it accounts for delays on both eNB 120 and gNB 130 sides. More information on the way in which c could be determined is discussed below. 206. Otherwise; the gNB 130 informs the LTE eNB 120 to immediately schedule UE 1 on the LTE uplink resources (using frequency f ₂ ). Again, 3GPP should specify this message to be sent for example through the x _g interface. The message should include an indication as to which UE the eNB 120 should allocate resources.

207. The gNB 130 informs the eNB 120 about the timing offset between the frequency f ₃ downlink and frequency f ₂ uplink. It is expected that 3GPP should also specify this message.

Option 2

Referring to Figure 4, there is shown a flowchart for deciding the base station for making a transmission and associated steps, when the FDD and TDD base stations are co- located and reserved resources for the UE are provided at the FDD base station. The steps will be discussed below.

301 . The 5G gNB 130 is serving UE 1 with low latency requirements.

302. The gNB 130 informs eNB 120 to reserve uplink resources using frequency f ₂ for UE 1 to send its NR ACK/NACK messages. A scheduling delay (T2) for the eNB 120 is equal to the Xn latency. The Xn latency in the case of co-located eNB 120 and gNB 130 is considered negligible (close to zero).

303. gNB 130 checks which sub-frame in the TDD uplink/downlink configuration is due next. This depends on the chosen configuration, as shown in Table 1 . Also it calculates the delay until the next uplink scheduling opportunity using frequency f ₃ (T1 ).

304. If T2 is larger or equal to (T1 +c), the gNB schedules UE 1 on the NR uplink

resources using frequency f ₃ . As with Option 1 , c is a value accounting for further delays in the base station processing (and/or optionally UE processing) and could be a positive or negative value, as it accounts for delays on both eNB 120 and gNB 130 sides.

305. Otherwise; the gNB 130 informs the LTE eNB 120 to immediately schedule UE 1 on the LTE reserved uplink resources using frequency f ₂ . Again, 3GPP should specify this message to be sent for example through the Xn interface. The message should include an indication as to which UE the eNB 120 should allocate resources.

306. The gNB 130 informs the eNB 120 about the timing offset between the frequency f ₃ downlink and frequency f ₂ uplink. It is expected that 3GPP should also specify this message. Case 2: NR qNB and LTE eNB are no co-located

Referring to Figure 5, there are schematically illustrated options for accessing non- co-located LTE and NR base stations. This drawing is similar to Figures 1 and 2 and similar (and in some cases, the same) reference numerals have been used. First User Equipment (UE) 1 can communicate with: a first LTE system 410 at a first site using a LTE uplink and downlink 41 1 using FDD RAT and frequency W, a second LTE system 420 at the first site using a LTE downlink with FDD RAT and NR uplink 421 using frequency f ₂ ; and a NR system 430 at a second site using a NR uplink and a NR downlink 431 using TDD RAT and frequency f ₃ . Frequency and frequency f ₂ will, in most circumstances, be the same frequency (so = h)- Another, second UE 2 is shown for illustration and can communicate with the second LTE system 420 using a LTE uplink and a LTE downlink 422. The first LTE system 410 and second LTE system 420 communicate with the NR system 430 using a Xn interface 440, which has an associated non-negligible latency. For say, first UE 1 , a decision to transmit an acknowledgement (ACK/NACK) message via either the second LTE system 420 or the NR system 430 is to be made.

Referring to Figure 6, there is illustrated a flowchart for deciding the base station for making a transmission and associated steps, when the FDD and TDD base stations are not co-located and no reserved resources for the UE are provided at the FDD base station. The steps will be discussed below.

501 . The gNB 430 is serving UE 1 with low latency requirements.

502. gNB 430 checks which sub-frame in the TDD uplink/downlink configuration is due next, which depends on the chosen configuration as shown in Table 1 . Also it calculates the delay (T1 ) until the next uplink scheduling opportunity using frequency f ₃ .

503. The gNB 430 informs eNB 420 to periodically send it the estimated time delay (T2) needed to schedule UE 1 on the uplink resources of frequency f ₂ .

504. The eNB 420 informs the gNB 430 of the delay (T2) needed to schedule UE 1 on the uplink resources of frequency f ₂ . At this time, the gNB 430 starts to measure the latency (T3) of the Xn interface (connection between the first and second sites), for example through a ping.

505. If T2 is larger or equal to T1 +C, the gNB 430 schedules UE 1 on the NR uplink

resources using frequency f ₃ . Value c is as defined with reference to case 1 above.

506. Otherwise, the gNB 430 calculates the Xn latency (T3) and recalculates the delay (T1 ) until the next uplink scheduling opportunity using frequency f ₃ . The gNB 430 recalculates T1 in order to take into account any changes in T1 during the time taken to measure the Xn latency (T3).

507. If T1 + c is less or equal to T3 + T2, the gNB 430 schedules UE 1 on the NR uplink resources on using frequency f ₃ .

508. Otherwise; the gNB 430 informs the LTE eNB 420 to immediately schedule UE 1 on the LTE uplink resources (using frequency f ₂ ).

509. The gNB 430 informs the eNB 420 about the timing offset between the frequency f ₃ downlink and frequency f ₂ uplink.

Based on the above, it will immediately be apparent how to modify the approach above for situations in which reserved resources for the UE are provided at the FDD base station, in line with Figure 4, but where the FDD and TDD base stations are not co-located.

Additional details about the parameter 'c' will now be presented and are equally applicable to any of the options or cases discussed above. Firstly, it should be noted that 'c' will typically be small (less than 50%, 40% or 30%, more preferably less than 25%, 20% or 15% and most preferably less than 10%, 5%, 2%, 1 %) of T1 , T2 and/or T3. The parameter may be set as a percentage of total delay (T1 +T2 or T1 +T2+T3, for instance) or it can be a fixed value based on experience (in other words, as a configurable parameter). The value of 'c' should not have a significant impact on the decision though. In many cases, c=0 will be used, particularly for an initial deployment and subsequent tuning of 'c' to take account of errors due to the non-zero nature of the additional processing delays may then be implemented. Additionally or alternatively, these delays additional processing delays may each be directly measured, estimated based on other measurements and/or iteratively adjusted to set 'c' accordingly. The value of c can be calculated, for instance, by substracting the eNB processing delay from the gNB processing delay and adding a term for error (that is, c = gNB processing delay - eNB processing delay ± error). The sign of c may depend on which of the processing delay values is larger.

In general terms, there may be considered a method of scheduling a transmission in a telecommunications network between a user device and a base station from a plurality of base stations in the telecommunications network, a first base station of the plurality of base stations using a Time Division Duplex (TDD) Radio Access Technology (RAT) and a second base station of the plurality of base stations using a Frequency Division Duplex (FDD) RAT. The method comprises: determining a first scheduling delay for transmissions between the user device and the first base station; determining a second scheduling delay for transmissions between the user device and the second base station; and identifying whether to schedule the transmission using the first or second base stations, based on the determined first and second scheduling delays.

A network device of telecommunications network that comprises a plurality of base stations is further considered. A first base station of the plurality of base stations uses a Time Division Duplex (TDD) Radio Access Technology (RAT) and a second base station of the plurality of base stations uses a Frequency Division Duplex (FDD) RAT. Then, the network device is configured to operate in accordance with any method as herein disclosed. The network device may be a user device, a base station, a controller configured to provide scheduling information for a user device and/or base station or a combination of these. Optional features in respect of the method will now be discussed, but it will be understood that these can additionally or alternatively be implemented as structural features of a corresponding network device, for example having a controller configured to operate in accordance with the method.

Preferably, the transmission is from the user device to the base station (uplink). Optionally the user device is already being serviced by the first base station. One or more of the following (or their opposite) may be the case: the FDD RAT is LTE; the TDD RAT is not LTE; the TDD RAT communicates at a higher bit rate than FDD RAT.

Optionally, the step of determining the first scheduling delay for transmissions between the user device and the first base station comprises determining a time until the transmission can be made using the TDD RAT.

Optionally, the step of determining the second scheduling delay for transmissions between the user device and the second base station comprises determining one or both of: a time taken to schedule transmissions between the user device and the second base station; and a latency for communication between the first and second base stations.

Preferably, the method further comprises: communicating a message from the first base station to the second base station to request an indication of the time taken to schedule transmissions between the user device and the second base station; and/or communicating an estimate of the time taken to schedule transmissions between the user device and the second base station from the second base station to the first base station in response to the request from the first base station. More preferably, the step of communicating an estimate of the time taken to schedule transmissions between the user device and the second base station is performed at periodic intervals. In some

embodiments: if the first and second base stations are co-located, the latency for communication between the first and second base stations is determined as zero; and/or the method further comprises: measuring the latency for communication between the first and second base stations (using a ping communication). In some cases, if the second base station has reserved resources for the user device, the time taken to schedule transmissions between the user device and the second base station is determined as zero.

In the preferred embodiment, the step of identifying whether to schedule the transmission using the first or second base stations comprises comparing the first scheduling delay with the second scheduling delay. Optionally, the step of comparing the first scheduling delay with the second scheduling delay further accounts for additional delays in processing at the first and second base stations (by adding a value to the first scheduling delay). In embodiments, if the first and second base stations are not co- located, the second scheduling delay consists of a time taken to schedule transmissions between the user device and the second base station. Then, if the comparison of the first and second scheduling delays indicates that the second base station should be used for the transmission, the step of identifying whether to schedule the transmission may further comprise comparing the first scheduling delay with the sum of the second scheduling delay and a latency for communication between the first and second base stations. The step of comparing the first scheduling delay with the sum of the second scheduling delay and a latency for communication between the first and second base stations may further account for additional delays in processing at the first and second base stations (for instance, by adding a value to the first or second scheduling delay).

Typically, the method further comprises scheduling the transmission in accordance with the identification. In some cases, the step of identifying causes the transmission to be scheduled using the second base station. Then, the method may further comprise one or both of: the first base station communicating to the second base station an instruction to schedule the transmission, the instruction specifying the user device; and the first base station communicating to the second base station a timing offset between transmissions in respect of the first base station in a first direction and transmissions in respect of the second base station in a second direction, opposite to the first direction. Typically, the first direction is downlink and the second direction is uplink.

Although specific embodiments have now been described, the skilled person will understand that various modifications and variations are possible. For example, this approach may be used for a range of different communication systems, which need not be cellular-based or wireless LAN-based. Other implementations may further be considered in practice from those suggested above. Also, combinations of any specific features shown with reference to one embodiment or with reference to multiple embodiments are also provided, even if that combination has not been explicitly detailed herein.