A clock synchronization method

Abstract

The invention relates to a clock synchronization method performed by a central unit (11) to provide time synchronization between nodes (13) in a network (1), Accordingly, the method is characterized in that comprising steps of; transmitting of time-stamped data by a base station (12) to each node (13) within the coverage area (15), synchronizing the node (13) clocks and calculating the clock errors based on the timestamped data received by each node (13), transmitting synchronized time information and calculated local clock errors of each node (13) to said central unit (11) by means of the base station (12), obtaining individual clock correction data for each node (13) based on clock errors received by the central unit (11), transmitting the clock correction data obtained to each node (13) to provide that the clockwork of the nodes (13) is regulated according to the individual clock correction data.

Description

[0001] A CLOCK SYNCHRONIZATION METHOD

[0002] TECHNICAL FIELD

[0003] The invention relates to a clock synchronization method performed by a central unit to provide time synchronization between nodes in a network.

[0004] BACKGROUND

[0005] LEO satellites do not stay at a fixed point in space, but constantly move in their orbits over time. Therefore, nodes / devices that will communicate with the LEO satellite can only communicate when the satellite is in the field of view of the device. For devices, this field of view or communication cone varies depending on the number of satellites, satellite orbit parameters and satellite antenna beam width. The device may have to wait for hours for a chance to communicate with a single LEO satellite.

[0006] For the reason mentioned above, it is very critical for devices to switch to the sleep mode (or turn off) when they are not communicating with the satellite to reduce power consumption. The power consumption may be further reduced by optimizing how often the device will go to sleep and the probability of reducing the amount of data to be sent or taking the data processing from energy-limited loT device to a central entity as much as possible. It is also important to define an efficient waking up method for the case the device is asleep.

[0007] For all these reasons, it is very important that the clocks of the devices on the network are tightly synchronized with the network reference clock. Setting the clocks of the devices in the network correctly and keeping them in sync does not only affect power management; reading the values of the sensors at the right time, data fusion, transmission Scheduling, communication protocols, security (limit use of keys, detect replay attacks), data consistency (caches, replicated data), interference mitigation (e.g., limiting cross interference at frame borders) etc. plays a critical role as well.

[0008] In the state of the art, clock correction and time synchronization are made with the GNSS chipset in some devices, so that the clocks of the devices on the network work in sync. However, the use of GNSS chipset turns devices into more expensive, more power-hungry, and complex devices. However, when too many devices are spread over remote areas and it is expected to work as long as possible without replacing the battery, or when the GNSS signal cannot be received for any reason (i.e. GNSS antenna or chipset malfunction, jamming, blocking... etc.) correcting the clocks and synchronizing their time is very important. These cheap, simple, and low-power devices without GNSS chipset usually have low-cost reference clocks. Also, the local time of the device shifts differently due to oscillator imperfection, environmental (temperature, pressure, humidity) changes, aging and voltage variation. A typical crystal oscillator has an error of 100 ppm. The clock drift 8.64 seconds per day. In a month it would be 259 seconds or 4.32 minutes per month.

[0009] All the problems mentioned above have made it necessary to make an innovation in the relevant technical field as a result.

[0010] BRIEF DESCRIPTION OF THE INVENTION

[0011] The present invention relates to a synchronization method to eliminate the above-mentioned disadvantages and bring new advantages to the relevant technical field.

[0012] An object of the invention is to realize a clock synchronization method for providing time synchronization between each node in a network.

[0013] An another object of the invention is to realize a clock synchronization method to reduce the energy consumption of nodes in a network.

[0014] To achieve all the objects mentioned above and that will emerge from the following detailed description, the present invention relates to a clock synchronization method performed by a central unit to provide time synchronization between nodes in a network, Accordingly, a clock synchronization method is characterized in that comprising steps of;

[0015] -transmitting of time-stamped data by a base station to each node within the coverage area -synchronizing the node clocks and calculating the clock errors based on the timestamped data received by each node

[0016] -transmitting synchronized time information and calculated local clock errors of each node to said central unit by means of the base station

[0017] -obtaining individual clock correction data for each node based on clock errors received by the central unit -transmitting the clock correction data obtained to each node to provide that the clockwork of the nodes is regulated according to the individual clock correction data. Thus, based on the clock correction data received from the central unit, the nodes are provided to establish clock synchronization without the help of GNSS reference.

[0018] A possible embodiment of the invention is characterized in that said clock correction data is a time dependent function and / or time-series data that enables the clockwork of the nodes to converge to network reference time based on clock drifts occurring in the nodes according to clock errors received.

[0019] Another possible embodiment of the invention is characterized in that transmitting synchronized time information and calculated local clock errors of each node to said central unit is performed with timestamp.

[0020] Another possible embodiment of the invention is characterized in that comprising steps of creating individual profiles for each node by the central unit with the data received from nodes and saving said individual profiles in a memory unit.

[0021] Another possible embodiment of the invention is characterized in that comprising step of updating the clock correction data for each data coming from the nodes and saving it to the corresponding profile.

[0022] Another possible embodiment of the invention is characterized in that comprising step of transmitting profile data of each node, containing at least one of the information oscillator clock type, clock precision, usage status and / or quality of service requirements, to the central unit.

[0023] Another possible embodiment of the invention is characterized in that in the step of obtaining the individual clock correction data, adding the profile data for each node in addition to the clock errors.

[0024] Another possible embodiment of the invention is characterized in that in the step of obtaining the individual clock correction data, adding the propagation times occurred in base station-to- node downlink and node-to-base station uplink signaling calculated by central unit. Another possible embodiment of the invention is characterized in that comprising step of transmitting the capability information data to regulate the clockwork according to the clock correction data of each node to the central unit.

[0025] Another possible embodiment of the invention is characterized in that said nodes are user equipment and / or loT device.

[0026] Another possible embodiment of the invention is characterized in that said base station is a non-terrestrial base station.

[0027] Another possible embodiment of the invention is characterized in that said base station is a low earth orbit satellite.

[0028] Another possible embodiment of the invention is characterized in that comprising step of repeating the mentioned steps in case any non-terrestrial base station is in the coverage area of at least one node.

[0029] Another possible embodiment of the invention is characterized in that comprising steps of transmitting future satellite arrival data to each node by central unit and updating sleep mode durations of nodes based on future satellite arrival data received by nodes.

[0030] Another possible embodiment of the invention is characterized in that said base station is a terrestrial base station.

[0031] BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a drawing illustrating a representative view of the network.

[0033] Figure 2 is a drawing illustrating a representative view of the synchronized clock.

[0034] DETAILED DESCRIPTION OF THE INVENTION

[0035] In this detailed description, the subject matter is explained with references to examples without forming any restrictive effect only in order to make the subject more understandable. Present invention is a clock synchronization method performed by a central unit (11) to provide time synchronization between nodes (13) in a network (1 ). Said method is characterized in that comprising steps of; transmitting of time-stamped data by a base station (12) to each node (13) within the coverage area (15) synchronizing the node (13) clocks and calculating the clock errors based on the timestamped data received by each node (13) transmitting synchronized time information and calculated local clock errors of each node (13) to said central unit (11 ) by means of the base station (12) obtaining individual clock correction data for each node (13) based on clock errors received by the central unit (11) transmitting the clock correction data obtained to each node (13) to provide that the clockwork of the nodes (13) is regulated according to the individual clock correction data. In a possible embodiment of the invention said clock correction data is a time dependent function and / or time-series data that enables the clockwork of the nodes (13) to converge to network reference time based on clock drifts occurring in the nodes (13) according to clock errors received.

[0036] In a possible embodiment of the invention said nodes mentioned in this description are different loT devices and / or network elements / appliances such as loT gateways, aggregators, etc.), wireless sensor networks (WSNs) (e.g., where nodes are different sensors and / or network elements / appliances such as WSN gateways, aggregators, etc.), closed loop control and / or industrial automation (e.g., where nodes are manufacturing equipment, robots, drones, sensors, Programmable Logic Controllers (PLCs), actuators, network elements / appliances, etc.), telecommunications networks (e.g., where nodes are different network access nodes / base stations, user equipment, network functions, edge compute nodes, etc.), data centers (e.g., where nodes are different servers and / or switch fabric including interconnections between difference servers and between servers and hardware accelerators), instrumentation (e.g., where nodes are individual instruments and / or measuring devices), components in autonomous or semi-autonomous vehicles, interconnect (IX) technologies including multi-socket processors, audio video (AV) including professional AV (pro AV) (e.g., synchronizing multiple AV devices such as WiFi speakers, Bluetooth earbuds, webcams, etc. in conference rooms, live performances (medium and large venues), etc.), media clock porting (e.g., satellite Rx streaming content over LAN), power grid (e.g., communication links used to selectively isolate faults on high voltage lines, transformers, reactors and other important electrical equipment where nodes are individual electrical equipment), and / or the like.

[0037] The clock synchronization method is further characterized in that transmitting synchronized time information and calculated local clock errors of each node (13) to said central unit (11) is performed with timestamp.

[0038] The clock synchronization method further comprises steps of creating individual profiles for each node (13) by the central unit (11 ) with the data received from nodes (13) and saving said individual profiles in a memory unit.

[0039] The clock synchronization method further comprises step of updating the clock correction data for each data coming from the nodes (13) and saving it to the corresponding profile.

[0040] The clock synchronization method further comprises step of transmitting profile data of each node (13), containing at least one of the information oscillator clock type, clock precision, usage status and / or quality of service requirements, to the central unit (11 ).

[0041] The clock synchronization method is further characterized in that in the step of obtaining the individual clock correction data, adding the profile data for each node (13) in addition to the clock errors.

[0042] The clock synchronization method is further characterized in that in the step of obtaining the individual clock correction data, adding the propagation times occurred in base station-to- node (12-13) downlink and node-to-base station (13-12) uplink signaling calculated by central unit (11 ).

[0043] The clock synchronization method further comprises step of transmitting the capability information data to regulate the clockwork according to the clock correction data of each node (13) to the central unit (11).

[0044] In a possible embodiment of the invention said base station (12) is a non-terrestrial base station. In another possible embodiment of the invention said base station (12) is a terrestrial base station. In another possible embodiment of the invention said non-terrestrial base station (12) is a satellite. In another possible embodiment of the invention said base station (12) is a low earth orbit satellite. In a possible embodiment of the invention the clock synchronization method comprises step of repeating the mentioned steps in case any non-terrestrial base station is in the coverage area of at least one node (13). In a possible embodiment of the invention the clock synchronization method comprises steps of transmitting future satellite arrival data to each node (13) by central unit (11 ) and updating sleep mode durations of nodes (13) based on future satellite arrival data received by nodes (13).

[0045] Hereinafter, the execution of one embodiment of the invention is described. In the main frame; the satellite transmits the time stamped data in the satellite downlink signal for clock correction and synchronization to the user equipments (UE) within the coverage area (15). UEs first correct their clocks and synchronize themselves with the time stamped data they receive from the satellite. They then calculate their local clock errors. Calculated local clock errors and synchronized time information of UEs are transmitted to the satellite in an uplink signal with time stamp. The clock errors, UEs reference times and satellite reference time information are sent by the satellite to the satellite gateway located in the ground segment. Satellite Gateway (14) sends this data to the data center so that necessary operations can be performed. The data of UEs including clock errors, UEs reference times and satellite reference time are first recorded in the database. After collecting sufficient amount of data, data processing activities are carried out. These processing activities include cleaning missing or outlier data, training the data, and then making predictions. As a result of all these processes, a time-based clock error profile is obtained for all UEs. Based on these clock error profiles, time-based clock correction profiles are created for each UE. This correction data is transmitted to the satellite gateway side to be sent to the UEs over the satellite. The Satellite gateway combines the satellite pass informations with the correction data and transmits it to the satellite. The clock correction data of the UEs and the arrival information of the future satellites are sent to the UEs in the downlink signal when the satellite is entered into the coverage area. UEs receive time-based clock correction and future satellite arrival data. While UEs can keep their clocks in sync with the clock correction data for a long time, they will be able to know exactly when they need to be turned on and off by using the coming satellite arrival data. In figure 2, it is shown that after UE receives clock correction data, it corrects its clock and keeps its long-term clock synchronized.

[0046] The UE may provide some relevant information (such as oscillator clock type, clock precision, use case, QoS requirements and so on) and the network may use them in clock profile generation, next synchronization time calculation and synchronization mode switching. Clock profile-related information may be inquired by the network over e.g., an RRC IE UECapabilityEnquiry, responded by the UE with e.g., an RRC information element (IE) ueCapabilitylnformation. This information exchange may take place in initial access after RRC connection is established or any time while UE is in connected state. This RRC messaging mechanism is already defined in 3GPP for both LTE and NR (see 3GPP TS 36.331 and 3GPP TS 38.331). An exemplary call flow describing the operation undertaken in phase 0 can be seen in the figure below. The network may also inquire in UECapabilityEnquiry about whether the UE has required capability to execute profile-based time synchronization (the mode proposed in this document) and UE may also indicate in ueCapabilitylnformation whether it is capable of executing profile-based time synchronization.

[0047] Initially the deployed loT device applies regular synchronization with the network by utilizing 5G clock synchronization or IEEE PTP. A signal consisting of a synchronization reference signal and a time tag is broadcasted by network. The time tag information indicates the time at the reference (when generated) and synchronization reference signal can be used to measure propagation delay to further align the oscillator clock. Configured UEs listen and each calculates local clock shift to be applied as follows: Tshift,i '>=Tre(k) TinternaiUE ' k) where Trefis the reference time tag value, Tinternal UE.(k) is the internal clock state of i’th UE at k’th time frame / slot and Tpropis propagation offset calculated with synchronization reference signal transmitted alongside the time tag.

[0048] UEs synchronize their clock for the next transmission time with calculated local clock shift ^clock,UEi '-

[0049] Equation 1: Time correction

[0050] Up until this part, this operation has been a regular time synchronization process. The novelty of this phase lies behind UEs sending calculated Tshift i(k) values to the network optionally along with a transmission time tag. The calculated Tshift i(k) values are forwarded from LEO satellites to satellite gateway and recorded at the network datacenter in a UE- specific database so they can later be used in the user-specific clock profile generation. Since the network and UE clocks are assumed to be synchronized after Equation 1 , the time tag sent from the UE to network could indicate to the network the propagation time Tpropat that moment. Assuming that the downlink reference signalling and uplink response are not far apart in time, and the distance between LEO and UE does not change significantly, then Tpropt may also be calculated at the network side.

[0051] The collected UE specific clock drift data is processed in a training or profiler block, which may be implemented with any optimization tool including artificial intelligence and machine learning methods. Some auxiliary information like clock type or clock precision, which may be collected during initialization step, can be also utilized in the training period. The profile block returns UE clock error profile, which is comprised by a series of expected error values, and some metric which indicates the reliability or convergence of the clock error profile, based on e.g., how well it fits to test data. While this metric may directly indicate whether the UE may switch to profile-based synchronization, there may be an additional entity, which uses this test metric and other parameters such as clock type, QoS requirements, use case, traffic model etc. and make a decision whether the user may switch to profile-based synchronization. This entity may be integrated into mission planning unit, which will be detailed later, and whether the UE may switch to profile-based time synchronization can be determined by taking satellite positions and constellation status are in consideration in addition to the training convergence. While switching decision can be enforced by the network, with an aim to e.g., reduce synchronization overhead, the network may just send an indicator revealing the UE clock profile for a specific UE is mature enough to be used, and UE may select whether to switch to profile-based synchronization, depending on its own implementation. It may be also the case that for certain conditions, the clock error profile may be mature while for some conditions it may not be yet, more data or training may be needed, or more tight requirements should be fulfilled. In these cases, an explicit signalling should be performed from network to UE. The inverter block inverts UE clock error profile, converting it to UE clock error correction, which will be shared with the UE. The input and output of this branch will be a series of data points. Upon determination of convergence of clock error profile, in the mission planning block at the satellite gateway, the validity time until which the UE clock error profile is expected to be valid and the UE could stop transmitting and receiving synchronization related signalling is calculated. The validity time may depend on the satellite availability and service status, reliability of clock error profile, current network conditions, UE capabilities, use case, and other parameters. The validity time should be well aligned with positioning of LEO satellites so that there should be a satellite serving when next synchronization information is expected by the UE. The UE is instructed by the network that the time synchronization profile is mature enough so it may switch to profile-based synchronization and perhaps turn off its receivers listening to synchronization signals until the end of the validity time if its use case and implementation allow. The validity time is communicated to the UE in terms of next synchronization time when it is anticipated to be in a communication cone of a LEO satellite (that is n’th from current). Depending on various parameters, such as profile maturity / convergence, satellite and constellation status, use case, QoS requirements, UE hardware and clock type etc. a confidence a confidence interval value indicating the expected accuracy of validity time. Depending on the implementation of UE, it may use this information e.g., for early / late wakeups. The confidence interval may be adaptively determined by the network based on e.g., earlier successful operation. The outputs of the training phase are 1 ) error correction offset in form of time series data, 2) profile-based time synchronization ready indicator, 3) time of the next synchronization information delivery and 4) confidence interval indicator. The network may communicate these information elements to UE over e.g., RRC signalling.

[0052] While operating in regular synchronization mode, the UE may receive an indication regarding the availability of the profile-based time synchronization. The message may also contain an instruction for the UE to (immediately) switch to the profile-based mode. Otherwise, UE may determine whether it would switch to the profile-based synchronization mode, depending on its implementation. Once the UE switch to the profile-based time synchronization mode, it UE will expect a DL transmission from the network comprising: 1 ) Clock shift correction values, 2) When the next information dispatch will occur (validity time), 3) Confidence value on the validity time. The UE applies clock shift correction value and if it does not have any other task to execute, it may switch to sleep mode until next activity, which may be the next synchronization time Confidence value, depending on the states of network, use case and UE hardware, may indicate to the UE the estimated accuracy of the provided next synchronization time. UE may use this information to wake up early / late, depending on its implementation.

[0053] Synchronization errors may be caught by UL / DL transmission errors. During expected on- time, the UE may have opportunity to inform the network that it is experiencing issues and / or to ask to switch to hard-synchronization mode. The network may also figure out issues in transmission link and then force UE to switch to the hard-synchronization mode.

[0054] The scope of protection of the invention is specified in the attached claims and cannot be limited to those explained for sampling purposes in this detailed description. It is evident that a person skilled in the art may exhibit similar embodiments in light of the above-mentioned facts without drifting apart from the main theme of the invention.

[0055] REFERENCE NUMBERS GIVEN IN THE FIGURE

[0056] 1 Network 11 Central Unit

[0057] 12 Base Station

[0058] 13 Node

[0059] 14 Satellite Gateway

[0060] 15 Coverage Area CD Clock Drift

[0061] CCD Clock Correction Data

[0062] SC Synchronized Clock

Claims

CLAIMS1. A clock synchronization method performed by a central unit (11) to provide time synchronization between nodes (13) in a network (1) characterized in that comprising steps of;- transmitting of time-stamped data by a base station (12) to each node (13) within the coverage area (15)- synchronizing the node (13) clocks and calculating the clock errors based on the timestamped data received by each node (13) transmitting synchronized time information and calculated local clock errors of each node (13) to said central unit (11) by means of the base station (12)- obtaining individual clock correction data for each node (13) based on clock errors received by the central unit (11)- transmitting the clock correction data obtained to each node (13) to provide that the clockwork of the nodes (13) is regulated according to the individual clock correction data.

2. The clock synchronization method according to claim 1 , characterized in that said clock correction data is a time dependent function and / or time-series data that enables the clockwork of the nodes (13) to converge to network reference time based on clock drifts occurring in the nodes (13) according to clock errors received.

3. The clock synchronization method according to claim 1 , characterized in that transmitting synchronized time information and calculated local clock errors of each node (13) to said central unit (11) is performed with timestamp.

4. The clock synchronization method according to claim 1 , characterized in that comprising steps of creating individual profiles for each node (13) by the central unit (11 ) with the data received from nodes (13) and saving said individual profiles in a memory unit.

5. The clock synchronization method according to claim 4, characterized in that comprising step of updating the clock correction data for each data coming from the nodes (13) and saving it to the corresponding profile.

6. The clock synchronization method according to claim 1 , characterized in that comprising step of transmitting profile data of each node (13), containing at least one of the information oscillator clock type, clock precision, usage status and / or quality of service requirements, to the central unit (11 ).

7. The clock synchronization method according to claim 6, characterized in that in the step of obtaining the individual clock correction data, adding the profile data for each node (13) in addition to the clock errors.

8. The clock synchronization method according to claim 1 , characterized in that in the step of obtaining the individual clock correction data, adding the propagation times occurred in base station-to-node (12-13) downlink and node-to-base station (13-12) uplink signaling calculated by central unit (11 ).

9. The clock synchronization method according to claim 1 , characterized in that comprising step of transmitting the capability information data to regulate the clockwork according to the clock correction data of each node (13) to the central unit (i n-10. The clock synchronization method according to claim 1 , characterized in that said nodes (13) are user equipment and / or loT device.

11. The clock synchronization method according to claim 1 , characterized in that said base station (12) is a non-terrestrial base station.

12. The clock synchronization method according to claim 11 , characterized in that said base station (12) is a low earth orbit satellite.

13. The clock synchronization method according to claim 11 , characterized in that comprising step of repeating the mentioned steps in case any non-terrestrial base station (12) is in the coverage area of at least one node (13).

14. The clock synchronization method according to claim 11 , characterized in that comprising steps of transmitting future satellite arrival data to each node (13) by central unit (11 ) and updating sleep mode durations of nodes (13) based on future satellite arrival data received by nodes (13).

15. The clock synchronization method according to claim 1 , characterized in that said base station (12) is a terrestrial base station.

Images