RANDOM ACCESS IN A SATELLITE COMMUNICATION SYSTEM

Abstract

The present invention provides a method of performing by a UE device a random access attempt in a communication system comprising at least one non-terrestrial transmission station, the method comprising receiving reference signals transmitted by the non-terrestrial transmission station; determining from the received reference signals a trip time between the UE device and the non-terrestrial transmission station; and using the trip time to control the random access attempt.

Claims

1. A method of performing by a user equipment, UE, device a random access attempt in a communication system comprising at least one non-terrestrial transmission station, the method comprising: receiving reference signals transmitted by the non-terrestrial transmission station; determining a trip time between the UE device and the non-terrestrial transmission station; and using the trip time to control the random access attempt, characterized in that the trip time is determined from the received reference signals.

2. The method of claim 1, wherein the determined trip time is used to determine a timing offset for an access attempt transmission such that the access attempt transmission by the UE device is received at the non-terrestrial transmission station within a pre-determined time frame.

3. The method according to claim 1, wherein the trip time is determined by at least one of: i) a measurement of a received reference signal power; ii) a determined difference between two or more successive reference signal power measurements; iii) a determination of a Doppler frequency shift of a received reference signal; iv) a determined difference between two or more successive reference signal Doppler shift determinations.

4. The method according to claim 1, wherein a second trip time between the UE device and the same or a different non-terrestrial transmission station is estimated using a pre-determined trip time and information relating to relative movement of one or more non-terrestrial transmission stations in the communication system.

5. The method according to claim 1 wherein the UE device makes a decision as to which of more than one random access methods to use for the access attempt transmission dependent of a reliability estimate of the trip time determination.

6. The method according to claim 1 wherein the trip time is used to determine a response time for reception of a response from the non-terrestrial transmission station.

7. The method according to claim 1, wherein the non-terrestrial transmission station is a satellite in low earth orbit.

8. A base station of a satellite enabled communication system, the base station being adapted to configure two sets of random access resources, wherein a first set of random access resources enables slotted ALOHA random access attempts by user equipment devices and a second set of random access resources enables unsynchronized random access by user equipment devices.

9. The base station according to claim 8, wherein the first set of random resources are provided for use by user equipment devices which are capable of deriving a reliable trip time to a non-terrestrial transmission station and the second set of random access resources are provided for use by user equipment devices which are not capable of deriving a reliable trip time to the non-terrestrial transmission station.

10. The method according to claim 2, wherein the trip time is determined by at least one of: i) a measurement of a received reference signal power; ii) a determined difference between two or more successive reference signal power measurements; iii) a determination of a Doppler frequency shift of a received reference signal; iv) a determined difference between two or more successive reference signal Doppler shift determinations.

11. The method according to claim 2, wherein a second trip time between the UE device and the same or a different non-terrestrial transmission station is estimated using a pre-determined trip time and information relating to relative movement of one or more non-terrestrial transmission stations in the communication system.

12. The method according to claim 3, wherein a second trip time between the UE device and the same or a different non-terrestrial transmission station is estimated using a pre-determined trip time and information relating to relative movement of one or more non-terrestrial transmission stations in the communication system.

13. The method according to claim 2 wherein the UE device makes a decision as to which of more than one random access methods to use for the access attempt transmission dependent of a reliability estimate of the trio time determination.

14. The method according to claim 3 wherein the UE device makes a decision as to which of more than one random access methods to use for the access attempt transmission dependent of a reliability estimate of the trip time determination.

15. The method according to claim 4 wherein the UE device makes a decision as to which of more than one random access methods to use for the access attempt transmission dependent of a reliability estimate of the trip time determination.

16. The method according to claim 2 wherein the trio time is used to determine a response time for reception of a response from the non-terrestrial transmission station.

17. The method according to claim 3 wherein the trip time is used to determine a response time for reception of a response from the non-terrestrial transmission station.

18. The method according to claim 4 wherein the trip time is used to determine a response time for reception of a response from the non-terrestrial transmission station.

19. The method according to claim 2, wherein the non-terrestrial transmission station is a satellite in low earth orbit.

20. The method according to claim 3, wherein the non-terrestrial transmission station is a satellite in low earth orbit.

Description

[0049] Preferred embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings in which:

[0050] FIG. 1 illustrates differing UE-satellite distances in a satellite based communication system;

[0051] FIG. 2 illustrates an exemplary multi-orbit satellite system;

[0052] FIG. 3 illustrates a variation of received signal strength with time and distance to an orbit plane;

[0053] FIG. 4 illustrates a variation of Doppler shift with time and distance to a satellite orbit plane;

[0054] FIG. 5 shows an exemplary variation of a distance from a satellite to a UE's projection point on the orbit plane of a satellite with time;

[0055] FIG. 6 illustrates relative distances for implementing the invention;

[0056] FIG. 7 shows the transmission of random access preambles by two UEs in a terrestrial communication system;

[0057] FIG. 8 shows the transmission of random access preambles by two UEs in a satellite based communication system in a conventional manner; and

[0058] FIG. 9 shows as an example a timing of a random access attempt when trip time compensation is applied according to an aspect of the invention.

[0059] FIG. 9 shows as an example a timing of a random access attempt when trip time compensation is applied according to an aspect of the invention. A beginning of the preamble transmit time T.sub.TX,2 of UE2 is selected so that a selected access slot begins twice the trip time later (at T.sub.0,2), i.e. UE2 is selecting the transmit time as

T.sub.Tx,2=T.sub.0,2−2×t.sub.trip2.  (1)

wherein
t.sub.trip2: is a trip time of signals transmitted from UE2 to the satellite

[0060] Therefore, the reception at the satellite of the random access attempt takes place within given access slot boundaries. A small guard time only is required to compensate for measurement errors.

[0061] FIG. 1 shows an example radio access network 10 based on LEO satellites. The figure depicts two satellites (SAT.sub.n,m and SAT.sub.n,m+1), where the index m iterates the satellites on the same orbit (orbit n), and two UEs (UE1 and UE2). In an exemplary manner, two typical distances for LEO satellites are referenced in FIG. 1: the height of the satellites over ground (781 km) and the typical distance of a satellite that becomes visible by a ground based point at typically about 10° over the horizon (2050 km). In this example UE1 is located at a distance of 2050 km to SAT.sub.n,m whereas UE2's distance is 781 km.

[0062] In the example, the time between a satellite appearing at the horizon and the same satellite disappearing on the opposite side is 9 minutes. It is apparent from FIG. 1 that the distance between a ground based UE and a satellite changes significantly, and therefore also the path loss and latency, within these 9 minutes in an essentially predictable way.

[0063] FIG. 2 shows a similar example setup with two orbits (orbit n and orbit n+1), where the index n iterates all the orbits a satellite radio access network may comprise, which is typically 6. On each orbit, only two satellites are shown (index m and m+1, respectively) where typically 11 satellites are present on the full 360°. The nearest satellites on neighbouring orbits may be offset by half the satellite distance on one orbit so that UEs that reside on the ground at a point between the orbit planes may be served by satellites of alternating orbits.

[0064] The setup of FIGS. 1 and 2 is an example similar to a LEO satellite based system currently deployed. The current invention is as well valid for other setups with different number of satellites, different number of orbits, different inclination of orbits, different height and satellite speed etc.

[0065] FIG. 3 depicts a diagram showing an example for an expected received signal strength (RSS) in dBm over the time in seconds during a single flyover of a satellite (9 minutes) shown for different distances to the orbit plane. The depicted time is the phase time t.sub.P, which is “0”, when the satellite becomes visible by a ground based UE, which is assumed to be when the satellite appears about 10° over the horizon. The phase time is 540 seconds, when the satellite disappears, i.e. is below 10° over the horizon.

[0066] The figure assumes a geometry according to FIG. 1 and the following formula for the RSS (P.sub.RSS).

P.sub.RSS=P.sub.0−10*n*log.sub.10(r/R),  (2) [0067] where [0068] r is the current UE-to-satellite-distance according to the geometry in FIG. 1, [0069] R=2050 km is the max UE-Satellite-distance with maintained communication link, [0070] n=3 is the path loss exponent (n=2 is vacuum line of sight, n=4 is typical urban environment) [0071] P.sub.0 is the minimum RSS detectable by a typical UE, here set to −113 dBm

[0072] For the LEO satellite examples used throughout this invention, FIG. 4 shows the Doppler frequency f.sub.Doppler in kHz for a flyover period for a ground-based UE exactly on the satellite orbit plane (solid line) and for a UE which is 1000 km apart from the satellite orbit plane (dashed line). The Doppler frequency is shown for a carrier frequency of 3 GHz, other carrier frequencies would show linear deviations with respect to FIG. 4. As apparent, the Doppler frequency in this example has maximum values of about 130 kHz with respective opposite signs at the beginning and the end of a flyover period. In the phase of a steep angel between UE and satellite (i.e. the Satellite is in UE's zenith), the Doppler frequency falls quickly from +60 kHz to −60 kHz.

[0073] It is possible to obtain an estimate of the trip time from a measurement of reference signals and these can be compared against expected values. The principle to obtain the initial trip time is to perform one or multiple measurements of reference signals transmitted by the satellite and to compare these measurements with the known expected values, which were stored in the UE previously. The comparison will deliver the best matching curve and therefore an estimation of the distance of the UE to the orbit plane, i.e. to UE's projection point in the orbit plane (cf. FIG. 6), D.sub.U-PP and the current phase time T.sub.P. From these values, the trip time is derived. The details about how this is done are explained below.

[0074] In a first aspect of the current invention the estimation is based on RSS measurements from multiple satellites and some knowledge of relative positions of the measured satellites. Satellites may broadcast an identity information, e.g. in their system information, that provides indication of an order of the satellites on an orbit (index m in FIG. 2) and/or the orbit, the respective satellite belongs to (index n in FIG. 2). The comparison of two measurements performed by a UE on distinct downlink signals of different satellites that are known to be in different orbits shifted by one half flyover period (or 4.5 minutes in our LEO satellite examples) or any other fraction may then lead to a good estimation of the position of either of the satellites. Of course, real measurements may lead to more exact time values than the estimated ones explained above, e.g. seconds or fractions thereof.

[0075] The first aspect above may be enhanced by measuring RSS differences of two or more successive measurements with a certain timely distance. A satellite of rising RSS can be expected in the first half of its flyover and a degrading RSS may point to the second half. More precisely, the exact difference together with the absolute value may contribute a precise estimation of the phase time and the orbit plane distance.

[0076] In a second aspect of the invention the estimation is based on measurement of the Doppler frequency, i.e. measurement of the frequency deviation between the received signal and the transmitted signal caused by the relative speed of satellite and UE. Using the Doppler frequency is beneficial, as it is resistant against environmental effects like fast fading and scattering.

[0077] Using a Doppler frequency measurement is a good indication of the phase time and therefore of the relative position of the UE and the satellite. Using multiple Doppler frequency measurements to derive a Doppler frequency difference would even increase the position estimation accuracy.

[0078] Combining the two described methods of the first and second aspect above using RSS and Doppler frequency may further increase accuracy. One way of combination would be to firstly make two or more successive measurements of the Doppler frequency f.sub.doppler and calculate the ratio of inclination Δf.sub.doppler thereof, and at the same time make two or more successive measurements of the RSS. The values f.sub.doppler and Δf.sub.doppler are then compared with the curves of the expected Doppler frequencies for the different orbit plane distances. The curve with the smallest deviation to the measured inclination is selected, and the current phase time T.sub.P is taken from the intersection point of the selected curve and the measured value f.sub.doppler.

[0079] As example, a measured value f.sub.doppler=100 kHz is shown in FIG. 4 (shown as circles) and a relating Δf.sub.doppler (shown as straight solid lines through the circles). It could be seen, that the curve for 800 km orbit plane distance matches best to the measured Δf.sub.doppler. The relating phase time is T.sub.P1=110 s.

[0080] In the next step, an average value RSS.sub.avg is derived from the multiple measured RSS values. RSS.sub.avg is then compared with the curves of the expected RSS for different orbit plane distances at the derived phase time T.sub.P1=110 s. This is depicted in FIG. 3, whereas RSS.sub.avg=−111 dBm is assumed in this example. The value is between the curves for 800 km and 1200 km orbit plane distance. Therefore, the current orbit plane distance D.sub.U-PP is derived by linear interpolation as 1000 km.

[0081] Once the phase time and the distance to the orbit plane are derived, the distance UE-to-satellite D.sub.U-S and therefore the trip time can be calculated.

[0082] The distance UE-to-satellite could be calculated as:

D.sub.U-S=√{square root over (D.sub.S-PP.sup.2+D.sub.U-PP.sup.2)}  (3)

wherein [0083] D.sub.U-S: The distance UE-to-satellite [0084] D.sub.S-PP: The distance from the satellite to UE's projection point on the orbit plane [0085] D.sub.U-PP: The distance from the UE to UE's projection point on the orbit plane

[0086] The value for D.sub.U-PP is derived by the UE as described above and the value D.sub.S-PP is derived from calculations using the phase time and the satellites constellation. The course of D.sub.S-PP for LEO satellites is depicted in FIG. 5. The UE is either pre-configured with the calculation formula and the required parameters (orbit period, orbit altitude number of satellites per orbit, etc.) or the related values of D.sub.S-PP for multiple phase times are pre-configured.

[0087] The trip time can be calculated as

[00001]$\begin{array}{cc}{t}_{\mathrm{trip}}=\frac{D_{U-S}}{c_l}& \left(4\right)\end{array}$

wherein
t.sub.trip: Is the trip time of signals transmitted from the UE to the satellite
D.sub.U-S: The distance from the UE to the satellite
c.sub.l: Is the speed of light that equals 299,792,458 m/s in vacuum

[0088] There may be situations, where a reliable trip time cannot be derived by the UE. In this case the UE will detect such situation and will perform a different random access procedure, as described below.

[0089] After the UE has derived the initial trip time, it can simply derive the trip time of any time instance in the future. Therefore, the UE stores the latest derived trip time t.sub.trip,latest, the time stamp T.sub.latest, where this trip time was valid, the latest distance to the orbit plane D.sub.U-PP, and The value n of the current orbit. The value n is derived e.g. from the satellite itself.

[0090] For the calculation of the future trip time, the related future phase time is required. Therefore, the UE firstly calculates the timely difference towards the latest and the future time instance and adds the latest phase time T.sub.P,latest. Then the remainder of an Euclidean division (also known as “modular arithmetic”) is calculated with mod ΔT.sub.sat, i.e. the timely distance of two successive satellites from the same orbit which is 540 seconds in case of Iridium satellites.

T.sub.P,new=(T.sub.future−T.sub.latest+T.sub.P,latest+T.sub.correct)mod ΔT.sub.Sat  (5)

[0091] The value T.sub.correct is used to consider different timings of satellites from different orbits. T.sub.correct=0, if the current satellite is on the same orbit as the satellite used for latest trip time derivation. It is ½ΔT.sub.Sat, if the new satellite is on a neighbouring orbit (n+1 or n−1) and neighbour orbits are shifted by half a flyover period, and it is in general:

T.sub.correct=T.sub.shift×(n.sub.latest−n.sub.new)ΔT.sub.Sat  (6)

wherein [0092] T.sub.shift: is the timely offset of satellites from neighboured orbits as perceived by a UE [0093] n.sub.latest: is the orbit number of the satellite of which the latest trip time was stored in the UE [0094] n.sub.new: is the orbit number of the satellite of which a new trip time should be derived
With the new phase time and the stored latest orbit plane distance, the UE derives the new trip delay as described above.

[0095] This calculation leads to useful results, as far as UE's movement since the last trip time estimation is small. Or, more precisely, the trip time difference between the latest and the new derivation resulting from UE's movement is smaller than the preamble guard time. This is valid in most cases. For example, in case the guard interval of 0.684 ms is configured, the calculated trip time is usable at least for a horizontal movement up to 400 km.

[0096] Preamble transmissions with wrong transmit timings may lead to additional preamble collisions. To avoid transmissions with invalid trip times, the UE may use means to detect, whether a calculation of a future trip time is reliable. This is done e.g. by defining a maximum age of the latest trip derivation, e.g. 18 min. The expiration of this maximum age will require a new measurement as described above to derive the current trip time.

[0097] The methods described above assume usage of the slotted ALOHA principle for random access, because this will most likely also be used for satellite communications to the 3GPP network. In case that a correct reception timing within the access slot boundaries cannot be guaranteed by a given UE, the number of successful random access attempts will be degraded for all UEs in the related cell or random access would not be possible at all.

[0098] Therefore, it is one aspect of this invention to detect situations, were the trip time cannot be derived reliable. This may be e.g. if the RSS has very low values that are sufficient to connect to the satellite but too low to derive the trip time prior transmission of the random access preamble. The UE may detect this situation e.g. by comparing the currently measured RSS with a configurable threshold. In case the RSS is below this threshold, the trip time will not be derived by the UE. Instead the UE will use a different type of random access procedure, which is not based on the slotted ALOHA principle and which is not requiring the preambles to be received within access slot boundaries. I.e. a timing correction about the trip time is not required. Pure ALOHA is a usable principle. Compared to slotted ALOHA, the throughput of pure ALOHA is smaller, but still better than using slotted ALOHA where preambles will arbitrarily overlap with access slot boundaries. Therefore, it is beneficial to use slotted ALOHA, if preamble reception within the access slot boundaries could be guaranteed, and to use pure ALOHA, if the reliability of synchronous reception is too low.

[0099] When simultaneously applying different types of random access, it is important, that the receiver of the random access preambles (i.e. the satellite or the base station on ground) is aware, which type of random access is applied by the UE. Therefore, a first proposal is to use two different sets of random access preambles. A first set is used for slotted ALOHA mode, and a second set, for unsynchronised mode (e.g. pure ALOHA). The second set could either be derived by dividing the already specified preambles in two groups, or by defining new preambles for the second group. The UE can easily indicate the used random access type, by selecting a preamble from the relating preamble set. An alternative second proposal is to allocate separate resources in the time-frequency-domain for the two different random access approaches, e.g. different sub-carriers for slotted and non-slotted ALOHA.

[0100] As described above, using a common configuration for the random access response interval will lead to an unnecessary long delay of the random access procedure. Therefore, this invention proposes to derive the response interval individually for each UE from the current trip time. Additionally, the cell-specific configuration from the base station is considered. The beginning of the response interval in slotted ALOHA mode T.sub.RAR,start,slotted is calculated to be two times the trip time after the transmission of the relating preamble has ended:

T.sub.RAR,start,slotted=T.sub.TX+2X t.sub.trip+T.sub.preamble  (7)

wherein [0101] T.sub.Tx: is a time where the transmission of the related preamble has begun [0102] T.sub.preamble: is a duration of the related preamble

[0103] The duration of the response interval is than taken from the configuration parameter as transmitted previously by the base station.

[0104] If a reliable trip time is not available and the UE has therefore selected the unsynchronised random access mode, the start of the response interval T.sub.RAR,start,unsynchronised is calculated from a parameter T.sub.RAR,delay transmitted by the base station:

T.sub.RAR,start,unsynchronised=T.sub.TX+T.sub.delay+T.sub.Preamble  (8)

wherein [0105] T.sub.delay: is a time offset between transmission of a random access preamble and the beginning of the random access response window. This value is transmitted by the base station to the UE. It is selected by the base station to be at least 2 times the maximum trip time UE-to-satellite, e.g. 14 ms in case of LEO satellites.

[0106] Also in this case, the duration of the response interval is taken from the configuration parameter as transmitted previously by the base station.

[0107] In the following examples, a radio interface according to the LTE standard is assumed. This assumption is not to be understood as limiting the invention. The invention is also applicable for future mobile communication standards like 5G as specified at 3GPP.

[0108] According to the main aspect of this invention, UE1 is configured to derive a current signal trip time towards the satellite currently used for communication. UE1 is located within the coverage area of satellite Sat.sub.n,m (cf. FIG. 1). It has no active connection, i.e. it is in idle mode. The UE is configured to perform the random access procedure for connections towards satellites. It has stored values for the expected RSS and Doppler frequencies for several phase times and several orbit plane distances. Further it is configured with thresholds to detect unreliable measurements and to detect invalidity of stored trip times.

[0109] Step 1

[0110] UE1 is switched on and performs a cell search procedure. It selects Sat.sub.n,m to connect to the network. From the received signals transmitted by the satellite, the UE obtains the information that the link will be provided by a satellite. Further the UE has derived the random access configuration from the broadcasted system information. In addition to the parameters already specified for LTE, the satellite transmits parameters that define a second set of random access preambles to be used for unsynchronised random access.

[0111] Step 2

[0112] UE1 wants to perform a random access to Sat.sub.n,m. As the link is provided via satellite, the UE will derive the trip time prior to preamble transmission. No stored value of the trip time is available in this example. Therefore, the UE starts measurement as described above to obtain the trip time.

[0113] Step 3

[0114] The UE selects a random access preamble and an access slot from the respective set for slotted-ALOHA mode. It calculates the transmit time using equation 1.

[0115] Step 4

[0116] The UE transmits the random access preamble and listens to the response. As the UE selected the slotted ALOHA mode, it calculates the beginning of the random access response interval T.sub.RAR,start,slotted as described above, i.e. it considers the current trip time. The duration of the interval is derived from the configured parameter as transmitted previously by the base station.

[0117] Step 5

[0118] The satellite receives the preamble. It calculates the time offset of the reception time from the access slot boundary and calculates a timing advance (TA) value. The behaviour is identical to a terrestrial base station. It sends a random access response message back to the UE including the TA value.

[0119] Step 6

[0120] The UE receives the response within the calculated time window and applies the TA value for the next transmission. Further it corrects the latest trip time value with the received TA value and stores the corrected value including a time stamp.

[0121] Step 7

[0122] Now the UE finalises the random access as usual for LTE.

[0123] The following steps may be used for a future trip time calculation:

[0124] Step 1′

[0125] UE1 is in idle mode and Sat.sub.n,m+1 will be used for random access. The UE wants to perform a random access now, e.g. in order to transmit a Tracking Area Update (TAU).

[0126] Step 2′

[0127] As this is a satellite connection, the UE will derive the trip time prior preamble transmission. It uses the latest stored value and derives the current trip time as described above.

[0128] Step 3′

[0129] The UE performs the random access and proceeds with step 4 described in the message flow above followed by steps 5-7.

[0130] The following steps may be used for an unsynchronised random access:

[0131] Step 1″

[0132] UE1 is in idle mode and Sat.sub.n,m+1 will be used for random access. The UE wants to perform a random access now.

[0133] Step 2″

[0134] As this is a satellite connection, the UE will derive the trip time prior preamble transmission. No stored value of the trip time is available in this example. Therefore, the UE starts measurement as described above to obtain the trip time. It detects, that the measurements will not be suitable for a reliable calculation of the trip time, e.g. because the measured RSS is below the configured threshold.

[0135] Step 3″

[0136] The UE will select the unsynchronised random access mode and selects a preamble and an access slot as configured for this mode.

[0137] Step 4″

[0138] The UE transmits the random access preamble and listens to the response. As the UE selected the unsynchronised random access mode, it calculates the beginning of the random access response interval T.sub.RAR,start,unsynchronised as described above, i.e. it considers the parameter T.sub.delay, which was configured by the base station previously. The duration of the interval is derived from another parameter as transmitted previously by the base station.

[0139] Step 5″

[0140] The satellite receives the preamble. It calculates the time offset of the reception time from the latest access slot boundary and calculates a timing advance (TA) value. It sends a random access response message back to the UE including the TA value and the number of the access slot. The timing of this transmission after reception of the preamble is a fix value T.sub.delay pre-configured in the UE and to be used exclusively for the unsynchronised random access method.

[0141] Step 6″

[0142] The UE receives the response within the calculated time frame and adjusts the timing of the next transmissions according to the TA value and the access slot number. Further, it derives the trip time from the reception time and the transmission time. It has to be corrected about the value T.sub.delay with the following equation:

t.sub.trip=½(T.sub.RX−T.sub.TX−T.sub.delay)  (9)

[0143] As one result from the availability of a reliable trip time, the UE may use the slotted random access method for future random access attempts, at least for a specified time.

[0144] The following steps may be used to remove stored trip time values which becomes invalid:

[0145] Step 1*

[0146] The UE is in idle mode and Sat.sub.n,m+1 will be used for random access. The UE wants to perform a random access now.

[0147] Step 2*

[0148] As this is a satellite connection, the UE will derive the trip time prior preamble transmission. It reads the latest stored value and detects, that the validity of this value is expired, i.e. due to the time stamp the value is older than allowed by the current configurations.

[0149] Step 3*

[0150] The UE will not use the stored value. Instead it deletes them and will derive a new trip time value from new measurements as described in steps 1 to 7 above.