Signal detection in the presence of the Doppler effect

Abstract

Method (200) for detecting a message in a Doppler-affected radio signal. The message comprises a Zadoff-Chu ZC sequence followed by a pseudo-random PN sequence. The method comprises: a first cross-correlation (201) between the signal and a local replica of the ZC sequence; a determination (202) of an approximate start time of the ZC sequence in the signal as a function of a maximum of the first correlation; a transformation (203) of the received signal by multiplying it by a conjugate sequence of the ZC sequence; a transformation (204) of a local replica of the ZC sequence followed by the PN sequence by multiplying it by the conjugate sequence; a second cross-correlation (205) between the transformed signal and the transformed local replica; a determination (206) of a precise start time of the ZC sequence as a function of a maximum of the second correlation. Figure for the abstract: Fig. 9

Description

Title of the invention: Signal detection in the presence of the Doppler effect Scope of the invention

[0001] The present invention relates to the field of telecommunications, and more particularly to the detection of a radio signal in the presence of the Doppler effect. The invention is particularly well suited for establishing communication between a ground terminal and a satellite belonging to a constellation of satellites orbiting the Earth, especially for a constellation in low Earth orbit. State of the art

[0002] In the field of IoT, connected objects generally communicate via terrestrial networks. However, the coverage of these terrestrial networks is not universal, and there may be "white zones" due to geographical, economic, or technical limitations that make it difficult to install the terrestrial infrastructure necessary to provide adequate network coverage.

[0003] This is why solutions have been devised to provide satellite access networks for these IoT communication systems. These satellite access networks generally rely on constellations of satellites in low Earth orbit (LEO). However, protocols initially developed for terrestrial use, such as LoRaWAN or NB-IoT, require relatively significant modifications to be compatible with satellite communication. These modifications result in increased complexity and implementation cost.

[0004] In particular, these protocols generally require continuity of the radio link between the object and a ground network server. The satellite must therefore be in simultaneous communication with the object and with a ground communication station, possibly through other satellites via inter-satellite links (ISL for "Inter-Satellite Link".

[0005] The LoRaWAN protocol relies on random access of the device to the communication medium. If there is no satellite within range of the device when the device attempts to establish communication with the satellite access network, communication cannot be established. The device must then repeat its connection attempt until successful. This results in higher energy consumption by the device and an increase in its cost.

[0006] The proposed adaptations for the NB-IoT protocol to support satellite communications require uninterrupted coverage, such that objects are always able to communicate with the network. In other words, these adaptations require that an object be constantly within range of at least one satellite in the satellite access network. This implies a larger number of satellites in the constellation, which increases the overall cost of the system.

[0007] On the other hand, the use of low Earth orbit satellites introduces a significant Doppler effect due to the satellites' speed. This results in substantial signal shifts in the frequency spectrum, exceeding what is originally supported by terrestrial protocols. To adapt the NB-IoT protocol to satellite communications, it is proposed to use multibeam antennas to compensate for the Doppler effect corresponding to the different beam directions. This necessitates the use of larger and more expensive satellites. Another solution relies on estimating and pre-compensating for the frequency shift due to the Doppler effect based on the objects' knowledge of their position and the positions of the network satellites. This solution considerably increases the complexity and cost of the objects.

[0008] Solutions based on geostationary satellites have also been proposed to provide global coverage, but these solutions are much more expensive in terms of the cost of the satellites and their launch. On the other hand, these solutions offer significantly more limited capabilities than solutions based on low Earth orbit satellites.

[0009] There are various solutions for detecting the start time of a message carried in a Doppler-affected signal, and for estimating the frequency shift affecting the signal. For example, it is known to form a preamble to the message by juxtaposing at least two known sequences, for example, Zadoff-Chu sequences. By cross-correlating the received signal with each sequence independently, and then combining the results obtained, it is possible to estimate the Doppler shift and improve the reliability of the detection.

[0010] Some solutions are based on the juxtaposition of a Zadoff-Chu sequence with the conjugate of that sequence. Some solutions rely on a frequency analysis of the Zadoff-Chu autocorrelation, or on a frequency analysis of the modulation of a cosine by a Zadoff-Chu sequence, or on the use of a sliding differential correlation.

[0011] These different solutions, however, assume that the receiving device implementing the detection can withstand relatively high computational complexity. Each sequence must be long enough for the results of the correlation associated with that sequence to exhibit satisfactory accuracy. On the other hand, the The juxtaposition of two such sequences leads to an increase in the length of the message to be detected, and therefore to an increase in the use of radio resources necessary for the transmission of this message.

[0012] There are also solutions based on a single pseudo-random sequence to detect the start time of a message carried in a Doppler-affected signal. Again, these solutions generally require a particularly high computational complexity for the receiving device. Description of the invention

[0013] The present invention aims to remedy all or part of the drawbacks of the prior art, in particular those set out above.

[0014] To this end, and according to a first aspect, a method is proposed for detecting a message carried in a radio signal emitted by a transmitting device, the signal being affected by a Doppler effect due to the relative motion between the transmitting and receiving devices. The message comprises a Zadoff-Chu ZC sequence followed by a pseudo-random PN sequence. The method comprises: - a first cross-correlation between a signal received by the terminal and a local replica of the ZC sequence, - a determination of an approximate starting time for the ZC sequence in the signal as a function of a maximum absolute value of the first correlation, - a transformation of the received signal by multiplying it by a conjugate sequence of the ZC sequence, - a transformation of a local replica of the ZC sequence followed by the PN sequence by multiplying said replica by the conjugate sequence, - a second cross-correlation between the transformed received signal and the transformed local replica, - a determination of a precise moment of start of the ZC sequence in the signal as a function of a maximum of absolute value of the second correlation.

[0015] Such arrangements allow the detection of a message in a signal affected by Doppler, with a relatively low computational complexity for the receiving device, and while limiting the length of the preamble in the message to be detected.

[0016] The expression "relative movement between the emitting device and the receiving device" covers the following different cases: the emitter and the receiver are both in motion, the emitter is fixed and the receiver is in motion, the receiver is fixed and the emitter is in motion.

[0017] In particular embodiments, the detection method may further comprise one or more of the following characteristics, taken individually or in all technically possible combinations.

[0018] In particular embodiments, the method includes an estimation of a Doppler frequency shift affecting the received signal as a function of a difference between the precise time and the approximate time, and as a function of a parameter which depends on the ZC sequence used.

[0019] In particular embodiments, the estimation of the Doppler frequency shift is refined as a function of a phase shift observed on the received signal over time. The phase shift is calculated as the difference between a phase measured for a beginning section of the set formed by the ZC sequence and the PN sequence and a phase measured for a ending section of this set.

[0020] In particular embodiments, at least one of the first cross correlation and the second cross correlation is followed by an interpolation step of the correlation results to optimize the accuracy of the determination of the maximum absolute value of the correlation.

[0021] In particular embodiments, the PN sequence has fewer symbols than the ZC sequence.

[0022] In particular embodiments, the number of symbols in the PN sequence is at least twice as small as the number of symbols in the ZC sequence.

[0023] In particular embodiments, the "Ping" message successively comprises the ZC sequence, the PN sequence, and at least one block formed by a sequence of useful data symbols and a sequence of pilot symbols, and the method comprises at least a third cross-correlation between the received signal and a local replica of the pilot symbol sequence.

[0024] In particular embodiments, the message is a "Ping" message transmitted by the transmitting device repeatedly on a plurality of different carrier frequencies comprising a center frequency f used as the reception frequency by the receiving device during radio listening periods to detect a "Ping" message, and 2N offset frequencies f. - f£ + i- A y, i being a non-zero integer index ranging from -N to N. The frequency offset A f is chosen to be less than or equal to twice a maximum frequency offset A supported by the receiving device for message detection. N is a strictly positive integer chosen such that ( \ , where Dmax is the absolute value of a maximum frequency difference due to the Doppler effect for a signal emitted by the transmitting device to the receiving device.

[0025] According to a second aspect, the invention relates to a computer program product comprising program code instructions which, when implemented by a processor, configures said processor to implement a detection method according to any one of the implementation modes mentioned above.

[0026] According to a third aspect, the invention relates to a receiving device comprising a processor configured to implement a detection method according to any one of the previously mentioned implementation modes.

[0027] According to a fourth aspect, the invention relates to a satellite communication system comprising at least one ground terminal and a constellation of satellites forming a satellite access network. Each satellite in the constellation is configured to repeatedly transmit a "Ping" message signaling the satellite's presence to a terminal within range of the satellite. The terminal includes a processor configured to implement a method for detecting a "Ping" message according to any one of the aforementioned implementation modes.

[0028] In particular embodiments, the ZC sequence corresponds to a satellite access network identifier.

[0029] According to a fifth aspect, the invention relates to a method for establishing communication between a ground terminal and a satellite belonging to a constellation of satellites orbiting the Earth. The satellite constellation belongs to a satellite access network of a communication system. Each satellite in the constellation is configured to repeatedly broadcast a "Ping" message signaling the satellite's presence to a terminal within range of the satellite. The method comprises: a. Reception by the terminal, from a satellite in the constellation, of a message containing a list of future times when a satellite in the constellation will pass within range of the terminal, b. a selection, by the terminal, based on a desired time to establish communication with the satellite access network, of a future time for a satellite to pass over the Earth from among all the future times for satellites to pass over the Earth included in the list, c. determining a wake-up time for the terminal based on the selected future transition time, d. a transition of the terminal into a "sleep" mode until the moment of awakening, e. when the wake-up time is reached, the terminal switches to an "attentive" mode in which the terminal repeatedly listens for radio signals to detect a "Ping" message from the satellite, f. a detection of a "Ping" message by the terminal using the detection method according to any one of claims 1 to 8 and, when a "Ping" message is detected, an emission of a "Pong" message, by the terminal, to the satellite. Presentation of the figures

[0030] The invention will be better understood upon reading the following description, given by way of non-limiting example, and made with reference to the following figures:

[0031] [Fig-1] an illustration of a communication system according to the invention, to a the moment when the terminal receives a list of future times when a satellite in the constellation will pass overhead.

[0032] [Fig.2] an illustration of a communication system according to the invention, at a moment when the terminal establishes communication with a satellite of the constellation,

[0033] [Fig.3] a schematic representation along a time axis of the main steps of a method according to the invention for establishing communication between a terrestrial terminal and a satellite,

[0034] [Fig.4] a sequence diagram illustrating the establishment of communication between a terrestrial terminal and a satellite,

[0035] [Fig. 5] another schematic representation of the main steps of a method according to the invention for establishing communication between a terrestrial terminal and a satellite,

[0036] [Fig.6] a first example of the implementation of a temporal spreading of random accesses of the terminals of the communication system,

[0037] [Fig.7] a second example of the implementation of a temporal spreading of random accesses of the terminals of the communication system,

[0038] [Fig.8] an example of implementing a recurrent broadcast of a "Ping" message on a plurality of different carrier frequencies,

[0039] [Fig.9] a schematic representation of the main steps of a method according to the invention for detecting a signal affected by Doppler,

[0040] [Fig. 10] an example of an implementation of a "Ping" message to be detected.

[0041] In these figures, identical reference numerals from one figure to another designate identical or analogous elements. For clarity, the elements shown are not necessarily to the same scale, unless otherwise stated. Detailed description of the invention

[0042] Figures 1 to 5 illustrate the main steps of a method for establishing communication between a terrestrial terminal and a satellite in a communication system.

[0043] As illustrated in Figures 1 and 2, we are considering a communication system 10 comprising at least one terrestrial terminal 11 and a constellation of satellites 12 in orbit 17 around the Earth 18. The constellation of satellites 12 belongs to a satellite access network for the terminals 11 of the communication system 10. These are, for example, nanosatellites in low Earth orbit (LEO).

[0044] In the example illustrated in Figures 1 and 2, N satellites are considered to be positioned in the same orbital plane of the constellation. Figure 1 represents three satellites SI, S2, and SN of this orbital plane at a time close to a time t1 at which satellite SI is within range of terminal 11 (meaning that terminal 11 and satellite SI are visible and sufficiently close to each other to establish radio communication). Figure 2 represents three satellites S3, S4, and S5 of this orbital plane at a time close to a time t4 later than time t1, at which satellite S4 is within range of terminal 11. This assumes a relatively small number of satellites 12 in the constellation, such that terminal 11 is not constantly within range of a satellite 12 in the constellation.

[0045] As illustrated in Figures 1 and 2, the communication system may also include a ground gateway station 15 connected to a terrestrial server 13 via a core network 14 (for example, the Internet). The communication system 10 may, in particular, allow data exchange between the server 13 and the terminal 11 via satellite.

[0046] By way of non-limiting example, the communication system 10 is a low-power wide area network (LPWAN) of the IoT or M2M type. However, nothing would prevent the consideration of other types of communication systems in variants.

[0047] In the example considered, terminal 11 corresponds to a connected object located at Earth's surface 18. However, it should be noted that the application could also apply to a "quasi-terrestrial" terminal 11, such as a connected object embedded in a stratospheric communication platform (HAPS, for "High Altitude Platform Station"). The term "terrestrial terminal" should be understood in this application as including "quasi-terrestrial" terminals, i.e., terminals located in the lower layers of the atmosphere, for example at altitudes below 50 km.

[0048] Terminal 11 includes a processing circuit comprising one or more processors and electronic storage means in which a computer program product is stored, in the form of program code instructions to be executed to implement the communication method 100. Terminal 11 also conventionally includes hardware and software means for radio communication (antenna, amplifiers, filters, oscillator, analog-to-digital converter, channel (de)coder, (de)modulator, (de)multiplexer, etc.).

[0049] Similarly, each satellite 12 of the constellation includes a processing circuit comprising one or more processors and electronic storage means in which a computer program product is stored, in the form of program code instructions to be executed to implement the communication method 100. Each satellite 12 also conventionally includes radio communication hardware and software means (antenna, amplifiers, filters, oscillator, analog-to-digital converter, channel (de)coder, (de)modulator, (de)multiplexer, etc.).

[0050] In particular, each satellite 12 in the constellation is configured to repeatedly broadcast a "Ping" message 50 signaling the satellite's presence to a terminal 11 within its range. Terminal 11, for its part, is configured to implement steps 101 to 106 of method 100, with the aim of establishing communication with a satellite 12. Steps 101 to 106 are illustrated in Figures 3 and 5.

[0051] Step 101 corresponds to the reception, by terminal 11, of a message from a satellite 12 of the constellation containing a list of future times when a satellite 12 of the constellation will pass within range of terminal 11. This list may, in particular, be transmitted in a "Ping" message. However, nothing would prevent the list from being received in a message other than the "Ping" message (for example, in a downlink traffic data message).

[0052] In the example considered and illustrated in figures 1 to 3, the terminal 11 receives a list 16 of future satellite pass times in a "Ping" message 50 broadcast by the satellite SL. This list 16 of future pass times is received by the terminal 11 at a time close to tl at which the satellite SI is within range of the terminal 11. The list 16 includes the time t2 (which corresponds to a time at which the satellite S2 should be within range of the terminal 11), the time t3 (which corresponds to a time at which the satellite S3 should be within range of the terminal 11), the time t4 (which corresponds to a time at which the satellite S4 should be within range of the terminal 11), etc.

[0053] Each future instant in the list 16 corresponds to a future instant of entry into visibility of a satellite 12 of the constellation with the terminal 11, that is to say a future instant at which another satellite (S2, S3, S4, etc.) of the constellation will take a position on the orbit 17 relatively close to the position of the satellite SI at the time when the list 16 is broadcast (it is assumed that the position of the terminal 11 varies little between the instant of reception of the list and the most distant future instant in the list).

[0054] Future transit times can be defined relative to the time of transmission of list 16 by the SI satellite (which also corresponds approximately to the time of reception of the list by the terminal). Thus, it is not essential to maintain a Time synchronization between terminal 11 and the satellite access network. Future transit times can be defined based on a transmission time of list 16 by satellite SI and a transit period of a satellite at a given position in orbit 17.

[0055] It should be noted that in the example considered and illustrated in Figures 1 to 3, list 16 contains future times of transit of satellites belonging to the same orbital plane of the constellation. However, nothing would prevent list 16 from also containing future times of transit of satellites 12 belonging to different orbital planes.

[0056] Future pass times are predictable because the system 10 knows the orbits 17 of all the satellites 12 in the constellation. A ground control station can provide each satellite with the necessary data so that it can repeatedly determine, for at least a certain period, the list of future satellite pass times that it must broadcast. According to another example, the satellites can know their own position and orbit, as well as those of all the other satellites in the constellation, and use this knowledge to autonomously construct the lists to be broadcast.

[0057] Step 102 corresponds to a selection, by terminal 11, according to a desired time to establish communication with the satellite access network, of a future time of passage of a satellite 12 from among those in list 16.

[0058] The desired time (named "te") can be "as soon as possible". The time te can also correspond to a future time that meets a constraint of recurring message transmission from terminal 11 to the terrestrial server 13. The time te can also be conditioned by the presence of uplink traffic data (UL) to be transmitted.

[0059] The selected future time of passage (named "ts") can correspond to the time in list 16 that is closest to the desired time te. In another example, the time ts can correspond to the time in list 16 that is closest after the time te. In yet another example, the time ts can be selected based on its distance from the time te.

[0060] In the example considered and illustrated in Figures 1 to 3, the selected instant ts corresponds to instant t4 in list 16. This is the closest future transit instant to instant te. At instant t4, satellite S4 should be within range of terminal 11, which will allow terminal 11 to establish communication with satellite S4.

[0061] Step 103 involves determining a wake-up time (denoted "tw") for terminal 11 based on the selected future pass time. The wake-up time tw may correspond to the selected time ts. It may also precede the selected time to account for a margin of uncertainty regarding the actual start of the visibility window 41 of satellite S4 by terminal 11.

[0062] Step 104 corresponds to the terminal 11 entering a "sleep" mode until it wakes up. The "sleep" mode is a mode in which the terminal 11 does not attempt to establish communication with the satellite access network. This is an energy-saving mode for the terminal, since in this mode the terminal does not need to turn on its radio to receive a message transmitted by a satellite or to transmit a message to a satellite.

[0063] Step 105 corresponds to the terminal 11 switching to an "attentive" mode when the wake-up time tw is reached. The "attentive" mode is a mode in which the terminal 11 repeatedly performs radio listening periods 40 to detect a "Ping" message 50 coming from the satellite S4.

[0064] When a "Ping" message 50 is received from the satellite S4, as in the example illustrated in Figures 2 and 3, the method 100 includes a step 106 of transmitting a "Pong" message 51 by the terminal 11 to the satellite S4.

[0065] The "Pong" message 51 establishes communication with the S4 satellite, for example by requesting radio resource allocation for data exchange between terminal 11 and the S4 satellite. The "Pong" message may also optionally contain uplink traffic data directly and be sufficient on its own for communication between terminal 11 and the S4 satellite.

[0066] When no "Ping" message 50 is detected at step 105 for a predetermined duration of the "attentive" mode, the method 100 may include a selection, by the terminal 11, of the next future time of passage of a satellite in the list 16, and a new iteration of steps 103 to 105.

[0067] The "attentive" mode and the "sleep" mode:

[0068] The "attentive" mode of a terminal of the communication system 10 can also to be activated when the terminal does not have a list of future satellite pass times in memory (for example, during its first activation, or following a long period of inactivity, or when the list it has in memory is about to expire), or when the terminal fails to establish communication with a satellite for a predetermined period (this can happen, for example, if the terminal has moved significantly and the list it has in memory is no longer adapted to its new position).

[0069] Generally, the "attentive" mode of a terminal in the communication system 10 is primarily activated when a satellite is likely to be within range of the terminal 11. The purpose of the "attentive" mode is indeed to allow the detection of a satellite 12 while conserving the energy of the terminal 11. When the "attentive" mode is activated, the terminal 11 periodically listens to the transmission channel and searches for a "Ping" signal for a certain duration corresponding to a predetermined probability of detection. The duration of a radio listen 40 is at least as long as The sum of the transmission time and the repetition period of a "Ping" message is used to ensure that a radio interception 40 completely overlaps with a transmission of a "Ping" message. The time interval between two radio interceptions 40 is shorter than the average duration of a satellite visibility window 41, in order to ensure at least one interception per satellite pass. However, the time interval between two radio interceptions 40 is long enough to conserve the power of the terminal 11 and to allow the first radio interception 40 of a visibility window 41 to occur relatively late in that window. For an average visibility duration of four minutes, for example, a radio interception 40 period can be opened approximately every three minutes.

[0070] As illustrated in [Fig. 6], upon activation of the "attentive" mode, the delay before the first radio ping 40 can advantageously be randomly defined with a uniform distribution relative to a predetermined maximum delay (this maximum delay can correspond to the time interval separating two radio pings 40). According to another example, and as illustrated in [Fig. 7], it is the wake-up time tw that can be randomly determined within a predetermined time window relative to the selected future time ts (and in this case, the first radio ping 40 can occur immediately upon switching to "attentive" mode). Such arrangements make it possible to optimize the temporal distribution of the "Pong" messages transmitted by the terminals 11 of the communication system 10. Indeed, if terminals are geographically grouped, they risk sharing the same satellite visibility and thus performing their detection simultaneously.This would then increase the probability of collision between the "Pong" messages sent by these terminals.

[0071] The "sleep" mode is activated at the end of a successful communication with a satellite, or after a predetermined period if no communication could be established. This mode saves energy when no communication is possible or necessary. A time margin (for example, on the order of a few tens of seconds) can be taken into account by terminal 11 to determine the next wake-up time tw, depending, for example, on the accuracy of the terminal 11 clock and / or on the accuracy of predictions of future satellite passes.

[0072] Random access:

[0073] The "Pong" message 51 corresponds to a random access of the terminal 11 issued in response to the detection of the "Ping" message 50. The "Pong" message is issued within a time interval that is, for example, defined relative to the start of the "Ping" message. Several terminals 11 of the communication system 10 could simultaneously issue a "Pong" message. To limit the risk of collision between several "Pong" messages, the time of emission and the frequency of emission of a message " "Pong" can be randomly determined by terminal 11 within a predetermined time interval and frequency band.

[0074] Allocation of uplink or downlink radio resources:

[0075] In the example considered and illustrated in Figures 3 and 4, the "Pong" message 51 corresponds to a request for resource allocation for uplink (UL) and downlink (DL) traffic data. However, nothing would prevent us from considering other examples in which the "Pong" message aims to obtain radio resources in only one direction (uplink or downlink), or in which the "Pong" message directly contains uplink traffic data and does not aim to obtain other radio resources. The "Pong" message 51 may, in particular, contain an identifier of the terminal 11 that transmits this message.

[0076] An "Allocation" message 52 is transmitted by the satellite S4 in response to the "Pong" message. The "Allocation" message 52 lists the time and frequency resources allocated to terminals 11 that are within range of the satellite S4 and that have requested allocation. The "Allocation" message 52 is, for example, transmitted in a frequency band centered around the transmission frequency of the "Ping" message 50, at a specific time relative to the start of the "Ping" message 50.The "Allocation" message 52 contains, for example, the list of identifiers of the terminals 11 for which a "Pong" message 51 has been received, each identifier being associated with all or part of the following information: a transmission time allocated to the terminal, a time and frequency to be used for transmitting a message 53 containing upstream traffic data, a time and frequency to be used for receiving a message 54 containing downstream traffic data, a time and frequency to be used for receiving an upstream acknowledgment message 55 (ACK UL), and a time and frequency to be used for transmitting a downstream acknowledgment message 56 (ACK DL). The transmission times are, for example, defined relative to a reference time in the "Allocation" message 52. The transmission frequencies are, for example, defined relative to the transmission frequency of the "Pong" message.The reception frequencies are, for example, defined relative to the reception frequency of the "Allocation" message 52. In the case of a "full duplex" system, the upstream traffic data message 53 and the downstream traffic data message 54 can be exchanged simultaneously (i.e., with a time overlap).

[0077] The uplink traffic data carried in message 53 is, for example, intended to be transmitted to the terrestrial server 13. As illustrated in [Fig. 2], the S4 satellite can be configured to store the uplink traffic data until it can transmit it to the gateway station 15 connected to the terrestrial server 13 via The core network 14. Here, we assume that there is no line of sight between satellite S4 and a ground gateway station at the time the communication takes place between terminal 11 and satellite S4. However, in a variant, nothing would prevent satellites 12 from supporting inter-satellite links (ISL for "Inter-Satellite Link") to allow for rapid transmission of uplink traffic data to the ground server 13.

[0078] The downlink traffic data carried in message 54 originates, for example, from the ground server 13. This downlink traffic data must be in the memory of satellite S4 when it enters line of sight with terminal 11. This downlink traffic data was, for example, transmitted to satellite S4 by a ground gateway station before satellite S4 entered line of sight with terminal 11. To achieve this, the following steps were implemented by method 100 before the establishment of communication between terminal 11 and satellite S4: - a determination of an approximate geographical position of terminal 11 based on the position of a satellite 12 of the constellation (for example, satellite SI) at a recent time of reception by that satellite of a message transmitted by terminal 11 (for example, at a time close to tl), - an identification, based on the approximate geographical position of terminal 11, of satellite S4 as being soon within range of terminal H, - a transmission of downstream traffic data intended for terminal 11 from a ground gateway station to the S4 satellite.

[0079] Indeed, a gateway station must know which satellite to transmit data to so that it can then be transmitted to a terminal. To this end, each time a satellite communicates with a terminal, the satellite records its position. The positions thus recorded by the satellites are transmitted to the ground during communications with gateway stations, and a database of the approximate positions of the terminals can then be created. When a message destined for a terminal enters the system, the terminal is searched in the database, and a list of satellites that will soon pass near the terminal can be defined. The downlink traffic data can then be integrated into the data to be loaded into the memory of these satellites during their next communication with a gateway station.Terminals attempt to detect satellites and respond to "Ping" messages regularly, even when they have nothing to transmit, in order to inform the satellites of their presence. A satellite carrying downlink traffic data destined for a terminal that sends it a "Pong" message allocates radio resources on the downlink to that terminal so that it can transmit the intended data.

[0080] Acknowledgment mechanism:

[0081] As illustrated in Figures 3 and 4, an acknowledgment mechanism may be provided, in which satellite S4 sends an acknowledgment message 55 to the terminal when an uplink traffic data message 53 transmitted by terminal 11 is correctly received by satellite S4, and / or in which terminal 11 sends an acknowledgment message 56 to satellite S4 when a downlink traffic data message 54 transmitted by satellite S4 is correctly received by terminal 11. Acknowledgments of downlink data messages may be transmitted to server 13. The various nodes of the communication system 10 (terminal 11, satellite 12, server 13) may attempt to send a message several times until an acknowledgment is received or until a predetermined timeout is reached. Such arrangements ensure a good level of message delivery reliability.

[0082] When large, upstream or downstream traffic data can be fragmented and transmitted in several successive messages. An acknowledgment can then be associated with each message. According to another example, only the last message received during an exchange is acknowledged.

[0083] The "Pong" message 51 can indicate the sequence number of the last downlink data message received by terminal 11. This can prevent satellite S4 from sending a downlink data message 54 that has already been received by terminal 11 (and possibly transmitted by another satellite). The "Pong" message can then be considered an acknowledgment of the downlink data message corresponding to that sequence number.

[0084] For upstream traffic data, if the allocation is insufficient to transmit all the data, a new communication can be established by terminal 11 at the end of the current communication, during the same pass, during the next pass of a satellite, or later depending on a possible constraint on a routing time to be respected.

[0085] For downlink traffic data, the "Allocation" message 52 may include an indication that additional data will still be available to terminal 11 after the current communication. This indication may force terminal 11 to shorten the time before establishing a new communication with the satellite network. Terminal 11 may also request the same satellite several times during the same pass.

[0086] Authentication mechanism:

[0087] The communication system 10 can advantageously support a terminal authentication mechanism 11. However, it is hardly feasible to transfer the list of all the system's terminals 11 into the memory of the satellites 12. In this case, it is not possible for a satellite 12 to authenticate the Terminal 11. One solution could be to include authentication parameters in the "Pong" message 51 or in an upstream traffic data message 53. These parameters would be used by the ground server 13 to authenticate terminal 11 at a later time. The authentication parameters could include a challenge and a challenge response, the challenge having been previously generated by the S4 satellite and transmitted to terminal 11. For example, the challenge could be transmitted in the "Ping" message 50. Alternatively, the challenge could be transmitted in the "Allocation" message 52 (possibly with a different challenge for different terminals). The challenge response is calculated by terminal 11 based on the challenge and a cryptographic algorithm. The challenge could be renewed regularly.

[0088] For the downlink, it is possible to transfer the DL messages destined for the terminals to the satellite along with the secrets enabling terminal authentication. The satellite can then authenticate a terminal before transmitting the DL data intended for it.

[0089] Resistance to the Doppler effect by the use of several carriers:

[0090] The signal carrying a "Ping" message broadcast by a satellite 12 is designed to be very easy to detect with a high probability of success, even in the presence of a significant Doppler effect. The Doppler effect causes a frequency difference between the frequency at which the signal carrying the "Ping" message is transmitted by the satellite 12 and the frequency at which the signal is received by the terminal 11. This frequency difference is due to the relative motion between the satellite and the terminal. The greater the difference in velocity between the satellite and the terminal, the greater this frequency difference. The Doppler effect is particularly significant for satellites 12 placed in a LEO orbit.

[0091] The "Ping" message contains a sequence of symbols (a preamble) known to the terminals 11 of the system. The detection of this known sequence of symbols allows the terminals 11 to detect the beginning of a "Ping" message.

[0092] Advantageously, and as illustrated in Figure 8, each satellite 12 of the constellation is configured to broadcast a "Ping" message 50 on a plurality of different carrier frequencies comprising a center frequency f used as the reception frequency by the terminal 11 during the radio listening periods 40 of the "attentive" mode, and 2N offset frequencies f. - fc + i. A y, i being a non-zero integer index varying between -N and N. The frequency offset A f is chosen to be less than or equal to twice a maximum frequency offset A supported by the terminal 11 for the detection of a "Ping" message 50. N is a strictly positive integer that can advantageously be chosen such that ( ff ' °ù ^max is the absolute value of a maximum frequency deviation due to the Doppler effect for a signal emitted by a satellite 12 to a terminal 11.

[0093] With such arrangements, if a frequency difference Df due to the Doppler effect existing between terminal 11 and satellite 12 is such that ( z. A) ,. 5 . r , A / \ r , where i is a non-zero integer index between -N and N, then a "Ping" message emitted by the terminal on the frequency carrier f. can be detected by terminal 11.

[0094] With such arrangements, the capabilities of the detection means of terminal 11 do not depend on the maximum frequency deviation D^x due to the Doppler effect. On the other hand, the number of different carrier frequencies that a satellite 12 must support depends on the value of Dmax.

[0095] As illustrated in [Fig.8], the "Ping" messages 50 emitted on the different carrier frequencies can be time-shifted so that two "Ping" messages emitted respectively on two neighboring carrier frequencies are not emitted simultaneously (i.e. so that the emissions of these two "Ping" messages do not have a time overlap).

[0096] The variation of the Doppler effect over time follows a known curve dependent on the positions of terminal 11 and satellite 12. The variation of the Doppler effect over time can be more or less estimated in the short term when the Doppler frequency difference is known at a given instant by calculating its expected value (its average) for possible positions of terminal 11 relative to that of satellite 12. This estimate can also be based on an estimate of the distance between terminal 11 and satellite 12. This distance can be estimated from a delay observed by satellite 12 for a message transmitted by terminal 11 (the message is delayed by twice the propagation delay). The relative frequencies transmitted in the "Allocation" message can then take into account the estimated evolution of the Doppler effect.Pilot symbols can also be used in exchanged messages to facilitate monitoring the evolution of the Doppler effect.

[0097] Time synchronization:

[0098] The propagation delay of radio signals exchanged between a terminal and a satellite varies over time (because the distance between the terminal and the satellite varies over time). Thus, similarly to what is proposed for taking into account the Doppler frequency difference, the evolution of the propagation delay over time can be estimated and taken into account in the exchanged messages.

[0099] Detection of a "Ping" message:

[0100] It is preferable that the "Ping" message be detectable with reasonable computational complexity for terminals 11, even in the presence of a large Doppler shift. It is also preferable that the size (and therefore the duration) of a The Ping message should be small, in order to optimize the use of the radio medium and thus maximize the system's capacity.

[0101] To this end, a method for detecting a "Ping" message is proposed. The detection method is implemented by a terminal 11. In particular, program code instructions can be stored in an electronic memory of the terminal 11 to configure a processor of the terminal to implement the detection method 200. To implement this method, a "Ping" message comprises a Zadoff-Chu sequence ZC followed by a pseudo-random sequence PN.

[0102] A pseudo-random sequence is a series of symbols that appears statistically random, but which is generated by a deterministic algorithm.

[0103] A Zadoff-Chu sequence is a mathematical sequence used in telecommunications for its specific correlation properties. These sequences are named after their inventors. They are used particularly in 4G and 5G wireless communication systems (fourth and fifth generation mobile telephony). They possess specific properties that make them ideal for detecting a synchronization preamble.

[0104] Zadoff-Chu sequences are CAZAC sequences (Constant Amplitude Zero Autocorrelation waveform). Cyclically shifted versions of a Zadoff-Chu sequence are orthogonal to each other. A generated Zadoff-Chu sequence that has not been shifted is called a root sequence. Each Zadoff-Chu root sequence parameterized by u can be defined by the formula:

[0105] / ~mbi+Cj+2q] X xu (n) = exp[- j—üyc— J

[0106] where:

[0107] xu(n) is a complex value at position n

[0108] Nzc is the sequence length,

[0109] 0 <n<Nzc,

[0110] 0 < u < Nzc and gcd ( Nzo u ) = 1 (gcd corresponds to the "greatest common divisor") [YES] Cf = Nzc mod 2,

[0112] ge€ Z.

[0113] Thanks to the properties of Zadoff-Chu sequences, the presence of the ZC sequence can be detected in a received radio signal by means of a cross-correlation operation. This cross-correlation is calculated between the received signal and a local replica of the ZC sequence stored by the terminal. The maximum absolute value of this correlation indicates, by its amplitude (for example, when it is greater than a predetermined detection threshold), the presence of the ZC sequence in the received signal.

[0114]

[0115]

[0116]

[0117]

[0118]

[0119] The position of the maximum provides an approximate indication of the start time of the ZC sequence in the received signal. Indeed, the Doppler shift experienced by the received signal affects the position of the correlation maximum. If we denote fo as the actual start time of the ZC sequence in the signal, the maximum of the absolute correlation value is observed at a time + Δft, where Δf is the Doppler frequency shift and α is a parameter that depends on the ZC sequence. For example, if we use a ZC sequence obtained with u = 3, Cf = 0, g = 0, and Nzc = 1250, with a symbol rate Rs of 180 kBd (one hundred and eighty kilobauds) for the receiving device, the parameter α corresponds to: α = α = 1.28 × 10-8 s / hz uKI The received signal should be sampled at a frequency sufficient to accommodate the maximum Doppler shift (the sampling frequency can be defined using Shannon's theorem). Signals from local replicas of the sequences are sampled or oversampled at the same frequency. To more precisely locate the ZC sequence within the received signal, we exploit the PN sequence that follows the ZC sequence. To do this, a transformation is applied to the received signal, which consists of multiplying it by a conjugate sequence of the ZC sequence that is coherent (i.e., temporally aligned) with the position of the maximum of the first correlation. The conjugate of a Zadoff-Chu sequence refers to the operation by which each complex element of the sequence is replaced by its complex conjugate. The conjugate sequence can be extended so that it spans a longer duration of the received signal than the combined duration of the ZC and PN sequences contained within the received signal. Note that the result of the transformation of the received signal reveals a gate function, or in other words, a constant time interval, in place of the ZC sequence that the signal initially contained. Searching for the constant interval through a new correlation based solely on the ZC sequence would yield a broad and imprecise cross-correlation peak. Utilizing the PN sequence, which is juxtaposed with the ZC sequence in the message preamble, allows for the creation of a more precise cross-correlation peak (a narrower peak with a greater amplitude). The same transformation (multiplication by the conjugate of the ZC sequence) is therefore applied to a local replica of a signal containing both sequences (the ZC sequence and the PN sequence). This local replica is stored by the terminal. A new cross-correlation is then calculated between the transformed received signal and the transformed local replica of the two sequences. The maximum absolute value of this new correlation precisely provides the start time of the ZC sequence. in the received signal. The start time of the "Ping" message can then be determined based on the precise start time of the ZC sequence (they coincide if the ZC sequence is at the beginning of the "Ping" message). The difference between the precise time and the approximate time determined previously with the first correlation corresponds to the Doppler shift Dj- that affects the received signal, up to a coefficient that depends on the ZC sequence used (parameter a, which depends on the ZC sequence used).

[0120] Figure 9 schematically summarizes the main steps of a method for detecting a "Ping" message carried in a Doppler-affected radio signal. It includes the steps described above, namely: - a first cross-correlation 201 between a signal received by terminal 11 and a local replica of the ZC sequence, - a determination 202 of an approximate start time of the ZC sequence in the signal as a function of a maximum absolute value of the first correlation, - a transformation 203 of the received signal by multiplying it by a conjugate sequence of the ZC sequence, - a transformation 204 of a local replica of the ZC sequence followed by the PN sequence by multiplying said local replica by the conjugate sequence, - a second cross-correlation 205 between the transformed received signal and the transformed local replica, - a determination (206) of a precise start time of the ZC sequence in the signal as a function of a maximum absolute value of the second correlation.

[0121] The method 200 may include an additional step of estimating 207 the Doppler frequency shift affecting the received signal as a function of a difference between the precise time and the approximate time, and as a function of the first sequence.

[0122] The duration of the ZC sequence can be sized to allow detection of the signal with a desired level of probability while ensuring that a given false alarm rate is not exceeded.

[0123] The detection threshold applied to the maximum of the first cross correlation can be defined as a function of the duration of the first sequence, the desired detection and false alarm rates, and / or as a function of a noise power level measured in the received signal.

[0124] The duration of the PN sequence can be chosen so that the maximum of the second cross correlation is at the beginning of the "Ping" message with a certain level of probability.

[0125] In order to reduce the duration of the sequences (at the cost of an increase in the computational complexity of the receivers), the reception frequency band can be divided in sub-bands. The signal search can then be performed in each sub-band for the first step of searching for the ZC sequence. The second step of estimating the parameters of the received signal (message start time, Doppler shift affecting the signal) is performed on the entire reception band.

[0126] The estimation of the Doppler frequency shift can optionally be refined by calculating the phase shift experienced by the received signal over time. For this purpose, a phase shift can be calculated as a phase difference between a phase measured from a starting section of the set formed by the ZC and PN sequences (i.e., from a number Ni of symbols located at the beginning of the set of two sequences) and a phase measured for an ending section of this set (i.e., from a number N2 of symbols located at the end of the set of two sequences). The numbers Ni and N2 are not necessarily identical. The starting section is not necessarily at the very beginning of the set of two sequences, and the ending section is not necessarily at the very end of the set of two sequences.However, the greater the distance between the start and end sections, the more relevant the phase shift measurement will be.

[0127] Optionally, at least one of the first cross-correlation 201 and the second cross-correlation 205 is followed by an interpolation step of the correlation results. This interpolation improves the temporal accuracy of message start detection. Indeed, the maximum absolute value obtained for a correlation does not necessarily correspond to the exact position of the start of the sequence being sought due to discretization (sampling frequency). By using interpolation (linear, spline, parabolic, or other), the precise position of the correlation peak can be estimated with a resolution higher than that provided by discrete samples.

[0128] The PN sequence may advantageously have fewer symbols than the ZC sequence, or even at least half as many symbols as the ZC sequence. By reducing the size of both sequences, the computing power requirements of terminal 11 are reduced, and the message size is reduced (thus optimizing the use of radio resources and system capacity).

[0129] Figure 10 illustrates an example of the implementation of a "Ping" message. In the example shown, the ZC sequence has 198 (one hundred and ninety-eight) symbols and the PN sequence has 72 (seventy-two) symbols. As illustrated in Figure 10, the "Ping" message can also consist of one or more blocks, each formed by a sequence of useful data symbols and a sequence of pilot symbols. In the example shown, each block has 500 (five hundred) useful data symbols and the pilot symbol sequence has 40 (forty) symbols. For each block, a new cross-correlation can be performed between the received signal and a local replica of the pilot symbol sequence to track the phase shift experienced by the signal over time, and to compensate for the associated phase error. For this example implementation of a "Ping" message 50, the numbers Ni and N2 (sizes of the start and end sections of the concatenated sequence ZC +PN used for the phase shift calculation) are, for example, equal to seventy-two (N! = N2 = 72).

[0130] The ZC sequence can correspond to a satellite access network identifier. It is conceivable that several satellite access networks in the system use different ZC sequences, and that a terminal can identify the access network that originated a detected "Ping" message based on the ZC sequence that enabled the detection of the "Ping" message. A terminal may be associated (subscribed to) only one network. In this case, it will attempt to detect "Ping" messages using only the ZC sequence corresponding to the network to which the terminal is associated, and it will not be able to detect "Ping" messages originating from another network. A terminal could also attempt to detect "Ping" messages using several different ZC sequences corresponding to several different networks to which it is subscribed.

[0131] In one variant, the different networks could be distinguished by the use of different PN sequences. According to yet another variant, the identifier of the access network to which a "Ping" message is associated can be encoded in the payload data of the "Ping" message.

[0132] It should be noted that the detection method 200 described above can be applied to any type of communication between a transmitting device and a receiving device in the presence of Doppler. More specifically, the detection method 200 described with reference to [Fig. 9] can be implemented independently of the method 100 for establishing communication between a terminal and a satellite described previously with reference to Figures 1 to 7. The detection method 200 could also be applied to communications that are not necessarily satellite-based. The detection method 200 can advantageously be combined with the use of several different carrier frequencies (as described with reference to [Fig. 8]) to enhance Doppler robustness.

[0133] The foregoing description clearly illustrates that, through its various features and their advantages, the present invention achieves the stated objectives. In particular, with the proposed solution, it is not necessary to ensure that a terminal of the communication system is constantly within range of a satellite in the constellation. This makes it possible to limit the number of satellites in the constellation. The proposed solution also does not require a continuous communication link between a terminal and a ground server with which data must be exchanged. Furthermore, the proposed solution does not require the communication system terminals to know their own position, nor the positions of the satellites in the constellation. Finally, the proposed solution is resistant to the Doppler effect.

Claims

Demands

1. A method (200) for detecting by a receiving device (11) a message (50) carried in a radio signal emitted by a transmitting device (12), said signal being affected by a Doppler effect due to the relative motion between the transmitting device (12) and the receiving device (11), the message comprising a Zadoff-Chu ZC sequence followed by a pseudo-random PN sequence, said method (200) comprising: - a first cross-correlation (201) between a signal received by the terminal (11) and a local replica of the ZC sequence, - a determination (202) of an approximate start time of the ZC sequence in the signal as a function of a maximum absolute value of the first correlation, - a transformation (203) of the received signal by multiplying it by a conjugate sequence of the ZC sequence,- a transformation (204) of a local replica of the ZC sequence followed by the PN sequence by multiplying said replica by the conjugate sequence, - a second cross-correlation (205) between the transformed received signal and the transformed local replica, - a determination (206) of a precise start time of the ZC sequence in the signal as a function of a maximum absolute value of the second correlation.

2. Method (200) according to claim 1 comprising an estimation (207) of a Doppler frequency shift affecting the received signal as a function of a difference between the precise time and the approximate time, and as a function of a parameter that depends on the ZC sequence.

3. Method (200) according to claim 2 wherein the estimation of the Doppler frequency shift is refined as a function of a phase shift observed on the received signal over time, said phase shift being calculated as a difference between a phase measured for a beginning section of the set formed by the sequence ZC and the sequence PN and a phase measured for a ending section of this set.

4. A method (200) according to any one of claims 1 to 3 wherein at least one of the first cross correlation (201) and the second cross correlation (205) is followed by a step of interpolating the results of the correlation to optimize the accuracy of the determination of the maximum absolute value of the correlation.

5. Method (200) according to any one of claims 1 to 4 wherein the PN sequence has fewer symbols than the ZC sequence.

6. Method (200) according to claim 5 wherein the number of symbols of the PN sequence is at least twice as small as the number of symbols of the ZC sequence.

7. Method (200) according to any one of claims 1 to 6 wherein the "Ping" message successively comprises the ZC sequence, the PN sequence, and at least one block formed by a sequence of useful data symbols and a sequence of pilot symbols, and the method (200) comprises at least a third cross-correlation between the received signal and a local replica of the pilot symbol sequence.

8. Method (200) according to any one of claims 1 to 7 wherein the message (50) is a "Ping" message emitted by the transmitting device (12) repeatedly on a plurality of different carrier frequencies comprising: - a center frequency fc used as a reception frequency by the receiving device (11) during radio listening periods (40) to detect a "Ping" message (50), - 2.N offset frequencies f.= fc+i- A / , i being a non-zero integer index varying between -N and N, the frequency shift A y being chosen to be less than or equal to twice a maximum frequency shift A max supported by the receiving device (11) for the detection of a message (50), N being a strictly positive integer chosen such that A / \ / , where DmUx is the absolute value of a maximum frequency shift due to the Doppler effect for a signal emitted by the transmitting device (12) to the receiving device (11).

9. Product computer program comprising program code instructions which, when implemented by a processor, configures said processor to implement a detection method according to any one of claims 1 to Q

10. O. Receiver device (11) comprising a processor configured to implement a detection method (200) according to any one of claims 1 to 8.

11. A satellite communication system (10) comprising at least one ground terminal (11) and a constellation of satellites (12) forming a satellite access network, each satellite (12) in the constellation being configured to repeatedly transmit a "Ping" message (50) signaling the presence of the satellite (12) to a terminal (11) within range of the satellite, the terminal (11) comprising a processor configured to implement a method for detecting a "Ping" message (50) according to any one of claims 1 to Q

12. O. Satellite communication system (10) according to claim 11 wherein the sequence ZC corresponds to a satellite access network identifier.

13. A method (100) for establishing communication between a terrestrial terminal (11) and a satellite (12) belonging to a constellation of satellites orbiting the Earth, said constellation of satellites belonging to a satellite access network of a communication system (10), each satellite (12) of the constellation being configured to repeatedly broadcast a "Ping" message (50) signaling the presence of the satellite to a terminal (11) within range of the satellite, the method (100) comprising: a. a reception (101) by the terminal (11), from a satellite (12, SI) of the constellation, of a message (50) containing a list (16) of future times of passage of a satellite of the constellation within range of the terminal (11), b.a selection (102), by the terminal (11), according to a desired instant (te) to establish communication with the satellite access network, of a future instant (ts) of passage of a satellite (12, S4) from among all the future instants of passage of satellites included in the list (16),. c. a determination (103) of a wake-up time (tw) of the terminal (11) as a function of the selected future transition time (ts), d. a transition (104) of the terminal (11) into a "sleep" mode until the moment of awakening (tw), e. when the wake-up time (tw) is reached, a transition (105) of the terminal (11) into an "attentive" mode in which the terminal (11) repeatedly performs radio listening periods (40) to detect a "Ping" message (50) from the satellite (12, S4), f. a detection (200) of a "Ping" message (50) by the terminal (11) using the detection method (200) according to any one of claims 1 to 8 and, when a "Ping" message is detected, an emission (106) of a "Pong" message (51), by the terminal (11), to the satellite (12, S4).

Images