Method and system for determining the location of a stopped vehicle on a storage track using a virtual beacon

Abstract

The invention relates to a method for determining the point position of a vehicle stopped on one of a set of storage tracks using virtual beacons. The method comprises determining (208, 210) and comparing (212) the likelihood of a plurality of hypotheses about the stopping position of the vehicle corresponding to the first set of predefined virtual beacons Be(i) NBe by correlating GNSS geolocation signals received by the GNSS receiver on the vehicle at respective synchronization reset times of a second set with predicted GNSS geolocation signals of copies of the positions expected at said respective times for predefined virtual beacons Be(i) NBe of the first set of predefined virtual beacons Be(i), the respective positions of which are known in an amount of at least one virtual beacon per storage track, i varying from 1 to NBe. The detected holding position of the vehicle is the position corresponding to the maximum likelihood.

Description

Technical Field

[0001] The present invention relates to the field of satellite geolocation. More specifically, the present invention relates to a method and system for determining a point position, i.e., instantaneously or in real time, the time-stamped position of a stopped vehicle held on one of a set of storage and / or departure tracks using a virtual beacon.

[0002] The invention also relates to a device for determining a point position of a vehicle parked and stopped on a departure track, which device is integrated into a system for determining a point position according to the invention, and to a computer program product for determining a point position to be executed or executed by the device.

[0003] The present invention is applicable to any vehicle or mobile means that is parked and stops on a storage and / or departure track. Background Art

[0004] As is well known, determining the storage and / or departure track of a stationary train after starting from a cold state (i.e., after turning on or preparing its onboard electronic and electrical equipment) is a key element in locating the train when starting from a cold state, where there is no a priori information about the train's parked location. Specifically, once the train's departure track has been determined, proprioceptive sensors such as an odometer and / or an inertial measurement unit, along with a track map, can be used to determine the train's position and keep the train's position updated during subsequent movement of the train.

[0005] Conventionally, in the context of the ETCS L2 standard (acronym for European Train Control System Level 2), the use of contactors or RFID beacons (RFID is the abbreviation for Radio Frequency Identification) placed on each storage track allows to almost certainly eliminate any doubt after a low-speed train movement as to which section of the storage track is involved.

[0006] RFID beacons (whose positions are known in all maintained tracks) enable the activation of an RFID signal that is detected and dated by the onboard positioning equipment when a train passes, and is used to reset other onboard positioning units (such as odometers) accurately to within one meter and with high integrity.

[0007] A disadvantage of using contactors or RFID beacons to determine the location of points in this system is that the system requires a large infrastructure in terms of the number of RFID beacons that need to be installed on the storage track and the maintenance of the installed beacons. Another disadvantage is that before the train can be located by the RFID beacons or contactors, it must begin "blind" movement at a very low speed, which requires prior authorization from the traffic authorities.

[0008] Furthermore, known solutions based on standard GNSS (Global Navigation Satellite System) receivers are generally envisaged as a way to help achieve a first level position with an "on-orbit" accuracy of sufficient granularity to identify the deposited orbit closest to the estimated position.

[0009] A number of documents, for example the first document US 5 375 059 A, describe systems and methods for improving the real-time positioning accuracy of autonomous land vehicles using a global satellite positioning system and a set of beacons.

[0010] However, there is always the possibility of distortion in the received signal, which may be related to multipath or local interference, and the train being stopped may result in a position error that is much larger than the spacing between the tracks, so there is a risk of making a determination that the track is corrupted due to loss of integrity even after a long integration of the calculated position (in terms of the integration duration).

[0011] In general, the main known limitations associated with using standard GNSS methods are as follows:

[0012] - the availability of GNSS signals may be insufficient in situations where the land vehicle is stored on the ground due to signal masking or unavailability of satellite signals; and / or

[0013] - the quality of the measurements may degrade, this degradation manifesting itself in the form of measurement deviations due to the effects of propagation disturbances (e.g. multipath or local interference in the reception environment); and / or - the accuracy and integrity of the position measurements may be much lower than that provided by more robust physical beacons and may not allow the safety objectives sought to be achieved by railway traffic, in particular those concerning delivery positions with an integrity level of SIL4 (acronym for Safety Integrity Level 4) such as that defined in standard IEC-61508.

[0014] Therefore, a first technical problem is to provide a method and a system for determining the point position of a vehicle parked on a storage track and identifying it before departure using virtual beacons associated with said departure track, which allows, using only measurements provided by a standard GNSS receiver, to identify, among a set of departure tracks close to each other, the departure track on which the stored train has been found, with a high safety integrity level, after the parked train has been unleashed on said departure track.

[0015] A second technical problem is to provide a method and a system for determining the point position of a vehicle stopped on a storage track using a virtual beacon associated with said departure track, said virtual beacon seeking to be identified before departure, which allows identifying the departure track on which the stored train has been found among a set of departure tracks close to one another, using only measurements provided by a standard GNSS receiver, and which guarantees high integrity performance and high robustness performance with respect to the effects of propagation outages or local interferences in the reception environment. Summary of the Invention

[0016] To this end, a subject of the invention is a method for determining the point position of a vehicle parked on one of a set of storage rails using virtual beacons.

[0017] The position determination method is implemented by a system for determining a point position, the system comprising: a GNSS receiver located on the vehicle, capable of directly measuring pseudoranges set as a synchronized local time base within a synchronized imprecise time range using geolocation signals transmitted by GNSS geolocation satellites visible to the GNSS receiver; and an electronic processing unit located outside the GNSS receiver or integrated into the GNSS receiver.

[0018] The method for determining the position of a point is characterized in that it comprises a set of steps, wherein: by correlating GNSS geolocation signals received by said GNSS receiver on board said vehicle at respective times of a second set with predicted GNSS geolocation signals of copies of the respective positions expected for a first set of integers NBe of predefined virtual beacons Be(i) at said respective times, i varying from 1 to NBe, the likelihood of a plurality of hypotheses regarding the stopping position of said vehicle corresponding to said first set is determined and compared by said electronic processing unit, the respective positions of said predefined virtual beacons Be(i) being known in an amount of at least one virtual beacon per storage track, and the detected parking position of said vehicle being the position corresponding to the maximum likelihood.

[0019] According to a particular embodiment, the method for determining a point position of a vehicle includes one or more of the following features, which may be implemented individually or in combination:

[0020] - in order to reduce mismatches due to poor synchronization of the local time reference of the GNSS receiver with respect to the time of the GNSS global positioning system, said imprecise time range is covered by the electronic processing unit by dividing the imprecise time range of the synchronization of the local clock into time periods respectively represented by synchronization reset times tk, k varying between 1 and Nk, wherein Nk specifies the number of synchronization reset times within said imprecise time range, correlation of the GNSS geolocation signals received by the on-board GNSS receiver being performed at synchronization reset times tk;

[0021] the correlation of the GNSS geolocation signal received by the on-board GNSS receiver at each position P(i) of the virtual beacon Be(i) and at each time tk is a multi-satellite correlation with the corresponding replica GNSS signal expected for each position P(i) at the time tk and from the satellites Sat(i,j,k) visible from the position P(i) of the virtual beacon Be(i) at the time tk, i varying from 1 to NBe and k varying from 1 to Nk, the multi-satellite correlation for the position P(i) at the synchronization reset time tk being equal to the sum of the squares of the elementary correlations over the set of satellites visible from the position P(i) at the reset time tk: the actual GNSS signal received by the GNSS receiver for the position P(i) at the time tk and the GNSS replica expected at the position Pi at the time tk and from the satellites Sat(i,j,k) visible from the beacon Be(i) at the time tk, where i varies from 1 to NBe and k varies from 1 to Nk;

[0022] - a set of steps comprising an initialization phase, a phase of searching for the position of the virtual beacon closest to the holding position of the vehicle using a predefined strategy, and a phase utilizing a search phase during which a storage track Vi0 is identified based on a priori knowledge of the virtual beacon B(ei0) estimated to be closest to the vehicle and the position P(i) of said virtual beacon Be(i), said position P(i) being set and predefined, i varying from 1 to NBe;

[0023] - an initialization phase of a set of steps comprising a first initialization step in which a first set of possible positions P(i) of the virtual beacons Be(i) corresponding to possible storage trajectories Vi of the vehicle is determined from a database of terrain data, i varying from 1 to NBe, said database of terrain data being external to the GNSS receiver or being included in a database memory of the GNSS receiver, and a second set of synchronization reset times tk forming hypotheses is determined so as to be substantially regularly spaced within the synchronization imprecision time range, k varying from 1 to Nk, Nk being an integer greater than or equal to 3, said synchronization imprecision time range being included in the time correlation field of the GNSS PRN code sequence or determined using an integrity protection time radius calculated from time data of a RAIM function integrated in the GNSS receiver or external to the GNSS receiver;

[0024] - the difference between the actual position of the GNSS receiver along the storage track on which the vehicle is parked and the position of the associated virtual beacon is smaller than an imprecision spatial interval, the equivalent time impact of which on the time correlation accuracy is significantly lower than the time correlation domain of the GNSS PRN code sequence used;

[0025] said phase of searching for the closest virtual beacon comprises a second step, in which: the GNSS receiver on board the vehicle receives real GNSS signals at synchronization reset times tk, k varying from 1 to Nk, and the electronic processing unit determines, for each synchronization reset time tk of said second set of hypotheses and for each possible position P(i) of a virtual beacon Be(i) of said first set, the GNSS replica signal expected by said GNSS receiver at said position P(i) and time tk for each visible satellite Sat(i,j,k) at the local time tk of said receiver, k varying from 1 to Nk, i varying from 1 to NBe, i varying from 1 to NBe and k varying from 1 to Nk; then, for each reset time tk of said second set and for each position P(i) of a virtual beacon of said first set, the electronic processing unit determines, on the one hand, a basic correlation between the signal of the GNSS replica expected at each holding position P(i) and each synchronization reset time tk and from each satellite visible at each time tk and the actual GNSS signal received by said GNSS receiver on board the vehicle at the synchronization reset time tk;

[0026] - said phase of searching for the virtual beacon closest to said vehicle comprises a third step, carried out after or in parallel with said second step, in which, for each reset time tk of said second set and for each position P(i) of the virtual beacons of said first set, said electronic processing unit determines a multi-satellite correlation forming a value of a likelihood function, and wherein, for each position P(i) of the virtual beacons of said first set and for each reset time tk of said second set, said multi-satellite correlation is equal to the sum of the squares of the elementary correlations between the GNSS signal received by said GNSS receiver at time tk and the signal of the replica expected at time tk for the position P(i) of said virtual beacon from the visible satellites Sat(i,j,k) visible at said position P(i) at time tk;

[0027] - said phase of searching for the closest virtual beacon comprises a fourth step, carried out after said third step, in which said electronic processing unit determines the position P(i0) of said virtual beacon and the reset time tk0 that maximizes the likelihood function of the positions P(i) of said first set and the reset times tk of said second set, whereby the position P(i0) is determined as the position of said virtual beacon Be(i0) detected as being closest to said on-board receiver;

[0028] - the method for determining a point position further comprises a fifth step, performed after the fourth step, in which: the electronic processing unit identifies the storage track on which the vehicle is stopped, based on the position P(i0) of the virtual beacon detected as being closest and a lookup table mapping all virtual beacons to all storage tracks; and / or the electronic processing unit transmits information about the detected synchronization reset time tk0 to a mechanism for correcting the drift of the local clock of the GNSS detector; and / or the electronic processing unit issues an alarm in case of failure to identify the storage track and starts exploration in a wider synchronization imprecision time domain;

[0029] - for each virtual beacon position (Pi), the index of the set of satellites visible from all virtual beacons Be(i) at a given time tk is the same and independent of the order of the virtual beacons i and i varies from 1 to NBe only depending on said time tk;

[0030] - checking the integrity of the position by an additional step of verifying that the maximum likelihood is above a predefined safety threshold guaranteeing the integrity of the position.

[0031] Another subject of the invention is a device for determining the point position of a vehicle stopped on one of a set of storage rails using virtual beacons, comprising: a GNSS receiver located on the vehicle and capable of measuring pseudoranges set to a synchronized local time base directly within a precise time range using geolocation signals emitted by GNSS geolocation satellites visible to the receiver; and an electronic processing unit located outside the GNSS receiver or integrated into the GNSS receiver.

[0032] The device for determining the position of a point is characterized in that the electronic processing unit is configured to: determine and compare the likelihood of multiple hypotheses about the stopping position of the vehicle corresponding to the first set by correlating GNSS geolocation signals received by the GNSS receiver located on the vehicle at respective times of a second set with predicted GNSS geolocation signals of expected copies of the respective positions of virtual beacons of a first set of integers NBe of predefined virtual beacons Be(i) at said respective times, i varying from 1 to NBe, the corresponding positions of the predefined virtual beacons Be(i) being known in an amount of at least one virtual beacon per storage track; and the current storage track position of the vehicle detected is the position corresponding to the maximum likelihood.

[0033] According to a particular embodiment, the device for determining a point position of a vehicle includes one or more of the following features, which can be implemented individually or in combination:

[0034] - in order to reduce mismatches due to poor synchronization of the local time reference of the GNSS receiver with respect to the time of the GNSS global positioning system, the electronic processing unit is configured to cover the imprecise time range by dividing the imprecise time range of the local clock into time periods respectively represented by synchronization reset times tk in a predefined manner, k varying between 1 and Nk, wherein Nk specifies the number of synchronization resets within the imprecise time range, the correlation of the GNSS geolocation signals received by the on-board GNSS receiver being performed at the synchronization reset time tk; and the correlation of the GNSS geolocation signals received by the on-board GNSS receiver at each position P(i) of the virtual beacon Be(i) and at each time tk is performed with respect to the time reference for the local clock at the time tk and Nk. and a multi-satellite correlation of the corresponding replica GNSS signal expected at each position P(i) from the satellites Sat(i,j,k) visible from the position P(i) of the virtual beacon Be(i) at time tk, i varying from 1 to NBe and k varying from 1 to Nk, said multi-satellite correlation for said position P(i) at a synchronization reset time tk being equal to the sum of the squares of the elementary correlations over the set of satellites visible from said position P(i) at the reset time tk between the actual GNSS signal received by the GNSS receiver for said position P(i) at time tk and the GNSS replica expected at time tk and at position Pi from the satellites Sat(i,j,k) visible from the beacon Be(i) at time tk, where i varies from 1 to NBe and k varies from 1 to Nk;

[0035] The electronic processing unit is configured to determine the synchronization imprecision time range by including the synchronization imprecision time range in a time correlation field of a GNSS PRN code sequence or by calculating an integrity protection time radius based on time data of a RAIM function, the RAIM function being integrated into the GNSS receiver or being external to the GNSS receiver.

[0036] Another subject of the present invention is a system for determining the point position of a vehicle stopped on one of a set of storage tracks using virtual beacons, comprising: a GNSS global satellite geolocation system; and an apparatus for determining the point position of a vehicle such as defined above, preferably enhanced by a RAIM function implemented within a GNSS receiver or within a second GNSS receiver separate from the first GNSS receiver; the vehicle being: a land vehicle parked on a land storage track, preferably a train stored on a storage track, or a marine transport stored on a seaway.

[0037] Another subject of the invention is a computer program product containing computer-readable instructions which, when executed on a processing unit, cause the processing unit to perform a method such as described above for determining the position of a vehicle stopped on one of a set of storage rails using virtual beacons. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The invention will be better understood from a reading of the following description of several embodiments, given by way of example only and with reference to the accompanying drawings, in which:

[0039] Figure 1 The relationship between temporal and spatial correlations is shown;

[0040] Figure 2 is a diagram of the general architecture of a system for determining the point position of a stopped vehicle according to the present invention, where the train is held on one of a set of storage and departure tracks of known topography;

[0041] Figure 3 is an overall flow chart of a method for determining a point position of a stopped vehicle according to the present invention, where the train is parked on one of a set of storage and departure tracks whose topography is known, said method for determining a point position being Figure 2 The system used to determine the position of a point is implemented in;

[0042] Figure 4 Is used to determine Figure 3 A flowchart of a first specific embodiment of a method for determining a point position of a train;

[0043] Figure 5 Is used to determine Figure 3 A flowchart of a second specific embodiment of a method for determining a point position of a train;

[0044] Figure 6 It is allowed to implement Figure 5 A partial view of a modular architecture of a system for determining a point position according to a second embodiment of the present invention;

[0045] Figure 7 is a graphical view of an example of a resulting multi-satellite correlation or detection function obtained after summing the squares of the outputs of a matched filtering operation or basic correlation function for satellites with various satellite visibility events. DETAILED DESCRIPTION

[0046] The overall objective of the method and system for determining the position of a land vehicle (e.g. a train) that remains stopped on a storage and / or departure track according to the present invention is to compare the likelihood of multiple hypotheses about the train's starting position, which correspond to predefined and known storage or parking positions, by verifying the consistency of GNSS signals received by a standard GNSS receiver located on the train with predicted GNSS signals of expected counterparts for these respective positions or correlating GNSS signals received by a standard GNSS receiver located on the train with predicted GNSS signals of expected counterparts for these respective positions.

[0047] It is assumed that the stowage position of the train for each departure track is accurate enough to ensure that the onboard GNSS receiver is positioned at the parking position within one meter of a visual marker located at a point along the departure track and corresponding to a single virtual beacon of the departure track.

[0048] However, if the train's stopping position cannot be guaranteed to within 1 meter, it is necessary to evaluate the likelihood of each stopping position hypothesis in spatial uncertainty intervals greater than 1 meter in length and explore spatial uncertainty intervals for each track in 1 meter increments.

[0049] In contrast to the use of RFID beacons, the method and system for determining the position of a train according to the invention allows detection of the departure track (before it passes a railway switch) without even requiring the train to move at low speed.

[0050] In cases where the train's stopping position can be guaranteed to within one meter, the likely location where the GNSS signal is expected to be received is given by the stopping point positions P(i) corresponding to the virtual beacons Be(i), with one virtual beacon per storage track. In this case, the beacon Be(i) and its associated storage track Vi can be designated by a common index i, ranging from 1 to NBe. In this particular case, NBe indicates both the number of beacons and the number of tracks in the set of possible storage and departure tracks where the train may be parked. The positions P(i) of the virtual beacons (i ranging from 1 to NBe) form a first set of hypotheses about the train's storage and departure positions and can be evaluated in parallel and / or serially, depending on the processing method being performed.

[0051] In the case where the train's stopping position cannot be guaranteed to within 1 meter, the possible receiving position of the expected GNSS signal is given by the parking point position P(i) corresponding to the virtual beacon Be(i), where each storage track Vs has at least one virtual beacon, s is the number index of the storage and departure track varying from 1 to NV, and i is the number index of the virtual beacon Be(i) varying from 1 to NBe, NV specifies the number of tracks Vs in the set of possible storage and departure tracks on which the train may be parked, and NBe specifies the total number of virtual beacons. Assuming that the number of virtual beacons per track Vs is NBV, where NBV is greater than or equal to 1, the beacons can be continuously separated by at most one meter, and the number index i of the virtual beacon can vary from 1 to NBe according to the following relationship:

[0052] i=(s-1)*NBV+r

[0053] The number index s of the storage track varies from 1 to NV, and r is the integer index of the sequence of beacons within the same given track.

[0054] The positions P(i) of the virtual beacons (i varies from 1 to NBe) form a first set of hypotheses about the storage and departure positions of the train and can be evaluated in parallel and / or serially depending on the processing mode performed.

[0055] A first advantage of the method and system for determining a point position according to the invention is that multiple hypotheses about the storage rails regarding possible parking positions P(i) of the train can be tested without moving the train.

[0056] A second advantage of the method for determining a point position and the system for determining a point position according to the invention, compared to conventional positioning methods and systems used with standard GNSS receivers, is the fact that it allows a more robust likelihood ratio to be established, evaluating various possible virtual beacon or deposited orbit solutions based on a purely energy linear criterion, without having to use an arrival time estimator (such as that conventionally implemented in GNSS receivers), which becomes nonlinear when the waveform of the received GNSS signal is distorted due to multipath or various interference sources.

[0057] A third advantage of the method for determining the position of a point and the system for determining the position of a point according to the invention lies in the fact that the estimation can be reduced to a single spatial dimension by ignoring time, whereas the temporal aspect is predefined a priori by dividing the local clock relative to the GNSS system clock into a set of time periods of imprecision intervals, the width of which is small enough so that in the process of estimating the position of the departure track on which the train is located, the contribution of the temporal uncertainty to the uncertainty of the joint position estimate (relative to the contribution of the spatial uncertainty of the actual position of the train relative to its position on the track) can be made negligible.

[0058] In the following, the expression "virtual beacon" is understood to mean a predefined position or reference point, the spatial coordinates of which are precisely known.

[0059] In the following, the expression "current time" is understood to mean the time measured in the local time base of a standard GNSS receiver included in the positioning device according to the present invention. t .

[0060] In the following, the expression “synchronization reset time” of level k is understood to mean the time comprised in the kth segment of the coverage of the imprecise time range of the local clock synchronization of a standard GNSS receiver, k being an integer comprised between 1 and Nk.

[0061] The GNSS (abbreviation for Global Navigation Satellite System) satellite positioning system is, for example, the GPS (abbreviation for Global Positioning System) system, the Galileo system, the GLONASS system or any other equivalent system.

[0062] It should be recalled that the C / A (short for Coarse / Acquisition) acquisition code in the case of GPS is a digital signal consisting of 1023 chips, repeated every millisecond. It should be noted that the term "chip" used in GNSS technology should not be confused with the term "bit", which is used to define a unit of information.

[0063] According to a first aspect, the concept of the invention is based on the joint implementation of:

[0064] - comparison of likelihood functions of set restricted integers NBe of predefined and known parking positions P(i) corresponding to the virtual beacon Be(i), i varying from 1 to NBe; and - an algorithmic core that implements processing by multi-satellite correlation, i.e. correlation involving signals from satellites of different inclinations, which has been disclosed in the second document EP 3306272 A1, and which checks the consistency of the GNSS signals received by the on-board GNSS receiver with the GNSS signals expected for various hypotheses regarding the position P(i) of the train on the corresponding departure track Vi.

[0065] Unlike conventional positioning methods and systems, which use a standard GNSS receiver on a stationary vehicle and direct position estimation, and therefore carry the risk of serious deficiencies in the integrity of the estimated position, a risk that increases if the train is stationary and the GNSS receiver receives very slowly varying multipath and local interference, the method and system for determining the position of a point according to the present invention implements multi-satellite (i.e. multi-dimensional) filtering along all boresights of visible satellites, and which is customized relative to the positions of a first set of predefined positions P(i) of virtual beacons and the reception times tk of a second set of predefined possible times (synchronization reset between the local time base and the time base of the GNSS system).

[0066] In particular, following the "search strategy" of the method for determining the position of a point according to the present invention, the energy of the position estimator (i.e. the amplitude of the sum of the squares of the amplitudes of the elementary correlators respectively associated with the available visible satellites) is maximized for various hypotheses about the stopping position of the train, i.e. it is maximized for a first set of positions Pi of the virtual beacons Be(i), i varying from 1 to NBe, assuming that the train has stopped.

[0067] According to a first aspect of the inventive concept, a matched multi-satellite or multi-dimensional filtering is performed by correlating a GNSS signal received by a GNSS receiver with a replica signal expected at the current reception time t for a spatial point Pi corresponding to a possible parking position of the GNSS receiver (i.e., the position of a virtual beacon Bei), i varying from 1 to NBe. The code of the local replica signal received by the GNSS terminal at the current reception time is based on a set-phase transmission from one or more visible satellites, and the reception phase of the replica signal expected at the current reception time t is adjusted according to the motion of said satellites, whose paths are calculated from the ephemeris, according to various assumptions about the position P(i), and according to the current reception time t, assuming that the GNSS receiver's time base has been previously synchronized using a conventional PVT (abbreviation for position velocity time) calculation, the imprecision of which is defined by an imprecision time interval whose width for conventional GNSS receivers of the GPS type is equal to, for example, a few hundred nanoseconds. In fact, the uncertainty in the GNSS time estimate is due to the same defects, i.e., the same bias, as the uncertainty in the position estimate in the presence of multipath or interference. The positioning method according to the invention allows reducing this uncertainty by providing an exploration of the domain using an estimation method that is not affected by these major measurement imperfections.

[0068] Assuming perfect synchronization between the local time base of the GNSS receiver and the time base of the GNSS system, the phase of the GNSS signal actually received by the GNSS receiver naturally varies over time, depending on the set position of the virtual beacon associated with the parking track; when the phase of the signal actually received locally corresponds to the phase of the replica signal expected at the parking position of the GNSS receiver, the optimal match and maximization of the likelihood function can be achieved and maintained over time.

[0069] However, in reality, for various parking positions P(i), the GNSS receiver is t This consistency between the actual GNSS signal received and the predicted and calculated expected replica signal cannot be fully ensured as i varies from 1 to NBe except for the following reasons:

[0070] - Errors in the synchronization of the GNSS receiver’s local time with the GNSS system’s time, which necessarily include estimated inaccuracies caused by the inherent inaccuracies of the GNSS receiver’s local time base (e.g., a temperature-compensated quartz oscillator) being higher than the inaccuracies of the GNSS system’s time base (typically an atomic clock to meet space flight constraints), and even instabilities over time;

[0071] - Errors specific to the GNSS system and radio signal propagation errors that cannot be fully corrected by various compensation models, such as propagation errors caused by the ionosphere and / or troposphere, antenna leverage effects, satellite ephemeris, and errors in time synchronization between satellites.

[0072] It will be noted that the residual propagation errors after model compensation remain on the order of a few meters, with only a small impact on the level of the spatial correlation function; for example, in the GPS C / A signal mode, the length of the spatial correlation span is 300 meters, and that these residual propagation errors vary little over the time required to start a stopped train (i.e., 10 minutes or less), and do not introduce noise or other bias in the determination of the maximum spatial correlation according to the various possible parking positions P(i) of the train.

[0073] Therefore, if the synchronization difference between the local time base of the GNSS receiver and the time of the GNSS system becomes larger than the chip duration of the pseudo-random GNSS code, for example equal to 1 μs in the case of the GPS C / A code, the local time base of the GNSS receiver is not accurately synchronized with respect to the separate GNSS system, which may lead to a significant reduction in sensitivity in the train point position determination.

[0074] In order to mitigate such synchronization errors, and according to a second aspect, the concept of the invention is based on resetting the time synchronization, for a given second set of synchronization times, a plurality of different local reception times distributed over a synchronization reset time interval. tk , a search for maximum spatial correlation of each hold position P(i) is employed, with the time interval centered on the resolved reference synchronization time in order to sample the residual uncertainty in the resolved time. For example, in the case of the C / A code of the GPS sequence, a synchronization reset interval of width equal to 1 μs can be explored by sampling in increments of 0.2 μs.

[0075] Thus, by performing in parallel a number of filtering operations in multiple dimensions with respect to visible satellites and their angles of incidence, these filtering operations being adjusted to various possible holding positions of the train based on predicted copies of the received signals, and these filtering operations being calculated for various assumptions about the required reset of the local time in the time uncertainty domain, the inaccuracies in the local synchronization are covered and the departure track on which the stopped train is located can be identified.

[0076] It will be noted that among the possible methods of determining the size of the time uncertainty domain (i.e. the time synchronization interval to be taken into account), one method consists in calculating the protection time radius of the time measurement based on the time information of the RAIM function (acronym for the well-known Receiver Autonomous Integrity Monitoring in GNSS systems), which allows that the deviations (i.e. the errors) in the satellite boresight measurements may have exceeded a maximum protection time radius or alarm radius corresponding to the detection of a detected GNSS system integrity failure.

[0077] Therefore, the time exploration domain or synchronization reset time interval is limited by the RAIM protection radius. The RAIM function guarantees the detection of any satellite boresight error that would result in a time error greater than this protection radius with the required confidence (in terms of the probability of error non-detection Pnd and the probability of false alarm Pfa). This is in contrast to the accuracy of the time measurement alone, which assumes that there are no errors in the measurement relative to the satellite boresight, i.e., no integrity failure in the GNSS system (failure of at least one satellite). It is assumed that the protection radius calculated for each measurement epoch (typically every second) is sufficiently stable to allow exploration of the time domain.

[0078] Figure 1 The figure shows the equivalence between time synchronization by time correlation and position synchronization by space correlation. GNSS (t) is a graphical representation of the time correlation function 12 as a function of time between the expected signal at the point of the abscissa Xi of the beacon Be(i) and at the synchronization time ti.

[0079] The span T of the time correlation function 12 corr14 corresponds to the duration of a chip of the GNSS code sequence. The term "chip" used in GNSS technology refers to the binary information that modulates the signal of the GNSS code sequence and is different from the concept of bits used to define information units. For example, the chip duration of GPS. The maximum value 16 of this time correlation function 12 is when the expected signal is the same as the GNSS signal received by the GNSS receiver. GNSS (t) The synchronization moment ti is obtained.

[0080] Figure 1 Also shown is a one-dimensional graphical representation of the equivalent spatial correlation function 22 between the GNSS signal received by the receiver from the geolocation satellite Sat(j) and the expected signal at the point Be(i) with the abscissa Xi at the synchronization time ti and at the position Xi of the virtual beacon Be(i) as a function of the abscissa X about the position P(i) associated with the virtual beacon Be(i).

[0081] The span X of the spatial correlation function 22 corr 24 corresponds to the time-dependent span T corr 14 onto the axis 25 of the change in the position of the GNSS receiver around the virtual beacon Be(i). The spatial correlation span X corr and the time-dependent span T corr The relationship is through the following formula:

[0082] X corr =c*T corr / cos(αj)

[0083] in:

[0084] -X corr Here is the length of the span of the spatial correlation function;

[0085] -c is the speed of light;

[0086] -αj is the angle of incidence of the satellite signal transmitted by the satellite Sat(j), j being between 1 and NSat with respect to the direction of motion of the vehicle;

[0087] - The operator "*" represents the multiplication symbol.

[0088] The maximum value 26 of the time correlation function 22 corresponds to the abscissa of the synchronization of the expected GNSS signal with the signal received by the GNSS receiver from the satellite Sat(j), which represents the position closest to the virtual beacon Be(i). Figure 1 In FIG, assuming that synchronization has been achieved accurately, the maximum value 26 corresponds to the abscissa Xi of the position P(i) of the virtual beacon Be(i).

[0089] exist Figure 2In and according to the invention, a system 152 for determining a point position of a train 154 stopped on a starting track 156 comprises a GNSS satellite global positioning system 172 and a device 174, the point position being unknown a priori to the train but forming part of a set 162 of starting tracks Vi, i varying from 1 to NBe, the topology of which is known, the device 174 for determining the point position of the train and for identifying the departure track on which said stopped train is parked.

[0090] The onboard equipment 174 for determining the position of a point comprises a conventional standard GNSS receiver 182 and a terrain data database 184 storing the terrain of the track Vi and the positions P(i) of the virtual beacons Be(i) respectively associated therewith.

[0091] The on-board device for determining a point position 174 further comprises an electronic processing unit for determining a point position 186 , which is connected to the standard GNSS receiver 182 and the terrain data database 184 .

[0092] The device 174 for determining a point position is configured to implement a method 202 for determining a point position. Figure 3 is shown in a general manner.

[0093] according to Figure 3 and the most general embodiment of the method for determining a point position, the method 202 for determining a point position according to the present invention comprises a set of steps 204 configured to implement a strategy for searching for the position of a virtual beacon that best matches the position of the train 154, i.e. further identifying the virtual beacon Be(i0) closest to the position of the GNSS receiver on the train 154 and thereby identifying the associated starting track Vi0 by utilizing the received GNSS signals and a priori knowledge of the position P(i) of the virtual beacon Be(i), the position P(i) of the virtual beacon Be(i) being set and predefined, the knowledge being provided by a database of terrain data on a set 162 of starting tracks Vi, i varying from 1 to NBe.

[0094] The search for the virtual beacon position that best matches the actual position of the train and its onboard GNSS receiver 182 stopped at its starting position is performed by correlating the GNSS signals received by the GNSS receiver 182 from the various visible satellites Sat(j) with the expected replica code signals of these various satellites at a given local reference time of the GNSS receiver. t The above is generated for each position P(i) of the virtual beacon Be(i), ie, various parking positions about the track Vi, where i varies from 1 to NBe.

[0095] Furthermore, in order to reduce the inefficiency caused by poor synchronization of the local time reference with respect to the system time of the GNSS global positioning system, the synchronization uncertainty domain can be covered by dividing it into time periods represented by reset times tk using predefined division increments, k varying between 1 and Nk, where Nk specifies the number of synchronization reset times. For example, for synchronization inaccuracies or uncertainty domains shorter than 1 μs in duration, this allows the PRN code of the GPS C / A sequence to be kept within the time correlation domain, and increments of duration less than or equal to 200 ps may be considered.

[0096] The various parking positions P(i) are stored in a terrain data database on the virtual beacon, which may be located outside the GNSS receiver, such as Figure 2 as shown, or may be contained in a database memory of a GNSS receiver.

[0097] The set of steps 204 of the general method 202 for determining the point position of a stopped train on a starting track of a set of parking tracks includes: an initialization phase, which includes a first step 206; a phase of searching for the best matching virtual beacon, which includes second, third and fourth steps 208, 210, 212; and a phase of utilizing the search results, which includes a fifth step 214.

[0098] In a first initialization step 206, a first set of possible positions P(i) of virtual beacons Be(i), corresponding to possible storage tracks Vi of the vehicle (here, a train), is determined from a database of terrain data, with i varying from 1 to NBe. It is assumed that the difference between the actual position of the GNSS receiver along any given storage track of the train and the position of the virtual beacon associated with that storage track is less than a predefined spatial uncertainty interval, which is ensured by the driving constraints imposed on the train driver when storing the train.

[0099] In the same initialization first step 206 , a synchronization reset time tk is determined to form a second set of hypothesis data within the synchronization imprecision time range contained in the time correlation domain of the GNSS code sequence, where k varies from 1 to Nk, where Nk is an integer greater than or equal to 3.

[0100] Next, in a second step 208 implementing a strategy for searching for the closest virtual beacon, the onboard GNSS receiver 182 receives a real GNSS signal at a synchronization reset time tk, with k varying from 1 to Nk.

[0101] In the same second step 208, the electronic processing unit 186, which may be external to the GNSS receiver 182 or integrated into the GNSS receiver 182, determines, for each synchronization reset time tk of the second set (k varies from 1 to Nk) and for each possible position P(i) of the virtual beacon Be(i) or each possible holding position of the first set (i varies from 1 to NBe), the GNSS replica signal expected by the GNSS receiver 182 at position P(i) and time tk for each satellite Sat(i, j, k) visible from position P(i) at the receiver local time tk, i varies from 1 to NBe, k varies from 1 to Nk, j being an index of the order of the set of satellites visible from the given position P(i) at a given synchronization reset time tk.

[0102] In the same second step 208, for each reset time tk of the second set and for each position P(i) of the virtual beacon of the first set, the electronic processing unit determines a basic correlation between the signal of the expected GNSS replica at the respective parking position P(i) and the respective synchronization reset time tk and for the various satellites Sat(i, j, k) visible at each time tk and the real GNSS signal received by the on-board GNSS receiver at the synchronization reset time tk.

[0103] In a third step 210, performed after or in parallel with the second step 208, the electronic processing unit 186 determines, for each reset time tk of the second set, a multi-satellite correlation or likelihood, which, for each position P(i) of the virtual beacons of the first set and each reset time tk of the second set, is equal to the sum of the squares of the basic correlations between the real signal received by the GNSS receiver at time tk and the replica signal expected at the reset time for the virtual beacon with position P(i) from the satellite Sat(i,j,k) visible from position P(i) at time tk.

[0104] Next, in a fourth step 212, the processing unit determines the position P(i0) of the virtual beacon Be(i0) and the reset time tk0 at which the multidimensional correlation or likelihood function is maximum, and transmits the position P(i0) and the synchronization time tk0 as output data, which is the most likely position of the virtual beacon and the most likely synchronization correction with respect to the location where the train is parked.

[0105] In an optional fifth step 214, following the fourth step 212, the electronic processing unit identifies the storage track for the train using the location of the virtual beacon determined to be closest to the train and a lookup table mapping a set of parking virtual beacons to a set of storage tracks.

[0106] It will be noted that the signal-to-noise ratio of the correlation function estimate can be improved in a conventional manner by incoherently integrating the output over a minimum duration sufficient to achieve a desired maximum estimate confidence.

[0107] It will be noted that the failure to isolate a single solution with sufficient confidence (no clear maximum) indicates an error in the completeness of the method and an alert may be generated. However, such an error may be related to the fact that the exploration interval of the local time is not wide enough; it is then recommended to restart the search while extending the search time interval outside the uncertainty domain of the time estimate.

[0108] according to Figure 4 and Figure 3 In a first specific embodiment of a method for determining a point position of a train, the positioning method 302 comprises Figure 3 a first initialization step 206 of, a first loop 312 of sub-steps 314, 316, 318, 320, 322, wherein a second set of reset times tk parameterized by an index k is scanned, k varying from 1 to Nk, and Figure 3 The fourth step 212.

[0109] Before executing the first loop 312 , in a step 310 of initializing the execution of the first loop 312 , the index k of the iteration of the reset time tk is pre-set to 1.

[0110] In a first sub-step 314, k is set, and at time tk and position P(i) of virtual beacon Be(i), the GNSS signals of the expected GNSS replicas Ci(j,tk) are generated by calculating, for all satellites Sat(i,j,k) visible from virtual beacon Be(i) at time tk, a first set of virtual beacons Be(i) is scanned by an index i (varying from 1 to NBe) in a second loop (not shown). The phases of the expected GNSS replicas Ci(j,tk) for virtual beacon Be(i) at reset time tk are calculated using the ephemeris of the GNSS satellites, designated Pd(j,i,tk) for each visible satellite Sat(i,j,k), where j is the index of the satellite visible from B(i) at time tk.

[0111] In a second sub-step 316 following the first sub-step 314, k is set such that for each position P(i) of the virtual beacon Be(i), i varies from 1 to NBe, and for the virtual beacon Be(i) at time tk (denoted as Γ SC (i, tk)), the multi-satellite or multi-dimensional correlation is calculated as the sum of the squares of the outputs of the matched filtering operation or cross-correlation function of the signals received at time tk from the various satellites visible to the same beacon using the following expression:

[0112] ΓSC (i,tk)=∑ j (Γ SC (j,i,tk)) 2

[0113] in:

[0114] (Γ SC (j,i,tk)) 2 is the GNSS signal SIS received by the GNSS receiver at reset time tk GNSS (tk) and the square of the basic correlation function of the replica signal Ci(j,tk) at the virtual beacon Be(i), at time tk and expected from the visible satellite, which is specified by the index j of the sequence of the set of satellites visible from the virtual beacon Be(i) at time tk.

[0115] In the third sub-step 318, which is continuous with the second sub-step 316, k is set and the multi-satellite correlation at the synchronization reset time tk is performed, ie, Γ SC (i, tk), stored in electronic memory at a first set of virtual beacons Be(i) and their associated geographic locations P(i), described by an index i (varying from 1 to NBe).

[0116] Next, in a fourth sub-step 320 of checking the iterations of the first loop, the iteration index k of the reset time tk is compared with the total number Nk of reset times of the second set.

[0117] In case the index k is strictly below the total number Nk, a fifth sub-step 322 of unit increment of the iterative index k is performed, and then the first, second, third and fourth sub-steps 314, 316, 318 and 320 are performed again.

[0118] In case the iteration index k of the first loop is equal to the total number Nk, the fourth step 212 is performed.

[0119] according to Figure 5 and Figure 3 A second specific embodiment 352 of a method for determining a point position of a train, the method 352 for determining a point position is from a method according to Figure 4 The positioning method 302 of the first embodiment is derived from and differs from it in that the first, second and third sub-steps 314, 316, 318 are replaced by a second loop 356 of sixth, seventh, eighth, ninth and tenth sub-steps 358, 360, 362, 364, 366 of scanning the positions P(i) of the first set of virtual beacons Be(i), the loop being parameterized by an index i varying from 1 to NBe.

[0120] The second loop 356 parameterized by index k is embedded in the first loop 372, which is parameterized by index i varying from 1 to NBe and is associated with Figure 4 The first loop 312 is similar and contains the same sub-steps 320 , 322 .

[0121] Before executing the second loop 356 , in a corresponding initialization step 354 , the index i of the sequence of virtual beacons of the first set is pre-set to 1.

[0122] In a sixth sub-step 358, i and k are set, and the GNSS signals of the expected GNSS replica Ci(j,tk) at time tk and position P(i) of the virtual beacon Be(i) are generated by calculation for all satellites Sat(i,j,k) visible from the virtual beacon Be(i), with index j parameterizing the set of satellites visible to the set. The phase of the expected GNSS replica Ci(j,tk) for the virtual beacon Be(i) at reset time tk is calculated using the ephemeris of the GNSS satellites, specifying Pd(j,i,tk) for each visible satellite Sat(j,i,tk), where j is the index of the satellite visible from B(i) at time tk.

[0123] In the seventh sub-step 360 following the sixth sub-step 358, with i and k being set, at the virtual beacon Be(i) and time tk (denoted as Γ SC The multi-satellite or multi-dimensional correlation is calculated as the sum of the squares of the outputs of the matched filtering operation or cross-correlation function of the signals received from the various satellites visible from the same beacon at time tk using the following expression:

[0124] Γ SC (i,tk)=∑ j (Γ SC (j,i,tk)) 2

[0125] in:

[0126] (Γ SC (j,i,tk)) 2 is the GNSS signal SIS received by the GNSS receiver at reset time tk GNSS (tk) and the square of the basic cross-correlation function for the virtual beacon Be(i), at time tk, and the replica signal Ci(j,tk) expected from the visible satellite, which is specified by the index j of the order of the set of satellites visible from the virtual beacon Be(i) at time tk.

[0127] In the eighth sub-step 362, which is continuous with the seventh sub-step 360, i and k are set, and the multidimensional correlation or likelihood function Γ at time tk is reset.SC (i, tk) is stored in electronic memory.

[0128] Next, in a ninth sub-step 364 of checking the iteration of the second loop, the iteration index i of the position P(i) of the virtual beacon Be(i) is compared with the total number NBe of virtual beacons of the first set.

[0129] In case the index i is strictly lower than the total number NBe of virtual beacons, a tenth sub-step 366 of unit increment of the iterative index i is performed, then the sixth, seventh, eighth and ninth sub-steps 358, 360, 362 and 364 are performed again.

[0130] In case the iteration index i of the second loop is equal to the total number NBe of virtual beacons, a fourth sub-step 320 of checking the iteration index k of the first loop 372 is performed.

[0131] exist Figure 6 In the embodiment, a method for partially implementing the Figure 5 An example of a modular architecture of the system 402 of the second embodiment of the method 352 for determining a point location.

[0132] The modular architecture shown here includes a set of modules 412 for implementing the sixth sub-step 358 and the seventh sub-step 360. These modules may be software modules executed by at least one electronic processing unit, or hardware electronic modules that perform specific functions.

[0133] Using a series of digitally controlled oscillators 4221,…422 j ,…422 Nj (NCO) generates a module for implementing the sixth sub-step 358, the series of numerically controlled oscillators being arranged in parallel to generate, at a local reset time tk, the expected replica PRN code Ci(j,tk) of the satellite in the set of visible satellites represented by index j and visible from the position P(i) of the virtual beacon Be(i), j varying from 1 to Nj, Nj being the total number Nj of visible satellites.

[0134] The module for implementing the seventh sub-step 360 is generated using:

[0135] - a series of modules 4321, ..., 432 j ,…,432 Nj , for computing in parallel a basic cross-correlation function between the GNSS signal received by the GNSS receiver at the reset time tk and a copy of the PRN code expected for the virtual beacon Be(i) from the satellites visible from said beacon Be(i) at the time tk, and - modules 4421, ..., 442 j ,…,442 Nj, each raising the input to a power of 2, followed by an Nj-input adder 444 for determining the sum of the squares of the NL basic cross-correlation functions.

[0136] The sum of the squares of the outputs of the matched filtering operations or basic cross-correlation functions on the various satellites visible to a given virtual beacon Be(i) at a given synchronization time tk is an important feature of the present invention. In particular, since the coherence of the carrier phase cannot be guaranteed with sufficient accuracy due to random propagation effects (delays, multipath), it is recommended not to coherently sum the outputs of the customized filtering operations or cross-correlation functions on the various satellites visible to a given beacon. In addition, after correcting for the modeled delays, the residual errors in the PRN code have a very small impact on the value of the multidimensional correlation function (only a small error of a few meters relative to the width of the correlation span), which makes the implementation of the incoherent or sum-of-squares coherent.

[0137] Finally, the summation of the correlation functions makes it possible to eliminate the uncertainty of the local time alone, since the projection of the time error of the local clock (for the pseudoranges measuring the boresight directions of all satellites) remains coherent (within the time error) with the single true position of the receiver, while for positions that do not correspond to the actual received position, the time error is projected incoherently in the boresight direction.

[0138] The spatial correlation field depends on the direction of incidence of the GNSS signal relative to an alignment axis passing between the observed virtual beacon B(i) and the beacons of the first set of virtual beacons and said observed beacon B(i).

[0139] Since a signal that is "orthogonal" to the motion cannot reduce the position uncertainty, one may choose to define an equivalent value of the DOP (Dilution of Precision) factor to evaluate the ability to use a virtual beacon as an "absolute reference" while taking into account the geometry of the satellite.

[0140] It should be recalled that for the standard three-dimensional solution, a matrix of direction cosines of the angles of arrival is used, which actually enters into the system of equations to be solved.

[0141] In the case of one-dimensional solutions, an equivalent strict criterion can be constructed. This criterion can take the following form:

[0142]

[0143] in:

[0144] -DOP is an equivalent standard,

[0145] -Nsat is the number of positioning satellites,

[0146] -α j is the angle of arrival or incidence of the jth satellite, and j is an index that varies sequentially from 1 to Nsat.

[0147] For example, the satellites considered may be restricted to those having an angle of incidence less than 60° relative to the alignment direction between the observed virtual beacon and the nearest virtual beacon of the first set, which corresponds to a maximum elongation factor of 2 for the correlation span.

[0148] By way of explanation, Figure 7 An example of the results obtained after summing the squares of the satellite spatial cross-correlation functions for various events is shown. Curves 471 to 473 are graphical representations of the correlation functions for the satellite signal's angle of incidence relative to the direction of motion of 60°, 45°, and 30°, respectively. Curve 474 is a graphical representation of the sum of the satellite correlation functions.

[0149] Advantageously, the incoherent summation of the cross-correlation functions makes it possible to improve the signal-to-noise ratio C / N0 or the detection gain by 5.log(Nsat) relative to conventional GNSS detection, Nsat being the number of satellites visible from the possible position involved, thus providing better sensitivity and better precision in terms of detection time and position of maximum correlation.

[0150] Advantageously, during the calculation of the multidimensional correlation function or likelihood function for a given position P(i) of a virtual beacon Be(i) and a given synchronization reset time tk, the basic cross-correlation functions obtained for the various satellites Sat(j) visible from the virtual beacon Be(i) at time tk are optionally weighted, taking into account the geometrical line of sight matrix of the visible satellites.

[0151] Typically, the elementary correlation functions are not all centered exactly at the same time (corresponding to the maximum power), and differences in the synchronization of the signals received from the satellites (synchronization differences between different GNSS satellite clocks, propagation delays due to the ionosphere and troposphere, and multipath of the signal transmitted by a given GNSS satellite) can cause errors in the position of the spatial correlation function.

[0152] These differences are of the order of tens of nanoseconds at most, for example 200 ns corresponding to a pseudorange of 60 m, leading to a spread in the position of the correlation maximum of the visible satellites.

[0153] In order to mitigate the effects of this spread, and therefore the noise encountered when determining whether the maximum of the correlation function has been passed, it is necessary to correct the expected ranges of the visible satellites during the generation of the local GNSS copy of the expected PRN code using available error models, that is, known models of clock errors and errors caused by the ionosphere and troposphere. These error models can be provided by the onboard GNSS receiver or through an auxiliary link and allow the error in the expected range of the satellites to be reduced to a few meters. These error models do not allow correction of errors caused by multipath and interference in the PRN code signal received from a given visible satellite.

[0154] In addition to the detection gain obtained by the sum of the squares of the basic cross-correlation functions on a set of satellites visible from a given virtual beacon Be(i) at a given synchronization reset time tk, the method for determining the spatial position according to the invention improves the robustness of the detection of the virtual beacon Be(i0) closest to the GNSS receiver located on the train with respect to mirror multipath, and therefore improves the robustness of the identification of the storage track Vi0 on which the train is located with respect to mirror multipath.

[0155] It is noteworthy that the position of the correlation maximum remains unchanged in the presence of specular multipath, which is delayed with respect to the direct signal, despite the fact that the time correlation function is deformed and thus becomes asymmetric, which illustrates the robustness of the position of the correlation maximum with respect to the presence of multipath.

[0156] Advantageously, the method for determining the point position of a virtual beacon according to the invention exploits this property to perform an estimate of the "highest" position likelihood, which is based solely on the estimate of the position of the correlation maximum and is more robust than a simple direct position estimate based on a conventional reception time estimate using a discriminator that is sensitive to distortions of the correlation function.

[0157] It is also worth noting that the spatial correlation domain, which defines the positional resolution capability of the maximum, is directly related to the waveform of the code and the spreading technique used.

[0158] For example, in the case of GPS C / A codes, the spatial correlation spans 300m, and for a 5m spacing between beacons, nearby locations in this domain can be compared with a resolution sensitivity of 0.15dB.

[0159] When using the Galileo BoC (1,6) PRN code, the correlation span to the first zero is 25m, and the resolution sensitivity is improved, up to 2dB for a 5m spacing between two nearby virtual beacons.

[0160] Fully utilizing the spectrum of the signals E5a-E5b separated by 30 MHz will allow obtaining a resolution capability of 10 m, ie a resolution sensitivity of 6 dB for a separation of 5 m.

[0161] Therefore, the increase in sensitivity obtained by implementing a method for determining the point position of the virtual beacon (associated with knowledge of the topographic map of the parking location) allows to improve the performance of the detection of the starting track in terms of the probability of false alarms and the probability of non-detection.

[0162] The method for determining the point position of a virtual beacon according to the present invention, by comparing the validity of various parking positions of the vehicle based on the phase consistency of the expected replica signal with the received signal, is significantly different from conventional positioning methods, which are based on a direct estimation of the vehicle's GNSS position, which may be biased, and then projecting it onto a known deposited track map.

[0163] With respect to the standard or conventional operating mode of the GNSS receiver, the method for determining the point position of the vehicle according to the invention, based on individual per-satellite searches, is more robust to multipath since, based on the assumed symmetry of the correlation function, the position of the single maximum correlation threshold observed for each visible satellite, as opposed to the discrimination in time, is not affected by the presence of multipath, which are all delayed with respect to the directly received satellite signal.

[0164] Advantageously, the search for the best spatial match is associated with eliminating the correlation between biases associated with multipath reflections from one receiving location to another, to which the GNSS signal may be subject, thereby reducing the risk of biases remaining stable throughout the spatial correlation function, which, as in the standard method of estimating the time of arrival by advance / delay correlation, would lead to biases in the determination of the location of the maximum correlation peak.

[0165] Preferably, but not restrictively, the method and system for determining a point position according to the present invention can be used in the field of railway transportation for determining the starting position of a train in a set of parking positions known a priori.

[0166] In general, the method and system for determining the position of a point according to the present invention can be applied to any type of locomotive or land vehicle that is stored or parked on a curved or straight segment of a set of curved or straight segments that are sufficiently spaced apart and form a departure track. This may be the case, for example, in the maritime field, when it is necessary to locate ships in a "sea lane," or in the land transport field, when it is necessary to locate mass transit vehicles held in a queue line.

[0167] Preferably, but not restrictively, the present invention will be applied in the field of railway transportation.

[0168] An electronic processing unit is herein understood to mean a system comprising one or more microprocessors, processors, computers or any other equivalent means, suitably programmed to carry out the various computing operations implemented in the context of the method according to the invention.

[0169] Another subject of the invention is a computer program product containing instructions readable by a computer or any other type of equivalent computing device which, when these instructions are executed on a processor, cause the processor to carry out a method for determining the point position of a vehicle stopped on a storage track.

Claims

1. A method for determining a point position of a vehicle stopped on one of a set of storage rails using a virtual beacon, the position determination method being implemented by a system for determining a point position, the system comprising: a GNSS receiver (182) located on the vehicle (154) capable of directly measuring pseudoranges set to a synchronized local time base within a synchronized imprecise time range using geolocation signals transmitted by GNSS geolocation satellites visible to the GNSS receiver; as well as - an electronic processing unit (186) external to the GNSS receiver or integrated into the GNSS receiver, said method for determining the position of a point being characterized in that it comprises a set of steps (204), wherein: - determining and comparing (212) by the electronic processing unit the likelihood of a plurality of hypotheses regarding the stopping position of the vehicle (154) corresponding to a first predetermined set of positions of an integer number NBe of predefined virtual beacons Be(i) by correlating GNSS geolocation signals received by the GNSS receiver (182) on the vehicle (154) at respective times of a second set of synchronized times with predicted GNSS geolocation signals of copies of respective positions of the virtual beacons Be(i) expected at said respective times, i varying from 1 to NBe, the respective positions of the predefined virtual beacons Be(i) being known in the amount of at least one virtual beacon per deposit track, wherein said virtual beacons are predefined positions having known spatial coordinates; and - the detected parking position of the vehicle (154) is the position corresponding to the maximum likelihood; wherein, in order to reduce mismatches due to poor synchronization of the local time reference of the GNSS receiver (182) with respect to the time of the GNSS global positioning system, the imprecise time range of the synchronization of the local clock is divided in a predefined manner into time periods respectively represented by synchronization reset times tk, k varying between 1 and Nk, so that the imprecise time range is covered by the electronic processing unit (186), wherein Nk specifies the number of synchronization reset times within the imprecise time range, and correlation of the GNSS geolocation signals received by the GNSS receiver (182) is performed at synchronization reset times tk.

2. The method according to claim 1, wherein The correlation of the GNSS geolocation signal received by the GNSS receiver (182) at each position P(i) of the virtual beacon Be(i) and at each time tk is a multi-satellite correlation with the corresponding replica GNSS signal expected for each position P(i) at the time tk and from the satellites Sat(i,j,k) visible from the position P(i) of the virtual beacon Be(i) at the time tk, i varying from 1 to NBe and k varying from 1 to Nk, the multi-satellite correlation for the position P(i) at the synchronization reset time tk being equal to the sum of the squares of the basic correlations over the set of satellites visible from the position P(i) at the reset time tk: the actual GNSS signal received by the GNSS receiver for the position P(i) at the time tk and the GNSS replica expected at the position Pi at the time tk and from the satellites Sat(i,j,k) visible from the beacon Be(i) at the time tk, where i varies from 1 to NBe and k varies from 1 to Nk.

3. The method according to claim 1, wherein The set of steps (204) includes an initialization phase, a phase of searching for the position of the virtual beacon closest to the holding position of the vehicle using a predefined strategy, and a phase of utilizing a search phase, in which the storage track Vi0 is identified based on the virtual beacon Be(i0) estimated to be closest to the vehicle and the prior knowledge of the position P(i) of the virtual beacon Be(i), the position P(i) being set and predefined, i varying from 1 to NBe.

4. The method according to claim 3, wherein: The initialization phase in the set of steps (204) comprises a first step (206) in which a first predetermined set of possible positions P(i) of the virtual beacon Be(i) corresponding to possible storage tracks Vi of the vehicle is determined based on a database of terrain data, i varying from 1 to NBe, the database of terrain data being able to be located external to the GNSS receiver or being able to be included in a database memory of the GNSS receiver, and synchronization reset times tk forming a hypothetical second set of synchronization times are determined in the synchronization imprecision time range so as to be substantially regularly spaced, k varying from 1 to Nk, Nk being an integer greater than or equal to 3, the synchronization imprecision time range being included in the time correlation domain of the GNSS PRN code sequence or determined using an integrity protection time radius calculated based on time data of a receiver autonomous integrity monitoring function, the receiver autonomous integrity monitoring function being integrated in the GNSS receiver or being external to the GNSS receiver.

5. The method according to claim 4, wherein The difference between the actual position of the GNSS receiver (182) along the storage track on which the vehicle (154) is parked and the position of the associated virtual beacon is less than the imprecision spatial interval, resulting in an equivalent time impact on time correlation accuracy that is significantly lower than the time correlation domain of the GNSS PRN code sequence used.

6. The method according to claim 3, wherein: The phase of searching for the position of the virtual beacon closest to the maintained position of the vehicle using a predefined strategy comprises a second step (208) in which: - the GNSS receiver (182) receives a real GNSS signal at a synchronization reset time tk, k varies from 1 to Nk, and - an electronic processing unit (186) resets the time tk for each synchronization of said assumed second set of synchronization times and determines, for each possible position P(i) of a virtual beacon Be(i) of said first set of predetermined positions, the GNSS replica signal expected by said GNSS receiver at said position P(i) and time tk for each visible satellite Sat(i,j,k) at said receiver's local time tk, i varying from 1 to NBe; and then - for each reset time tk of the second set of synchronization times and each position P(i) of the virtual beacon of the first set of predetermined positions, the electronic processing unit (186) determines, on the one hand, a basic correlation between the signal of the expected GNSS replica at each holding position P(i) and each synchronization reset time tk and from each satellite visible at each time tk and the actual GNSS signal received by the GNSS receiver (182) located on the vehicle at the synchronization reset time tk.

7. The method according to claim 6, wherein: The phase of searching for the position of the virtual beacon closest to the maintained position of the vehicle using a predefined strategy comprises a third step (210) performed after or in parallel with the second step (208), wherein, for each reset time tk of the second set of synchronization times and for each position P(i) of the virtual beacon of the first predetermined set of positions, the electronic processing unit (186) determines a multi-satellite correlation forming a value of a likelihood function, and wherein, for each position P(i) of the virtual beacon of the first predetermined set of positions and each reset time tk of the second set of synchronization times, the multi-satellite correlation is equal to the sum of the squares of the basic correlations between the GNSS signal received by the GNSS receiver at time tk and the expected replica signal of the position P(i) of the virtual beacon at time tk from the visible satellites Sat(i,j,k) visible from the position P(i) at time tk.

8. The method according to claim 7, wherein: The phase of searching for the position of the virtual beacon closest to the maintained position of the vehicle using a predefined strategy comprises a fourth step (212) performed after the third step (210), wherein the electronic processing unit (186) determines the position P(i0) of the virtual beacon and a reset time tk0 that maximizes the likelihood function of the positions P(i) of the first set of predetermined positions and the reset time tk of the second set of synchronization times, whereby the position P(i0) is determined as the position of the virtual beacon Be(i0) detected as closest to the GNSS receiver.

9. The method according to claim 8, further comprising a fifth step (214) performed after the fourth step (212), wherein: - the electronic processing unit (186) identifies the storage track on which the vehicle is stopped based on the position P(i0) of the virtual beacon detected as being closest and a lookup table mapping all virtual beacons to all storage tracks; and / or - the electronic processing unit (186) transmits information about the detected synchronization reset time tk0 to a mechanism for correcting the drift of the local clock of the GNSS detector; and / or - the electronic processing unit (186) issues an alarm if the storage track cannot be identified and starts a search in a wider field of time synchronization inaccuracy.

10. The method according to any one of claims 6 to 9, wherein: For each position P(i) of the virtual beacon, the set of satellites visible from all virtual beacons Be(i) at a given time tk is identical and independent of the index i of the sequence of the virtual beacons and depends only on the time tk.

11. The method according to any one of claims 1 to 9, wherein: The integrity of the position is checked by an additional step in which it is verified that the maximum likelihood is above a predefined safety threshold guaranteeing the integrity of the position.

12. An apparatus for determining a point position of a vehicle stopped on one of a set of storage tracks using a virtual beacon, comprising: a GNSS receiver (182) located on the vehicle capable of measuring pseudoranges set to a synchronized local time base directly within a precise time frame using geolocation signals emitted by GNSS geolocation satellites visible to the receiver; and - an electronic processing unit (186) external to or integrated into the GNSS receiver, the device for determining the position of a point being characterized by: The electronic processing unit (186) is configured to: - determining and comparing the likelihoods of a plurality of hypotheses about the stopping position of the vehicle corresponding to a first predetermined set of positions of an integer number NBe of predefined virtual beacons Be(i) by correlating GNSS geolocation signals received by said GNSS receiver on board said vehicle at respective times of a second set of synchronized times with predicted GNSS geolocation signals of expected copies of respective positions of virtual beacons at said respective times, i varying from 1 to NBe, for said first predetermined set of positions of predefined virtual beacons Be(i), the respective positions of said predefined virtual beacons Be(i) being known in the amount of at least one virtual beacon per deposit track, wherein said virtual beacons are predefined positions with known spatial coordinates; and - the detected current storage track position of the vehicle is the position corresponding to the maximum likelihood; In order to reduce the mismatch caused by poor synchronization of the local time reference of the GNSS receiver with respect to the time of the GNSS global positioning system: - the electronic processing unit (186) is configured to cover the imprecise time range of the local clock by dividing the imprecise time range in a predefined manner into time periods respectively represented by synchronization reset times tk, k varying between 1 and Nk, wherein Nk specifies the number of synchronization reset times within the imprecise time range, the correlation of the GNSS geolocation signals received by the GNSS receiver being performed at synchronization reset times tk; and - the correlation of the GNSS geolocation signal received by the GNSS receiver (182) at each position P(i) of the virtual beacon Be(i) and at each time tk is a multi-satellite correlation with the corresponding replica GNSS signal expected for each position P(i) at time tk and from satellites Sat(i,j,k) visible from the position P(i) of the virtual beacon Be(i) at time tk, i varying from 1 to NBe and k varying from 1 to Nk, the multi-satellite correlation for the position P(i) at the synchronization reset time tk being equal to the sum of the squares of the basic correlations over the set of satellites visible from the position P(i) at the reset time tk: the actual GNSS signal received by the GNSS receiver for the position P(i) at time tk and the GNSS replica expected at time tk and from satellites Sat(i,j,k) visible from the beacon Be(i) at time tk, where i varies from 1 to NBe and k varies from 1 to Nk.

13. The device for determining a point position of a vehicle according to claim 12, wherein The electronic processing unit (186) is configured to determine the synchronization imprecision time range by including the synchronization imprecision time range in a time correlation field of a GNSS PRN code sequence or by calculating an integrity protection time radius based on time data of a receiver autonomous integrity monitoring function, the receiver autonomous integrity monitoring function being integrated into the GNSS receiver or being external to the GNSS receiver.

14. A system for determining a point location of a vehicle stopped on a storage track in a set of storage tracks using a virtual beacon, comprising: - GNSS Global Satellite Positioning System (172), and - Device for determining a point position of a vehicle according to any one of claims 12 to 13; The vehicle (154) is: - a land vehicle parked on a land storage track, or - Maritime transport vehicles stored on sea lanes.

15. The system for determining a point position of a vehicle according to claim 14, wherein the means for determining the point position is enhanced by a receiver autonomous integrity monitoring function implemented within the GNSS receiver or within a second GNSS receiver separate from the first GNSS receiver.

16. The system for determining a point position of a vehicle according to claim 15, wherein the vehicle comprises a train stored on a storage track.

17. A computer program product comprising computer readable instructions which, when executed on a processing unit, cause the processing unit to perform the method for determining a point position of a vehicle stopped on one of a set of storage rails using a virtual beacon according to any one of claims 1 to 10.

Images