Device and method for autonomous positioning of a moving vehicle on a railway track

Abstract

Embodiments of the present invention provide a method for locating a mobile vehicle traveling on a railway track of a railway network in a railway reference system. The method determines one or more railway track segments where the mobile vehicle may be located by querying a map database based on at least one position parameter. The map database includes data representing a vector description of at least one railway track of the railway network and data representing descriptions of one or more objects in the environment adjacent to the railway track of the railway network that may interfere with navigation signals received by the mobile vehicle. The method comprises the steps of generating the map database, which comprises the following steps: - receiving (201) a cloud of survey points classified and geolocated in a coordinate system, the cloud of survey points being associated with the railway network and the adjacent environment, - determining (203) a topology of the railway network based on the geolocated and classified survey points associated with the railway network and a plurality of additional elements of the railway network, - modelling (204) the geometry of railway tracks of the railway network into a plurality of modelled railway track segments, - determining (205) one or more neighbouring objects from the geolocated and classified survey points associated with the adjacent environment, each of the neighbouring objects being associated with one or more modelled railway track segments, - identifying (206) for each association between a neighbouring object and one or more modelled railway track segments information representing interference affecting the navigation signal.

Description

Technical Field

[0001] The present invention generally relates to positioning systems, and in particular to an apparatus and method for autonomously positioning a moving vehicle in a railway network comprising at least one railway using a map database. Background Art

[0002] The positioning of railway vehicles travelling on railway tracks of a railway network is usually performed in a non-autonomous manner using sensors deployed along the railway tracks, so-called track sensors. The railway vehicle can be positioned using odometry between two consecutive track sensors. By determining the railway segment on which the railway vehicle is located, positioning can be performed in a railway reference system. If the risk that the railway vehicle is not on the relevant segment is below an acceptable limit, the returned position can be considered reliable. The smaller the distance between two consecutive track sensors, the higher the reliability of the positioning technique combining track sensors and odometry. However, such known positioning techniques are expensive to deploy and involve high maintenance costs.

[0003] Among other existing approaches, integrated and autonomous positioning systems have been proposed that do not require the deployment of tracking sensors. The position provided by such positioning systems can be generated in a terrestrial reference system by means of positioning equipment installed on railway vehicles, which receives and processes navigation signals transmitted by geolocation satellites such as GPS (Global Positioning System) and GNSS (Global Navigation Satellite System) satellites. The conversion from a position in a terrestrial reference system to a position in a railway reference system can be performed using a map database of the railway network. Such a map database includes the geographical coordinates of various elements of the railway network, such as railway tracks. The map database may also include a description of certain elements of the environment adjacent to the railway network that may interfere with the navigation signals received by the railway vehicles. Therefore, the credibility of such an autonomous positioning system depends on the credibility of the map database used.

[0004] Building a reliable map database requires performing a topographic survey of the railway network and its immediate surroundings before processing the resulting cloud of survey points. This allows for modeling of elements of the railway network and certain objects in the immediate surroundings. This processing requires classifying each survey point according to whether it represents the railway network or its immediate surroundings.

[0005] Measurement points are typically classified manually. However, this manual classification is complex and time-consuming to perform. The classified points must then be processed by equipment that often lacks sufficient storage memory and computing speed to effectively handle the large amounts of data (on the order of terabytes) corresponding to these classified points. This results in reduced reliability and response time for vehicle positioning equipment using such databases.

[0006] Modeling the tracks of a railway network is usually done using computer-aided design (CAD) tools that were originally designed for modeling road infrastructure. Railway applications require sub-decameter accuracy when acquiring terrain data, as well as a protection radius representing an error ellipse of several meters. However, railway applications have very high reliability requirements (approximately 10 per hour). -7 Furthermore, modeling the railway tracks is insufficient for reconstructing the topology of the railway network, as it requires identifying other elements of the network, such as intersections. Design tools require users to have higher levels of professional skills and expertise. Numerous operations are required to extract information from the collected measurements.

[0007] Furthermore, considering the particularity of railway applications, existing tools cannot effectively reconstruct the topology of the railway network from geolocated and classified survey point clouds.

[0008] To model certain objects in the environment adjacent to the railway network, learning-based tools are available that can extract and model these objects from geolocated and classified survey point clouds. These tools allow for the identification of objects in the adjacent environment. However, their functionality is limited to extracting these objects and cannot analyze their impact on the navigation signals received by railway vehicles moving near these objects.

[0009] Therefore, there is a need for an improved apparatus and method for locating vehicles in a railway network. Summary of the Invention

[0010] The present invention aims to improve the above situation by proposing a method for positioning a moving vehicle travelling on a railway track of a railway network in a railway reference system, said method comprising the following steps:

[0011] - determining, based on a plurality of navigation signals received by the mobile vehicle, position parameters associated with the position of the mobile vehicle in a geographical reference system,

[0012] - determining one or more railway track segments at which the mobile vehicle may be located by querying a map database based on at least one of the position parameters,

[0013] - positioning the mobile vehicle based on at least one of the railway track segment and the position parameter provided by the map database,

[0014] The map database comprises data representing a vector description of at least one railway track of a railway network and data representing a description of one or more objects of an environment adjacent to the railway track of the railway network that may interfere with navigation signals received by the mobile vehicle, the method for generating the map database comprising the following steps:

[0015] - receiving a cloud of measured points classified and geolocated in a coordinate system, said cloud of measured points being associated with said railway network and said adjacent environment,

[0016] - determining the topology of the railway network based on geolocated and classified measurement points associated with the railway network and a plurality of additional elements of the railway network,

[0017] - modelling the geometry of a railway track of said railway network as a plurality of modelled railway track segments,

[0018] - determining one or more neighboring objects from geolocated and classified measurement points associated with the neighborhood, each of the neighboring objects being associated with one or more modeled railway track segments,

[0019] - identifying, for each association between a neighboring object and one or more modeled railway track segments, information representative of interference affecting the navigation signal.

[0020] In an embodiment, the method further comprises the step of saving the element comprising the modelled railway track segment, the additional element of the railway network and / or the neighbouring objects in a file having a given representation format.

[0021] Additional elements of the rail network may include junctions and stops.

[0022] In one embodiment, the geolocated and classified survey point cloud may be received from at least one mobile terrain system associated with a hybrid geolocation system comprising a positioning device associated with an inertial unit.

[0023] The position parameters may include position in three dimensions, movement speed and / or orientation parameters.

[0024] Preferably, then, the step of determining the topology of the railway network may comprise the following steps:

[0025] - converting elements of a geolocated point cloud associated with the railway network into a plurality of geospatial vectors, each of the geospatial vectors corresponding to a railway track,

[0026] - Segmenting the rail network by defining for each geospatial vector a plurality of elements comprising intersections and / or stops.

[0027] In one embodiment, each railway track may include two railway lines, and the step of vectorizing the geolocated and classified measured point cloud associated with the ground and the railway track into a plurality of geospatial vectors may include the following steps:

[0028] - identifying points of the geolocated point cloud associated with a selected railway line of said railway track,

[0029] - subsample the identified points using a predefined subsampling factor,

[0030] - grouping the sub-sampled points into subsets of points, two consecutive subsets of points sharing at least one measurement point,

[0031] - using a moving average to calculate said average position in three dimensions for each subset of points,

[0032] - Create a vector associated with the selected railway line, including the calculated average position in three dimensions,

[0033] - eliminating those components of the created vector that may be affected by noise coming from the automatic classification, which provides a geospatial vector,

[0034] - determining one or more vectors associated with another railway line of said railway track based on said geospatial vector determined for said selected rail, and

[0035] - Calculating a geospatial vector associated with the median axis of the railway track.

[0036] Specifically, each geospatial vector may also include bearing, slope, and tilt measurements extracted from each identified subset of points.

[0037] In one embodiment, the subdividing step may use a graph comprising a set of edges representing railroad tracks associated with geospatial vectors and nodes connecting the edges, the edges representing intersections or stops on the railroad network.

[0038] The step of modeling the geometry of said center axis of the railway track represented by a geospatial vector may comprise one or more iterations of the following steps:

[0039] i. determining a current analytical model of the geometry of the center axis of the railway track represented by a geospatial vector using at least one analytical equation,

[0040] ii. measuring the standard deviation between the median axis of the railway track and the associated curve returned by the analytical model,

[0041] iii. If the standard deviation of the measurement is greater than a predefined error threshold, then splitting the geospatial vector into two elements,

[0042] Steps i to iii are iterated as long as the standard deviation between the analytical model and the geospatial vector is larger than the predefined error threshold.

[0043] In one embodiment, the method further comprises determining a network format corresponding to the railway network by associating each modeled railway track segment with one or more elements of the railway network.

[0044] The steps of defining neighboring objects and defining associations between the neighboring objects and one or more modeled segments may comprise the following steps:

[0045] - subsampling said measurement points associated with said neighbourhood by a predefined subsampling factor,

[0046] - converting the subsampled measurement points into one or more neighboring objects,

[0047] - associating the neighboring objects with the modeled railway track segment.

[0048] Navigation signal propagation characteristics may be assigned to each association between a neighboring object and one or more modeled segments based on a number of parameters including the neighboring object's height and the neighboring object's distance from a medial axis of the modeled segment.

[0049] A system for positioning a moving vehicle travelling on a railway track of a railway network in a railway reference system is also proposed, the system comprising a positioning device, the system being configured to:

[0050] - determining, based on a plurality of navigation signals received by the mobile vehicle, position parameters associated with the position of the mobile vehicle in a geographical reference system,

[0051] - determining one or more railway track segments at which the mobile vehicle may be located by querying a map database based on at least one of the position parameters,

[0052] - positioning the mobile vehicle based on at least one of the railway track segment and the position parameter provided by the map database.

[0053] The map database comprises data representing a vector description of at least one railway track of the railway network and data representing a description of one or more objects in the environment adjacent to the railway track of the railway network that may interfere with the navigation signal received by the mobile vehicle. The positioning system further comprises means for generating the map database, comprising:

[0054] a receiving module configured to receive a cloud of measured points classified and geolocated in a coordinate system, said cloud of measured points being associated with the railway network and the adjacent environment,

[0055] a topology determination module configured to determine the topology of the railway network based on geolocated and classified measurement points associated with the railway network and a plurality of additional elements of the railway network,

[0056] a modelling module configured to: model the geometry of a railway track of the railway network as a plurality of modelled railway track segments,

[0057] a module for determining neighboring objects, configured to determine one or more neighboring objects based on geolocated and classified measurement points associated with the neighboring environment, each of the neighboring objects being associated with one or more modeled railway track segments,

[0058] - an identification module configured to identify, for each association between a neighboring object and one or more modeled railway track segments, information representative of interference affecting the navigation signal.

[0059] Thus, embodiments of the present invention provide methods and apparatus for achieving accurate and reliable positioning of a mobile vehicle traveling on railroad tracks in a railway network. Such positioning can be guaranteed at any point in the railway network and at any time. By taking into account the specific characteristics of the railway network in terms of accuracy and reliability, such positioning is advantageously performed with optimal response time and reduced computational complexity. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Other features and advantages of the present invention will become apparent with the help of the following description and accompanying drawings, in which:

[0061] - Figure 1 A system for locating a moving vehicle traveling in a railway network according to an embodiment of the present invention is shown.

[0062] - Figure 2 is a flow chart showing the steps of a method for generating a map database,

[0063] - Figure 3is a flow chart illustrating steps of modeling implemented to determine the topology of a railway network and the geometry of the railway tracks of the network according to some embodiments of the present invention,

[0064] - Figure 4 is a flow chart illustrating steps implemented to determine a geospatial vector based on a geolocated and classified survey point cloud according to some embodiments of the present invention,

[0065] - Figure 5 is a flow chart illustrating steps implemented to model the geometry of a segment of a railway according to some embodiments of the present invention,

[0066] - Figure 6 is a flow chart illustrating steps implemented to achieve association between a modeled railway track segment and objects of the adjacent environment according to some embodiments of the present invention,

[0067] - Figure 7 and Figure 8 shows two examples of associations between modeled railway track segments and objects of the adjacent environment according to an embodiment of the present invention, and

[0068] - Figure 9 is a positioning system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0069] Figure 1 An autonomous positioning and navigation system 100 is shown that may utilize an integrated map database 102 according to some embodiments of the present invention.

[0070] The autonomous positioning and navigation system 100 can be used in a railway vehicle traveling on a railway track of a railway network to provide the railway vehicle's position and navigation data that can be used for navigation in an autonomous mode. More generally, the autonomous positioning and navigation system 100 can be used in any type of moving vehicle (e.g., a train) traveling on a path confined by a dedicated track.

[0071] The autonomous positioning and navigation system 100 may further include a positioning device 101 and a tracking device 103 .

[0072] The positioning device 101 may include a navigation signal receiver configured to receive navigation signals sent by a positioning system (e.g., a GPS or GNSS system). The positioning device 101 may use such signals to determine position parameters associated with a moving vehicle. The position parameters may include position data representing the position of the vehicle in three dimensions in a global geographic reference system linked to the earth, and / or the orientation of the vehicle and / or the moving speed of the vehicle. For each position in the determined three dimensions, the positioning device 101 may be associated with an error ellipse and / or an alarm signal that may be triggered if conditions associated with the position determined about the ellipse are met. In one embodiment, the alarm signal is triggered if the position of the vehicle in question is within the error ellipse. The alarm signal may be stopped when the position of the vehicle in question moves out of the error ellipse after a predetermined period of time.

[0073] The positioning device 101 may also include an inertial unit configured to provide less accurate position parameters than those determined from the navigation signal without using any external information. The inertial unit can be used in situations where no navigation signal is available to allow for continuous positioning of the moving vehicle. This situation may occur, for example, when the moving vehicle is traveling through a tunnel. The inertial unit can also be used in situations where the navigation signal quality deteriorates. This degradation may result from multipath propagation caused by objects in the vicinity of the moving vehicle. Objects in the vicinity that may cause this degradation include, for example, buildings, vegetation, and traffic signs.

[0074] In one embodiment of the present invention, the positioning device 101 may further include a module for mixing positioning parameters provided by the navigation signal receiver and the inertial unit respectively.

[0075] The integrated map database 102 is configured to store a data set describing the topology of a railway network (hereinafter referred to as "topology description data"). The railway network topology includes at least one railway track on which mobile vehicles can travel. The data of the integrated map database 102 may also include data describing the format of such a network (hereinafter referred to as "network format description data"). The topology and network format of the railway network are key elements for autonomous navigation of mobile vehicles. As used herein, the term "autonomous navigation" refers to navigation in a vehicle without an active driver.

[0076] The integrated map database 102 may also include data describing certain objects in the environment adjacent to the tracks of the railway network that may interfere with navigation signals received by a moving vehicle (hereinafter referred to as “environmental object description data”).

[0077] The topographic description data may include data identifying some of the constituent elements of the railway network associated with the database, such as railway tracks, intersections, and / or stops (also known as "terminals"). The identification data associated with the constituent elements of the network may also include interconnection data representing the interconnections between the various constituent elements.

[0078] Each element of the railway network described by data from the database 102 may be associated with position data associated with the positioning device 101 in a global geographic reference system linked to the earth.

[0079] A network format corresponding to a railway network can represent a set of possible routes in the railway network, where, for a given departure point and a given arrival point, each route comprises a set of segments corresponding to the railway tracks of the network. The network format can be determined based on the characteristics of the links in the network by, for example, identifying, for each railway track connecting the departure point and the arrival point, the railway track segments that constitute it and the intersections encountered by a moving vehicle traveling in the railway network. Intersections are defined relative to the direction of travel and can be considered as zero-length segments with one entrance and two exits.

[0080] Tracking device 103 can be configured to determine the position of the mobile vehicle in the railway reference system by identifying the railway segment on which the mobile vehicle is located. The tracking device can receive the position parameters provided by positioning device 101 and then query the integrated map database 102 using an input request including at least some of these parameters to identify candidate railway track segments on which the mobile vehicle may be located. Tracking device 103 can use navigation rules in the railway network to determine the railway track segment on which the mobile vehicle is located. Tracking device 103 can also be configured to determine the horizontal coordinate of the mobile vehicle on the current railway track segment in the reference system in question.

[0081] Embodiments of the present invention advantageously allow for semi-automatic generation of the integrated database 102 .

[0082] Figure 2A method for generating a database according to some embodiments of the present invention is shown, which is implemented to generate an integrated map database 102. In step 201, a geolocated and classified survey point cloud describing a railway network and its adjacent environment in three dimensions is received. Each survey point can be geolocated in the sense that its position in a given survey reference frame is precisely known. Such a survey reference frame can be a terrestrial survey reference frame. The coordinates of the survey points can be provided in a GPS / GNSS tracking file. The geolocated survey point cloud can be obtained from a topographic survey previously conducted using one or more topographic techniques. Examples of topographic techniques include, but are not limited to, mobile mapping systems or MMS, topography, satellite topography, and traditional topographic techniques that implement mobile topographic equipment. An advantage of MMS is that, by using a hybrid geolocation system comprising a GPS / GNSS positioning device associated with an inertial unit, very high-quality measurements (less than 10 centimeters) can be made, thereby improving geolocation quality. MMS can use multiple optical cameras and laser remote sensing devices to construct a three-dimensional model of the analyzed environment by implementing specialized tools and vision algorithms.

[0083] Each geolocated measurement point may be further classified into at least two groups of geolocated measurement point clouds, including a group of point clouds associated with the ground and railway network, and a group of point clouds associated with the adjacent environment.

[0084] This classification into two groups of measurement point clouds can be performed by applying one or more classification algorithms. Such classification algorithms can be executed on a computer system or cloud computing service. Compared to traditional classification methods performed manually by operators, this classification of geolocated measurement points saves time and computing resources and improves accuracy.

[0085] In step 201, coordinates of certain elements of the railway network in the same measurement reference system as the geolocated measurement point cloud may also be received. Such elements may include intersections and stops.

[0086] In step 201, a subsampling operation may be applied to the received geolocated and classified measurement point cloud. Such a subsampling operation allows for a reduction in the size of the geolocated and classified measurement point cloud processed by the method for generating a database. The subsampling factors associated with such a subsampling operation may be the same for a three-dimensional geolocated and classified measurement point cloud. Alternatively, a subsampling factor may be associated with each dimension of the measurement point cloud.

[0087] The geolocated measurement point cloud associated with the ground and the railway network may be separated from the geolocated measurement point cloud associated with the environment adjacent to the railway network in step 202. This separation advantageously enables parallel processing of the two geolocated measurement point clouds.

[0088] In step 203, the topology of the railway network is determined based on the geolocated and classified survey point cloud associated with the ground and the track. Determining the topology of the railway network may include identifying the railway tracks, intersections, and stops that comprise the railway network. The railway tracks may be modeled using geospatial vectors. Each geospatial vector may include coordinates of a central axis of the associated railway in a survey reference system.

[0089] Intersections can be identified by analyzing intersections between various railroad tracks identified as being modeled by geospatial vectors. Stops can be identified by detecting the ends of the identified railroad tracks. According to an embodiment of the present invention, the positions of the intersections and stop points in the measurement reference system can be provided by step 201.

[0090] At step 204, a railroad track represented by a geospatial vector associated therewith is received so that the geometry of the railroad track can be modeled using a standard or universal geometry. Modeling the geometry of the railroad track can include subdividing the railroad track into a plurality of track segments. An analytical equation can be associated with each track segment to describe its geometry. An association can be determined between each modeled railroad track segment and elements of the railroad network to which it is connected (other segments, intersections, stops, etc.), which defines the format of the railroad network.

[0091] Step 205 is performed based on a geolocated cloud of measured points associated with a neighboring environment and a modeled railway track segment is received. The environment neighboring the railway network may have a width of one hundred meters centered on a railway track of the railway network. Such an environment may include stations, platforms, bridges, tunnels, etc. The cloud of measured points associated with the neighboring environment may be converted into a plurality of objects of the neighboring environment. The objects of the neighboring environment may be modeled using standard or common 3D geometries by determining a set of geometric parameters for each object of the modeled neighboring environment, which include the size of the object and the distance between the object and the modeled railway track segment. In step 205, an association is determined between each modeled railway track segment and objects of the modeled neighboring environment that may interfere with navigation signals received by a moving vehicle traveling on the railway track segment in question.

[0092] In step 206, the modeled railway track segment, the identified objects of the adjacent environment, and the associations between the segment and the objects provided in step 205 are used to characterize the propagation risk affecting navigation signals received by a mobile vehicle traveling on the railway track segment. Characterization of the propagation risk can be performed at several points on each modeled railway track segment associated with one or more objects of the adjacent environment. Characterization of the propagation risk provides propagation risk parameters such as, but not limited to, the type of propagation risk and the size of the mobile vehicle associated with the identified propagation risk.

[0093] According to one embodiment of the present invention, elements of a map database may be described in one or more description files having a selected representation format, such as XML. These database elements may include modeled railway track segments, intersections, stops, objects in the surrounding environment, and associations between the modeled segments and objects in the surrounding environment.

[0094] Figure 3 The following steps are shown according to some embodiments of the present invention to generate a railway track segment that is modeled based on a geolocated and classified survey point cloud associated with the ground and the railway track. Figure 2 The point cloud can be measured using a mobile terrain system associated with an inertial unit. The coordinates of the measured points can be organized in geolocation files with a GPS / GNSS track structure. Each geolocation file can contain a complete set of data describing the coordinates of the measured points associated with a railway. The coordinates of secondary railways can be described separately in other geolocation files.

[0095] In step 301 , geolocated measurement points associated with the ground and with a railway track are received, as well as locations of intersections and stops on the railway track.

[0096] In step 302, the received geolocated measurement points are converted into one or more geospatial vectors. Each geospatial vector can be associated with a railway track and can include the three-dimensional coordinates of a predetermined number of points on the rails of the associated railway track. Alternatively, each geospatial vector can include the three-dimensional coordinates of a predetermined number of points on the center axis of the associated railway track. The spacing between points on the track or the spacing from the center axis can be constant. The geospatial vectors can also include heading, grade, and inclination measurements associated with each point on the track or the center axis of the railway track. Step 302 can be implemented by applying a vision algorithm executed on a local computer or a cloud computing service.

[0097] In step 303, the geospatial vectors associated with the various railroad tracks are subdivided to identify other elements of the railroad network, including intersections and stops. These elements can be identified based on the coordinates of the elements provided by step 301. The subdividing step 303 can also include constructing a graph representing the topology of the railroad network. The graph can include a set of edges, nodes connecting the edges, and leaf nodes connected to individual edges. The edges of such a graph can represent railroad track segments associated with the geospatial vectors, and the nodes can represent intersections or stops (leaf nodes) in the railroad network. In the subdividing step 303, for example, the railroad track segments connected to each node (intersection and stop) of the railroad network can be determined, as well as the segments associated with each railroad.

[0098] In step 304, the geometry of the railroad tracks of the railroad network represented by the geospatial vectors can be modeled using analytical equations. Modeling step 304 can include subdividing each geospatial vector of the railroad network into a plurality of segments to facilitate modeling. A standard deviation measuring the error between a measured curve of the railroad track and a curve derived from the analytical equations can be determined. Modeling step 304 can also be configured such that the standard deviation associated with each segment of the railroad network is less than a predefined modeling error threshold.

[0099] In step 305, the railway network is partitioned based on the modeled railway track segments and other elements of the railway network. For example, step 305 may include defining the switching direction at each intersection of the railway network and the travel direction of moving vehicles on each modeled railway track segment. Step 305 may also include saving the elements of the railway network and the format of the network in a description file having a selected description format (e.g., XML format).

[0100] Figure 4 is a flow chart illustrating steps implemented to create a geospatial vector based on a geolocated and classified survey point cloud associated with the ground and a railway track, according to some embodiments of the present invention. In such embodiments, the geospatial vector represents the center axis of the railway track, which specifies a guide rail for a railway vehicle (e.g., a train) comprising two rail lines, the spacing of which is maintained constant by being fixed to a crossbar. According to other embodiments of the present invention, the geospatial vector may represent one or more rail lines of the railway track.

[0101] In step 401, one of the rail lines of each rail track is selected by analyzing the associated measurement points and their coordinates provided by the geolocation file of the measurement point cloud. For a rail track comprising two rail lines, for example, the right rail defined relative to the direction of movement of the moving vehicle may be selected. The remainder of the description will be provided with reference to such an example of rail line selection by way of non-limiting example.

[0102] The measurement points associated with the selected railway line are then subsampled in step 402. The subsampling factor may be of the order of ten points. Such a subsampling factor constitutes a trade-off between measurement accuracy and computational complexity.

[0103] In step 403, the subsampled measurement points are grouped into subsampled measurement point subgroups such that two consecutive subsampled measurement point subgroups share at least one measurement point. This overlap between measurement point subgroups ensures the continuity of the measurements to be extracted.

[0104] In step 404, a moving average is applied to the subset of subsampled measurement points to calculate an average position in three dimensions for each subset. Such an average position may be defined in the same measurement reference system associated with the geolocated measurement point cloud.

[0105] In step 405, a geospatial vector representing the right side track of the railway track is created by combining the calculated average positions in three dimensions while observing the order in which the positions were obtained.

[0106] In step 406 , the geospatial vectors thus obtained are filtered to eliminate locations resulting from measurement points that may be affected by noise from the automatic classification.

[0107] In step 407 , survey points associated with the left side rail line of the railway track are analyzed in a similar manner to the right side rail line of the same railway track to determine a second geospatial vector representing the left side rail line.

[0108] In step 408, a geospatial vector representing the center axis of the railway is determined based on the geospatial vectors representing the associated left and right railway lines. The coordinates of the center axis of the railway can be calculated by taking the average between the left rail vector point and the right rail vector point.

[0109] Figure 5 A flow chart illustrating a modeling method implemented to model a geospatial vector representing a railway track using one or more analytical equations. The modeling method may include one or more iterations of a set of modeling steps and may include a modeling error threshold defined per unit length. The analytical equations that may be used to model the geometry of the centerline of the railway track may correspond to a set of basic geometric shapes, such as a straight line, a circular arc, and a clothoid. In practice, even if an exact correspondence is not achieved, the shape of the railway track may be approximated by one of the basic geometric shapes. A modeling error, which may be of the average orthogonal type, may be associated with the correspondence.

[0110] In step 501 of the modeling method, a geospatial vector representing a railway track is received for modeling using one of the identified basic geometric shapes. The basic geometric shape associated with the minimum modeling error for modeling the geospatial vector is retained. Step 501 may also be configured to provide parameters for each geospatial vector being modeled, the parameters comprising the retained geometric shape, parameters of an analytical equation for fitting the geometry of the geospatial vector closer to the retained geometric shape, and a modeling error associated with such fitting.

[0111] In step 502 of the modeling method, the modeling error provided in step 501 is compared to a modeling threshold. Such a modeling threshold can be determined by multiplying the modeling threshold per unit length by the length of the railway track represented by the geospatial vector. If the modeling threshold is greater than the modeling error, the iteration of the modeling method can be stopped.

[0112] In step 503, a modeled geospatial vector having a modeling error greater than a modeling threshold is divided into two segments that may have the same length. Each of the two segments may be modeled in a manner similar to the complete geospatial vector according to an iterative method, the iterative method including dividing each segment into two sub-segments having a modeling error greater than the modeling error threshold. The method for modeling a geospatial vector representing a railway provides a plurality of segments, each segment being modeled using an analytical equation that observes the modeling error threshold.

[0113] Figure 6 Shown are steps implemented to identify and associate modeled segment objects adjacent to a railway network that may interfere with navigation signals received by a mobile vehicle moving on a modeled railway track segment of the railway network, according to one embodiment of the present invention.

[0114] In step 601, a cloud of geolocated and classified measurement points associated with the railway network's immediate surroundings is filtered according to predefined filtering criteria. This filtering operation can, for example, eliminate measurement points corresponding to nearby objects with low altitudes relative to the altitude of the mobile vehicle receiving the navigation signal. The risk of low-altitude nearby objects interfering with the navigation signal can be negligible. Step 601 can also include applying a subsampling operation to measurement points associated with nearby objects with altitudes of the same order of magnitude as the altitude of the mobile vehicle. The subsampling factor can be selected to enable identification of nearby objects while reducing processing complexity.

[0115] In step 602, the filtered and subsampled measurement points are converted into neighboring objects. The neighboring objects are associated with object parameters, such as the object's geographic location, dimensions (e.g., length, width, and / or height), the distance between the object and the nearest railway track, and / or the object's relative orientation with respect to the track. The obtained neighboring objects can also be further synthesized into standard geometric shapes such as parallelepipeds, cylinders, and pyramids.

[0116] In step 603, the obtained neighboring objects are associated with railroad track segments of the railroad network. If the navigation signal received by a mobile vehicle traveling on the railroad track segment is likely to be interfered with by the object in question, then the object may be associated with the railroad track segment. Thus, a neighboring object may be associated with more than one railroad track segment. A railroad track segment may not be associated with any neighboring objects.

[0117] In step 704, for each association between a railway track segment and a neighboring object, the primary physical phenomena that may interfere with the navigation signal are determined. Such physical phenomena may include, for example, multipath propagation, propagation without a direct line of sight between the transmitter and receiver of the navigation signal, scattering, and the like. The physical phenomena that cause interference in the navigation signal may also be characterized by specifying geometric parameters associated with the association between the modeled segment and the neighboring object, as well as with the railway vehicle receiving the navigation signal. Such geometric parameters may include the separation distance between the two elements of the association, the size of the neighboring object, and the size of the railway vehicle.

[0118] Figure 7 An example of the association between a railway track segment and a nearby object is shown, where multipath propagation is the main physical phenomenon causing interference in the navigation signal. The object is synthesized as a parallelepiped geometry with a height "H" and a distance "D" from the central axis of the railway track sub-segment. For a moving vehicle at a height "M", the multipath propagation phenomenon can be quantified using a mathematical equation that relates the parameters "M" and "D" to the elevation angle "alpha" between the track height and the direction from which the navigation signal is received. Such a formula can be written as follows:

[0119] (HM) / D>tangent(alpha) (1)

[0120] Figure 8A second example of association is shown, in which a railway track segment is associated with two adjacent objects. In such an association, propagation without a direct line of sight between the transmitter and receiver of the navigation signal is the primary physical phenomenon causing interference with the navigation signal. The two objects implemented in this association are combined into a parallelepiped-type geometric shape. The two adjacent objects have heights H1 and H2 and are spaced apart by corresponding distances D1 and D2 from the center axis of the railway track segment. For a moving vehicle at a height M, the phenomenon of propagation without a direct line of sight between the transmitter and receiver of the navigation signal may occur when the elevation angle defined above satisfies the following relationship:

[0121] (H1-4M) / D1>tg(alpha)ET(H2-4M) / D2>tg(alpha)ET tg(alpha)>(H2-H1) / (D1+D2) (2)

[0122] Figure 9 is a diagram illustrating an autonomous positioning system 100 for a mobile vehicle traveling on a railway track of a railway network implementing a positioning method according to one embodiment of the present invention.

[0123] The autonomous positioning system 100 may include a positioning device 101 configured to:

[0124] - determining position parameters associated with the moving vehicle in a geographical reference system,

[0125] - determining one or more candidate railway track segments by querying a map database based on at least some of the location parameters, and

[0126] - Positioning the mobile vehicle based on the candidate railroad track segments returned by the map database 102.

[0127] The positioning system 100 may also comprise a unit for generating the map database 91 , for generating information or warnings about possible modifications (environmental changes) to the database 102 .

[0128] The unit 91 for generating a map database may include:

[0129] a receiving module 9100 configured to receive a cloud of measured points classified and geolocated in a coordinate system, said cloud of measured points being associated with the railway network and the adjacent environment,

[0130] a topology determination module 9102 configured to determine a topology of the railway network based on geolocated and classified measurement points associated with the railway network and a plurality of additional elements of the railway network,

[0131] a modelling module 9104 configured to model the geometry of a railway track of the railway network as a plurality of modelled railway track segments,

[0132] a module 9106 for determining neighboring objects, configured to determine one or more neighboring objects based on geolocated and classified measurement points associated with the neighboring environment, each of the neighboring objects being associated with one or more modeled railway track segments, and

[0133] - an identification module 9108 configured to identify, for each association between a neighboring object and one or more modeled railway track segments, information representative of interference affecting the navigation signal.

[0134] In one embodiment, the unit 91 for generating a map database may further include:

[0135] a comparator 9111 capable of determining changes in the risk level 704 associated with the railway network and the adjacent environment. Such a comparator 9111 may in particular signal elements of the geolocated point cloud associated with the adjacent environment as a number of adjacent objects that may interfere with the navigation signal;

[0136] - an alarm registration module 9112 capable of detecting changes in the risk level ( 704 ) in the database;

[0137] - an alarm analysis module 9113 capable of defining alarms by measuring redundancy 9112;

[0138] An association module 9114 capable of associating a confirmed alarm 9113 with at least one modeled segment 102 .

[0139] Thus, embodiments of the present invention make it possible to reconstruct the geometry and topology of a railway network based on geolocated and classified survey point clouds. By converting the 3D measurements of the classified point clouds and their trajectories (GPS / GNSS) into semi-automatic, large-scale 3D geospatial vectors, they can also define risks associated with the 3D characteristics of the railway network's immediate surroundings without requiring drawings or manual guidance using CAD software.

[0140] Those skilled in the art will appreciate that the systems or subsystems according to embodiments of the present invention can be implemented in a variety of ways, including hardware, software, or a combination of hardware and software, and in particular in the form of program code that can be distributed in a variety of forms in the form of a program product. Specifically, the program code can be distributed using a computer-readable medium, which can include computer-readable storage media and communication media. The methods described in this specification can be implemented in the form of computer program instructions that can be executed by one or more processors in an information technology computer device. These computer program instructions can also be stored in a computer-readable medium.

[0141] Furthermore, the present invention is not limited to the embodiments described above by way of non-limiting example, but encompasses all variant embodiments that may be conceived by a person skilled in the art.

Claims

1. A method for positioning a moving vehicle travelling on a railway track of a railway network in a railway reference system, the method comprising the following steps: - determining, based on a plurality of navigation signals received by the mobile vehicle, position parameters associated with the position of the mobile vehicle in a geographical reference system, - determining one or more railway track segments at which the mobile vehicle may be located by querying a map database based on at least one of the position parameters, - positioning the mobile vehicle based on the railway track segment provided by the map database and at least one of the position parameters, The map database comprises data representing a vector description of at least one railway track of the railway network and data representing a description of one or more objects of the environment adjacent to the railway track of the railway network that may interfere with navigation signals received by the mobile vehicle, the method comprising the steps of generating the map database, comprising the steps of: - receiving a cloud of measured points classified and geolocated in a coordinate system, said cloud of measured points being associated with said railway network and said adjacent environment, - determining the topology of the railway network based on geolocated and classified measurement points associated with the railway network and a plurality of additional elements of the railway network, - modelling the geometry of a railway track of said railway network as a plurality of modelled railway track segments, The step of modeling the geometry of the center axis of the railway track represented by the geospatial vector comprises one or more iterations of the following steps: i. determining a current analytical model of the geometry of the center axis of the railway track represented by a geospatial vector using at least one analytical equation, ii. measuring the standard deviation between the medial axis of the railway track and the associated curve returned by the analytical model, iii. If the standard deviation of the measurement is greater than a predefined error threshold, then splitting the geospatial vector into two elements, Steps i to iii are iterated as long as the standard deviation between the analytical model and the geospatial vector is greater than the predefined error threshold; - determining one or more neighboring objects from geolocated and classified measurement points associated with the neighborhood, each of the neighboring objects being associated with one or more modeled railway track segments, - identifying, for each association between a neighboring object and one or more modeled railway track segments, information representative of interference affecting the navigation signal.

2. The method according to claim 1, characterized in that The following steps are also included: The elements comprising the modelled railway track segment, the additional elements of the railway network and / or the neighbouring objects are saved in a file having a given representation format.

3. The method according to claim 1, characterized in that The additional elements of the rail network include intersections and stops.

4. The method according to any one of claims 1 to 3, characterized in that The geolocated and classified survey point cloud is received from at least one mobile terrain system associated with a hybrid geolocation system comprising a positioning device associated with an inertial unit.

5. The method according to any one of claims 1 to 3, characterized in that The position parameters include three-dimensional position, movement speed and / or orientation parameters.

6. The method according to any one of claims 1 to 3, characterized in that: The step of determining the topology of the railway network comprises the following steps: - converting elements of a geolocated point cloud associated with the railway network into a plurality of geospatial vectors, each of the geospatial vectors corresponding to a railway track, - Segmenting the rail network by defining, for each geospatial vector, a plurality of elements comprising intersections and / or stops.

7. The method according to claim 6, characterized in that Each railway track comprises two railway lines and is characterized in that the step of vectorizing a cloud of geolocated and classified survey points associated with the ground and said railway track into a plurality of geospatial vectors comprises the following steps: - identifying points in the geolocated point cloud associated with a selected railway line of said railway track, - subsample the identified points using a predefined subsampling factor, - grouping the subsampled points into subsets of points, two consecutive subsets of points sharing at least one measurement point, - using a moving average to calculate said average position in three dimensions for each subset of points, - creating a vector associated with the selected railway line, said vector comprising the calculated average position in three dimensions, - eliminating those components of the created vector that may be affected by noise from the automatic classification, which provides a geospatial vector, - determining one or more vectors associated with another railway line of the railway based on the geospatial vector determined for the selected railway track, and - Calculating a geospatial vector associated with the median axis of the railway track.

8. The method according to claim 7, characterized in that Each geospatial vector also includes bearing, slope, and tilt measurements extracted from each subset of identified points.

9. The method according to claim 6, characterized in that The subdividing step uses a graph comprising a set of edges representing railroad tracks associated with geospatial vectors and nodes connecting the edges, the edges representing intersections or stops on the railroad network.

10. The method according to any one of claims 1 to 3, characterized in that It also includes determining a network format corresponding to the railway network by associating each modeled railway track segment with one or more elements of the railway network.

11. The method according to any one of claims 1 to 3, characterized in that The steps of defining a neighboring object and associating the neighboring object with one or more modeled segments include the following steps: - subsampling the measurement points associated with said neighborhood by a predefined subsampling factor, - converting the subsampled measurement points into one or more neighboring objects, - associating the neighboring objects with the modeled railway track segment.

12. The method according to any one of claims 1 to 3, characterized in that Navigation signal propagation characteristics are assigned to each association between a neighboring object and one or more modeled segments according to a plurality of parameters including a height of the neighboring object and a distance of the neighboring object from a medial axis of the modeled segment.

13. A system for positioning a moving vehicle travelling on a railway track of a railway network in a railway reference system, the system comprising a positioning device (101) configured to: - determining, based on a plurality of navigation signals received by the mobile vehicle, position parameters associated with the position of the mobile vehicle in a geographical reference system, - determining one or more railway track segments at which the mobile vehicle may be located by querying a map database based on at least one of the position parameters, - positioning the mobile vehicle based on the railway track segment provided by the map database and at least one of the position parameters, Its characteristics are: The map database comprises data representing a vector description of at least one railway track of the railway network and data representing a description of one or more objects of the environment adjacent to the railway track of the railway network that may interfere with the navigation signal received by the mobile vehicle, the system further comprising means for generating the map database (91), the means for generating the map database comprising: a receiving module configured to receive a cloud of measured points classified and geolocated in a coordinate system, said cloud of measured points being associated with the railway network and the adjacent environment, a topology determination module configured to determine the topology of the railway network based on geolocated and classified measurement points associated with the railway network and a plurality of additional elements of the railway network, a modelling module configured to model the geometry of a railway track of the railway network as a plurality of modelled railway track segments, the modelling comprising: i. determining a current analytical model of the geometry of the center axis of the railway track represented by a geospatial vector using at least one analytical equation, ii. measuring the standard deviation between the medial axis of the railway track and the associated curve returned by the analytical model, iii. If the standard deviation of the measurement is greater than a predefined error threshold, then splitting the geospatial vector into two elements, As long as the standard deviation between the analytical model and the geospatial vector is greater than the predefined error threshold, Then the determining i, the measuring ii and the dividing iii are iterated; a module for determining neighboring objects, configured to determine one or more neighboring objects from geolocated and classified measurement points associated with the neighboring environment, each of the neighboring objects being associated with one or more modeled railway track segments, - an identification module configured to identify, for each association between a neighboring object and one or more modeled railway track segments, information representative of interference affecting the navigation signal.

Images