FR2965065A1 - Method for geographic locating person in e.g. shopping center, to realize personalized marketing communication, involves calculating route connecting source point to destination point, and determining probability of route - Google Patents

Abstract

The method involves covering an environment model (1), and determining a source point (14). A center of interest is determined, and a destination point (15) is determined. The route connecting the source point to the destination point is calculated. The probability of the route connecting the source point to the destination point is determined. A marketing target (9) is defined by a catchments zone. Independent claims are also included for the following: (1) a computer program product comprising instructions for implementing a method for geographic location of a person in a dark zone environment (2) a system for geographic location of a person in a dark zone environment, comprising a storage unit.

Description

The present invention relates to a method of geolocation of a person. It finds for particularly advantageous application the geolocation of a person in an environment, such as a shopping center, being in a shadow zone, or for which a location by a satellite type of locating means is not usable. Such location information can be used to achieve personalized marketing communication. In the field of personalized marketing, it is known for example WO 2008/081420 or WO 2008/056350 to transmit advertising information to a person via its mobile terminal according to its geographical location relative to a marketing target. In open areas, such as open-air streets, where satellite signals are available, this geolocation is typically performed by a satellite location system such as GPS (Global Positioning System) or equivalent. However some environments are covered and thus constitute gray areas where the GPS system can not be used. These environments in shadow areas include, among others, shopping malls, shopping malls, covered and may also present several floors. The present invention overcomes this problem by proposing a probabilistic geographical location method in shadow zone. The subject of the invention is a method of geographically locating a person in an environment in a shadow zone, comprising the following steps: recovery of a model of the environment, determination of a source point, determining at least one point of interest, - determining at least one destination point, 30 - calculating the possible routes connecting the source point to a destination point through all, none, or part of said at least one center of interest, - determination of a probability for each route. Thus, the invention provides a probabilistic location of the person. It is particularly advantageous for the location of people operating in an environment in a shadow zone, that is to say in which the radio signals are not transmitted properly. Advantageously, this location covers both two-dimensional and three-dimensional displacements. The invention is thus particularly well suited to the location in a building comprising several levels. It is specified that in the context of the present invention, geo-location means geographic location involving or not an interaction with the person to be located, for example by signal exchanges. According to another characteristic of the invention, the centers of interest are weighted / ordered before the calculation of the routes. According to another characteristic of the invention, the method further comprises a step of estimating the position as a function of time, by considering an average travel time, an average time of stopping at the centers of interest, and a displacement on the along the possible route or routes, to determine one or more positions, respectively with a probability equal to the probability of the route. According to another characteristic of the invention, the method further comprises, for at least one marketing target defined by a catchment area, a step of determining a probability of passage in said catchment area of the marketing target. According to another characteristic of the invention, a catchment area is defined circular by a point / center stitch and a radius. According to another characteristic of the invention, the model of the environment is three-dimensional, and rectilinear, comprising stages, and junctions between 25 stages of elevator, escalator or staircase type, each stage being meshed by a grid, each point / mesh of said grid that can be free, busy, input / output or junction. According to another characteristic of the invention, the source point is determined by a last known position of a mobile terminal associated with the localized person determined by: a last GPS position obtained before entering the shadow zone; a last GPS position obtained before entering the darkened zone at the entry / exit of the nearest environment, - the entry / exit of the most likely environment, 35 - an entry / exit of the environment confirmed by a dialogue with the person, - a destination obtained by a previous geographical location calculation, or - a position determined by another means, such as a localizing interaction. According to another characteristic of the invention, said at least one destination point is determined among the inputs / outputs of the environment. According to another characteristic of the invention, a center of interest is a location selected on the basis of a user profile or for which the person has made an interaction among: - purchase, - request for information, - subscription to a program loyalty card / registration of a loyalty card, - interest / response for an offer, - frequent / recent visit, - sponsorship. According to another characteristic of the invention, the calculation of the possible routes connecting the source point to a destination point uses a point-to-point or "maze router" type propagation algorithm used for the circuit routing.

According to another characteristic of the invention, the calculation of the possible routes connecting the source point to a destination point further comprises the definition of at least one route segment, a route segment connecting the source point to a center of the route. interest or point of destination, or a point of interest to another point of interest or point of destination. The method further comprises a step of calculating a cost of travel adding: - a cost of progression of at least one increment of progression, preferably +1 per mesh, - a cost of crossing junction: - d ' at least one elevator increment, preferably +1 regardless of the number of stages for an elevator, - at least one escalator increment, preferably +1 per floor + the horizontal distance traveled for an escalator, - at least one stair increment, preferably +2 per floor + the horizontal distance traveled for a staircase. According to another characteristic of the invention, the calculation of the possible routes connecting the source point to a destination point further comprises a confirmation step retaining for each segment of the route, only the minimum cost path or routes. According to another characteristic of the invention, the calculation of the possible routes connecting the source point to a destination point further comprises a recombining route reconstruction step said minimum cost path segments. According to another characteristic of the invention, the determination of a probability for each route comprises the determination of a segmental probability for each route segment, equal to the inverse of the number of said minimum cost paths for said segment of the route. route.

According to another characteristic of the invention, the determination of a probability for a route is carried out by multiplying the segmental probabilities of the various segments of the itinerary constituting said route and dividing by the number of routes. According to another characteristic of the invention, the determination of a probability of passage of the person located in a catchment area of a marketing target calculates the sum of the products of the probabilities of the routes passing in said catchment area by the number of said itineraries passing through said catchment area. According to another characteristic of the invention, the method also comprises a step of determining a date of entry of the person located in a catchment area of a marketing target for a route passing through the catchment area comprising the steps following: - determination of a date of entry into the environment, - addition of a travel time per mesh traversed along the route 25 before the catchment area, - addition of a visit time per center of interest crossed along the route before the catchment area. According to another characteristic of the invention, the method also comprises a step of transmitting an advertisement relating to a marketing target, to a portable terminal associated with the localized person, when the probability of passing through the catchment area of said marketing target is greater than a threshold. According to another characteristic of the invention, the method also comprises a step of receiving an advertising message relating to a marketing target, by a portable terminal associated with the localized person, when the probability of passing through the catchment area of said marketing target is greater than a threshold.

According to another characteristic of the invention, said message is transmitted substantially on the date of entry of the person located in said catchment area of the marketing target or at the entry date removed from a predefined period. Said predefined duration is advantageously determined so that the person can read the message before having exceeded the catchment area. The invention also relates to a computer program capable of implementing the steps of the method according to any one of the preceding embodiments.

The invention also relates to a geographic location system of a person in an environment in a shadow zone, comprising means for storing a model of the environment and a processor configured to: determine a source point, determine at least one center of interest, determining at least one destination point, calculating possible routes connecting the source point to a destination point through all, none, or part of said at least one center of interest, determining a probability for each route. Other features, details and advantages of the invention will emerge more clearly from the detailed description given below as an indication in relation to drawings in which: FIG. 1 presents, in perspective view, an exemplary model of 2 shows, in plane view, an example of an environment model comprising three stages, FIG. 3 presents a detailed view of the first stage, FIGS. 4-6 present on the model of FIG. example of propagation for route determination, - Figures 7-15 present on another model of environment a complete example. The context of the invention is a tracking system of an affiliated consumer, for proposing with a high degree of relevance at least one targeted marketing offer when said consumer passes or is close to a marketing target associated with said offer. Such a system typically comprises a server segment comprising computer processing means and communication means and at least one customer segment, associated with a person / consumer, for example an application running on a mobile terminal (mobile phone, smartphone, etc.) of said person. The server segment can communicate with said mobile terminal.

The server segment still has interfaces that allow it to obtain information about the environment or to interact, among other things, with marketing operators, for example to configure marketing campaigns. According to a known mode, said person is localized by geographical location of his mobile terminal, for example by means of a Global Positioning System (GPS) satellite location system receiver integrated into said mobile terminal. Said mobile terminal then transmits this geographical location to the server segment. The problem to be solved by the present invention is to continue to have a geographical location of a person when the person enters a shadow zone no longer allowing the reception of signals, such as GPS signals, allowing a location GPS. Such a case typically occurs when the person enters a covered environment, such as a shopping center.

Thus, when the person enters such an environment in a shadow zone, the mobile terminal can no longer act as a GPS receiver. At this instant when the GPS location is lost, the server segment switches to shadow localization mode and implements the method described later to continue to estimate the location of the person in said environment, in the absence of any GPS location. Conversely, when the person leaves the shadow zone and returns to the discovery zone and a GPS location can be obtained again, the system can automatically switch back to GPS location mode, as before. The method of geographically locating a person in an environment in a shadow zone, once initiated for example by the loss of a GPS signal, comprises the following steps: - recovery of a model 1 from the environment, - determination a source point 14, - determining at least one point of interest 8, 30 - determining at least one destination point 15, - calculating the possible routes 11 connecting the source point 14 to a destination point 15 passing by all, none, or part of said at least one center of interest 8, - determining a probability for each route 11. 35 Given the lower number of certain geographical location information, the selected method adopts a probabilistic approach where, among all the possible routes, the most probable one or more are determined based on a displacement economy. The method considers factual elements of topography of the environment, summarized in the form of a modeling, typically done in advance, possibly updateable by updating, for a given environment and accessible by the server. In order to limit the very large number of possible route choices 11 possible in advance, the method advantageously uses the interest center considerations 8 collected from the person in order to direct the route choices 11 towards the most probable ones. A center of interest 8 of a person is a location in the environment, where it is assumed that the person wants to go. In the case of a shopping center, it is typically a shop, constituting one of the objects of the visit to the environment by said person. A shop is typically located by its point of entry. The method calculates possible routes 11 connecting a source point 14 to a destination point 15 through all, none, or part of said at least one center of interest 8. Thus, with a source point S, two destination points D1, D2 and two centers of interest, 11, 12, at least one route 11 connecting in sequence: S-D1, S-D2, S-11-D1, S-11-D2, S-12-D1, S is calculated -12-D2, S-11-12-D1, S-I1-12-D2, S-12-I1-D1, S-12-I1-D2. We see that all combinations without 11 and 12, with 11 and 12, alone, in order or in disorder, are a priori possible. It is conceivable that the combinatory increases very rapidly with the number of centers of interest 8 and / or destinations 15. Advantageously, it is possible to reduce this combination by considering that the centers of interest 8 are weighted respectively, for example by a graduation of interest. Such a weighting can thus reduce the number of routes, for example by ordering in interest order said centers of interest 8, and considering only routes 11 respecting this order. In the previous example, an imposed order 11, 12 makes it possible to delete the last two routes. The method comprises an exploratory step determining possible routes 11, said exploration making it possible to obtain all the possible routes 11 or at least all those whose probability is a priori high, for each source pair 14, destination 15. The method further comprises a step of valuation, where is associated with each route 11 thus obtained, a probability that said route 11 is 35 borrowed by the person. From these routes 11, each provided with a probability, it is possible to use the method according to at least two embodiments.

According to a first embodiment, said routes 11 are used to locate geographically, probabilistically, the person, as a function of time. According to a second alternative embodiment or combined with the first embodiment, said routes 11 are used to determine whether or not said person crosses a catchment area 10 of one or more marketing targets 9, and according to which probability. According to the first embodiment, it is considered an average time of displacement tD by mail and / or junction, a mean time of stopping at points of interest t ,. These parameters can be average for any person, according to certain typological characteristics such as the sex or the age range of the person or personalized according to the past behaviors observed for the person or function of the environment considered. This information is typically included in a user profile stored in the mobile terminal fitted to the person. Then an estimate of the position can be calculated, considering that the person is moving, at an average speed corresponding to said mean travel time tD, stopping at the centers of interest said mean time t ,, along the or routes, depending on the weather. It can be assigned to said position a probability for example equal to that of the route assumed. According to the second embodiment, from the probabilized possible routes, it is possible to determine whether and with what probability the person has chances, during his visit to the environment to pass in a catchment area 10 defined for a given time. marketing target 9. Such a probability of passing through said catchment area 10 of the marketing target 9 is taken as equal to the sum of the probabilities of the routes that cross said catchment area 10. It is still possible to combine the two embodiments. to determine / estimate a date of entry into a catchment area 10.

According to one possible embodiment, a model of the environment 1 summarizes the dimensional and topographical elements of the environment in which the person is evolving. An example of such an environment model 1 is illustrated in perspective view in FIG. 1. This deliberately simplified model 1 represents an environment comprising three levels or stages, 2-4. Given that the environments that we want to model can be multi-stage, a three-dimensional model, 3D, or exactly 2.5D is used. The model comprises stages 2-4, connected by junctions 7. The junctions 7 are detailed in elevators A, escalator E or stair E. A junction 7 comprises a set of coordinates defined in a stage 2-4 and at least one set defined in another floor. Since said environments are most often orthogonally organized, a rectilinear modeling is adapted. Each floor plan is then squared by an orthonormal mesh, defining meshs, smaller units considered and therefore equated to points. The size of the mesh is a compromise between a fineness of definition and modeling of the environment and a complexity of calculations, for example of route and / or probability.

For example, a mesh of 1 mxl m is a compromise often satisfactory. Thus, in the example of FIG. 1, two junctions 7 of lift type A are shown, a first input / output point 5 in the environment used here as an input and a second input / output point 6 out of the environment used here as an output. The modeling 1 of the environment contains only immutable elements and can thus be determined beforehand, once and for all, with a possible update, and stored in a database accessible to the server. Each mesh is free-typed 12 when it allows passage and circulation, occupied 13 when it is not free: wall, private area, shop, etc., junction 7 when it corresponds to a junction, or entry / exit 5.6.

Any mesh / point is designated by its coordinates X, Y, for example of the center of the mesh, in the grid with reference to a common given origin. It can still be added a coordinate Z designating the floor. Figure 1 further illustrates the assignment in the same system mesh representation and floor, temporary elements specific to the process.

Thus the stages 2, 3, 4 respectively comprise two, two and a center of interest 8, I1-15 specific to a person and a given period of time. In the second embodiment for determining the passage in a catchment area 10, such a zone 10 may be placed in the representation system and designated by the same mesh / coordinate principle.

An example of route 11 is still shown in FIG. 1. Such a route 11 may be, when it is designated by the sequence of points / meshes traversed by said route 11 and thus by the set of coordinates of these points. sts. A catchment area 10 can be designated by the set of meshes, and therefore coordinates, it covers. It can then take on any form. The membership test is then performed by comparing each point of a route 11 with a point in the catchment area 10. Another designation, 2965065 Zo more condensed, considering a circular zone 10, can be defined by a point of center 9 or marketing target 9 designated by a point / mesh and its coordinates Xc, Yc, Zc and a radius Rc. The membership of a point in the catchment area 10 is then tested simply by observing that the distance from said coordinate point X, Y, to the center / marketing target 9 is less than said radius in the plane of the floor, ie (X-Xc) 2 + (Y-Yc) 2 less than Rc2. Preferably a mesh is considered to be included in the catchment zone 10 if at least a portion of this mesh is superimposed with said circular zone. Figures 2-6 illustrate in an example with three stages 2-4 10 shown in plan each by a grid 16x17, the various steps of the method. The coordinate system is illustrated in Figures 3 or 6 with an origin in the lower left corner and coordinates from 0 to 16 on the abscissa and 0 to 15 on the ordinate. On the model 1 of the environment are the fixed elements: 15 - free boxes represented by a white box, with an example 12, - occupied squares represented by a X cross, with an example 13, - input / output boxes 5, 6, - junction boxes 7, which can be type elevator A, staircase or escalator E. 20 Moreover are represented on the same support elements of the process: - points of interest 8, 11, 12, 13, 14, 15 - marketing target 9, M, - travel costs figured not numbers in the mesh, - routes 16-19. One of the parameters necessary for the method is a source point 14, origin of the route (s) 16-19. This source point 14 is determined by the location point, the last known of the person, assimilated to the last known location of his mobile terminal. This last location may be the last GPS position obtained before entering the shadow zone, at the moment when the GPS signal is lost and the process begins. Given the arrival in the environment, and the entry into the environment can be done only by one of the inputs / outputs 5,6, the modeling of this environment indicates the coordinates of the possible points / meshes . The last location is thus advantageously the last GPS position obtained before entering the darkened zone at the input / output 5, 6, of the nearest environment. In addition, said registration makes it possible to remove the possible doubt on the floor through which the entry was made. In the absence of a reliable or recent GPS position, the last known position can be determined, among the positions of the inputs / outputs 5, 6 of the environment, present in a limited number, choosing the most likely. Here the probability can be linked to a habitual choice of the person. It can be based on historical data or preference data contained in a user profile associated with the person. It is still possible to put the person in contribution by proposing, on his mobile terminal, to validate the input used. Such a dialogue can be proposed to the person among all the inputs / outputs of the environment 5, 6 or from a list already reduced by elimination or probability. Thus, the person confirms with the server of the system his effective entry point. According to another mode, the last known position can still be derived from a previous calculation of geographical location, by any method of calculation, including that according to the invention, the previous destination of the previous calculation becoming the new source 14.

The last known position of the person can still be determined by another means, such as a localizing interaction. An example of such a localizing interaction is an interaction performed in a shop adherent to the system. The interaction can be, for example, a purchase or validation of a loyalty subscription, typically to a terminal located at the entrance of the environment. Since there is an interaction between the mobile terminal of the person and a shop terminal or a fixed or known terminal, this interaction is localizing in that it determines in a certain way the location of the person, or at least its mobile terminal, at the moment of the interaction.

In the example, the source point 14 is located at the stage 2 in the position designated by the coordinates (0.14). Another of the parameters necessary for the method is at least one destination point 15, the culmination of route (s) 16-19. As much, there is only one source point 14, as many destination points 15 may exist. In addition, for completeness, all possible destination points should be considered. At the end of his visit to the environment, the person must necessarily come out through one of the entrances / exits 5, 6. All entrances / exits 5, 6 are candidates to be destination 15, including that used to enter the environment. The set of possible destinations 15 is therefore most often confused with all the inputs / outputs 5, 6 of the environment. Some of these entries / exits 5, 6 candidates may be withdrawn, if necessary, on the basis of other information (for example: entry / exit condemned) or on a probabilistic basis (entry / exit never or rarely borrowed before by the person ). Another parameter useful to the method is the knowledge of at least one center of interest 8, I. These points of interest 8 will make it possible to limit / orient the route search in that said routes 11 will preferably go through all, none, or part of these areas of interest 8. A center of interest 8 is a location, associated with a shop, a sign, a mark, etc., for which the person, identified by his mobile terminal, is presumed to mark an interest. It is then assumed that a visit of the person to the environment may have for purpose a passage to said location. The interest is detected by an action of the person registered by the system. It can be an interaction, via its portable terminal, or via another computer or network (Internet). It can still be an operation performed by the person and recorded in the system by another person, such as a sponsor, a marketing player, a sales manager affiliated with the system, etc. It may also be an act of purchase made at the center of interest or at another point of sale belonging to a network to which the area of interest belongs. The focus can also be defined based on a user profile. The latter can be determined by a selection of user preferences made by the user and / or analysis of the behavior of the latter via a questionnaire and / or historical behavior data. Thus an operation that signs an interest of the person for a brand, sign, franchise, shop can be associated with a location in the environment as soon as said brand, sign, etc., is present for example by a shop, a concession , a branch or a representative. Said location is known to the system and is then eligible to become a center of interest 8. As an example of such an operation: a purchase. The latter, made in a shop, is registered with the system by the seller and may subsequently be used to define a potential center of interest 8 on the said shop or on another branch of the same brand. Another example of such an operation may be a request for information. In this case the operation is performed by the person himself from his mobile terminal or from the internet.

The fact that the person subscribes to a loyalty program or has registered a loyalty card of a given brand, is an operation indicating an interest in the brand / shop.

Another operation of interest may be a response to a promotional offer, which in this case is indicative of a potential next visit. Another operation indicative of an interest of the person is the existence of frequent and / or recent visits. Thus the system can produce centers of interest 8 for a person by consulting the statistics of frequentation of the environment. Such statistics are obtained by the tracking and geographical location of the person identified with the system via his mobile terminal. Another operation indicative of interest is a sponsorship action. Thus a sponsorship operation of a brand / shop, which can advantageously be carried out via the system, is indicative of a possible interest for the brand / shop both from the sponsor and the godson. A sponsorship makes it possible to create a center of interest 8 towards the brand / shop for these two people.

A focus can also be chosen regardless of the user's desire or profile. This may be the case when the center of interest is a particular attraction attracting the attention of a large number of people. A center of interest can be a commercial, cultural, fun, carried out in a shopping center. Since these operations are generally of limited duration, it is then necessary to regularly update the data relating to the centers of interest. According to an important characteristic of the invention, the calculation of the possible routes 11 connecting the source point 14 to a destination point 15, via the centers of interest 8, uses a point-to-point or "maze" type routing algorithm. router ", used for circuit routing. Indeed, in the context of the development of the present invention, it turned out that the problem of determining routes could be compared to that of the determination of routing interconnection on a printed circuit. Of course, the fields of integrated circuit routing and the geographical location of a person are very far apart, and a person skilled in the field of geographical location would never have sought to establish a link between these two fields. Part of the invention lies in the non-obvious identification of this link. The problem is close in that it includes: - search for a path linking a source and a destination, - mainly orthogonal directions, - bidirectional paths, - multiple stages (layers) connected by junctions (vias), - rectilinear obstacles or similar, - non-zero travel cost. To develop the present invention, it has been envisaged to implement a point-to-point or "maze router" type of algorithm such as the so-called Lee algorithm, and described in: Lee, CY (1961), " An Algorithm for Path Connections an I Applications, IRE Transactions on Electronic Computers EC-10 (2): 346-365. This algorithm is however modified to integrate the concepts of "center of interest" or intermediate points of likely (but not systematic) passage, as well as "marketing target". Another important difference is the purpose. For routing it is searched for a single route, if possible optimal. The method according to the invention retains all optimal routes. Moreover, the whole probabilistic part is peculiar to the invention.

It should be noted that the marketing target is not necessarily a focus of interest. The fact that the marketing target is, for a particular person, a center of interest does not change the process for determining the probability of crossing the catchment area 10. It will now, with reference to FIGS. an example of determining possible routes. From the environment model 1 shown in FIG. 2, and supplemented by the centers of interest 8, five in number, and a marketing target 9, the method starts at the source point 14, with coordinates (0, 14). . From this source point 14, as illustrated in Figure 2 and in detail for the beginning in Figure 3, the route is built step by step. Each step adds a free mesh adjacent to the previous mesh or the crossing of a junction 7. The neighborhood considers only the meshes reached by a horizontal or vertical displacement, and not the diagonal displacements. Since the source 14, S this step-by-step propagation attempts to achieve a next objective, among a center of interest 8 or a destination 15. In order to produce only the most economical routes, a cost is added at each step of propagation. This cost is equal to at least one increment of progression, for example of +1, for a progression from one mesh to a neighboring mesh. For a crossing of junction 7, it is distinguished between 35 lift, escalator and staircase. The cost of using an elevator A is for example independent of the number of floors crossed and is taken at least equal to an elevator increment, for example +1. The cost of using an escalator E is taken at least equal to an escalator increment, for example +1 per floor exceeded in addition to the horizontal distance traveled, ie n + JdX2 + dY2, where n is the number the difference between the abscissa of the escalator entry point and the escalator exit point, and dY the difference between the escalator entry point and the exit point of the escalator. the escalator. The cost of using a staircase E is taken at least equal to an increment of stairs, for example +2 per floor exceeded in addition to the horizontal distance traveled, ie 2n +, / dX2 + dY2. With reference to FIG. 3 is thus illustrated a route determination from the source point 14. The propagation is represented by the annotation in the meshes, of numbers indicating the cumulative cost. It eliminates some choices by not reusing a mesh / junction already used with a lower cost. A first center of interest 8, 11 of coordinates (4, 9) can be reached with a cost of 9. A second center of interest 8, 12 of coordinates (2, 4) can also be achieved with a cost of 12 To reach the first point of interest 8, 11, as illustrated in detail in Figure 3, there appear four possible courses 16-19 with the same minimum cost of 9. There are therefore four courses 16-19 equivalents, of minimum cost, here 9, that is kept to connect the source point 14 to the first point of interest 8, 11. All the other paths to reach, from the source point 14, the first point of interest 8, 11 would necessarily be longer / expensive and are therefore eliminated. As illustrated more fully in FIG. 4, the propagation is used, through the meshes and the junctions 7, to reach the centers of interest 8, I and all the destinations 15.

A stage 2-4 is left via a junction 7. Thus the first stage 2 is left via the elevator 7 of coordinates (0, 1) or (0, 2) or again via the escalator / escalator 7 of coordinates (2 , 12). The propagation continues in the same way from the junction 7 corresponding to said stair / escalator, from coordinates (2, 10) to the second stage 3, until, for example, to reach the third center of interest 13 of coordinates (5). , 13) with a cumulative cost of 13 or the fourth center of interest 14 of coordinates (4, 6) with a cumulative cost of 14. Similarly the propagation is still used and explores the third stage 4, until reaching the destination 15, here unique, passing if necessary 35 by one or more centers of interest 8, I. The propagation is used to achieve segmental paths connecting the source point 14 to a destination point 15, the source point to a center of interest 8, a center of interest 8 to another center of interest 8, or a center of interest 8 to a destination 15. All combinations of such segments are contemplated, since a route 11 connecting in the order for example the source, the second 12 and the first center of interest 11 may be less expensive or equivalent to a route 11 linking the source, the first center of interest 11 and the second center of interest 12. II it is possible to define n-tuples (S, {I}, D), where S is the source point 14, D one of the destination points 15, {I} = {11, ... 1; ..., Ip } a set comprising p centers of interest among those identified for the person, p being at least equal to 0 and at most equal to the total number of centers of interest 8, I identified. Each such tuple is composed of a succession of pairs (S, D), (S, I;), (I;, Ii), (Ib) connectable by at least one route segment. For each such pair, the propagation as described above makes it possible to determine at least one route segment accompanied by a minimum cost of implementation. For such a pair (S, D), (S, I;), (I ;, Ii), (Ib D) given, any segment / path having a cost greater than this minimum cost can be rejected. On the other hand, for each pair, there can be several segments / paths totaling this minimum cost. All these segments / courses are then equivalent and are all preserved.

For each such segment i of a given pair, an elementary or segmental probability P is allocated; equal to the inverse of the number of equivalent segments of minimum cost for this given pair. Thus, in the example of the pair (S, 11) illustrated in FIG. 3, four equivalent cost segments are each allocated a segmental probability of 1/4.

Then for each tuple or route 11, a global probability is calculated. A route 11, defined by a tuple, comprises successive segments which, concatenated form said route 11 between the source 14, S and a destination 15, D. The probability of a route 11 resulting from the concatenation of segments is equal to the product of the segmental probabilities of these segments divided by the number N of possible routes / n-tuples. Having thus determined and probabilized all the possible routes 11, it is possible to determine, for a catchment area 10 associated with a marketing target 9, a probability of passage of the localized person. For this purpose, among the possible routes 11 previously determined and composed by concatenation of minimum cost route segments, are considered only the routes 11 which pass through the catchment area 10 and thus close to the marketing target 9. The probability of crossing / crossing of the catchment area is calculated as the sum of the products of the NF numbers of routes passing / crossing the catchment area by the probability of the said route. An itinerary 11, which is a succession of points / stitches, is considered to pass or cross a catchment zone 10, itself constituted of a set of points / stitches, if at least one point / stitch of the route 11 is confuses with a point / stitch of the catchment area 10, or is located at a distance less than the radius R of the marketing target point 9. All possible routes are provided with a probability.

It is still possible, for each route 11 passing through the catchment area 10, to statistically specify on which date said person enters said catchment area 10. For this the method performs the following steps. An initial reference date is determined, such as a date of entry into the environment or a date when a certain geographical location is known, for example by a localizing interaction. If it is the date of entry into the environment, it is known by the system at the time of loss of GPS signals and can be stored. Then along the said route 11 is considered, from the starting point corresponding to the date of reference to and before entering the catchment area 10, the number of meshes crossed, as well as the centers of interest 8 crossed. . It is added to the date of reference to a travel time tD per mesh crossing and a visit time t, per point of interest crossed. As seen previously, the travel time tD can be a function of many factors related to the person (age, sex, etc.) or the environment. Similarly, the visit time t, by interest 8 may be a function of factors related to the person and / or individuated by interest 8. The method can be used to transmit a commercial message relating to a target marketing 9. This message is transmitted to a portable terminal associated with the localized person. Its transmission can be selectively triggered by the system when the probability of passing through the catchment area 10 of said marketing target 9 is greater than a threshold. In order to further increase the relevance of said advertising message, the transmission of the latter may advantageously be triggered substantially on the date of entry tF of the person located in the catchment area 10 of the marketing target 9 or shortly before or after . The offset for example before the date of entry is advantageously calculated to allow the person to read the message before it leaves the catchment area 10. Referring to Figures 7 to 15 will now be described another example complete implementation of the method. The environment, as modeled, is deliberately simplified to maintain a reasonable development size. Figure 7 shows the modeling of the environment. The latter comprises two floors, a ground floor, DRC, z = 0, to the left of the figures and a first floor, Z = 1, arranged to the right of the figures. Each stage is meshed by an 8x8 grid and marked in coordinates from 0 to 7. Each mesh can be empty for a free mesh, denoted by a B crossed with a cross for a locked or inaccessible mesh, denoted by an S for a source mesh (a single mesh source of coordinates (x, y, z) = (0, 6, 0)), denoted by a D for a target mesh (a single target mesh (7, 0, 0), denoted an I followed by a number for a mesh of interest (two points of interest 11 (4, 1, 0) and 12 (5, 5, 1)), denoted by an E for a stitch / escalator (a staircase joining (2, 4, 0) and (2, 3, 1), denoted by an A for an elevator mesh (two elevators respectively joining (7, 3, 0) and (7, 3, 1 ), and (7, 4, 0) and (7, 4, 1)), denoted by an M for a marketing target cell (a marketing target (2, 6, 1)) whose catchment area is represented by Hatched stitches.

The enumeration of all possible routes leads to the identification of N = 5 routes: S-D, S-11-12-D, S-12-11-D, S-11-D, S-I2-D. Referring to Figure 8 is illustrated the propagation for determining the minimum cost paths for the segment S-I1. For this segment 9 possibilities are found with a cost of 9. The segmental probability of S-I1 is thus 1/9. Referring to Figure 9 is illustrated the propagation for determining the minimum cost paths for the segment I1-D. For this segment 3 possibilities are found with a cost of 4. The segmental probability of I1-D is thus 1/3.

With reference to FIG. 10 is illustrated the concatenation of the two preceding segments S-I1 and I1-D to construct an S-11-D route. The probability of the S-I1-D route is calculated by the product of segmental probabilities divided by the number of possible routes: P (S-I1-D) = P (S-11) * P (11-D) / N = 1/91/31/5 = 1/135.

Referring to Figure 11 is illustrated the propagation for determining the minimum cost paths for the segment S-I2. For this segment 9 possibilities are found with a cost of 11. The segmental probability of S-I2 is thus 1/9. Referring to Figure 12 is illustrated the propagation for determining the minimum cost paths for the segment I2-D. For this segment 4 possibilities are found with a cost of 12 (both lifts are possible and equivalent). The segmental probability of I2-D is thus 1/4. Referring to Figure 13 is illustrated the propagation for determining the minimum cost paths for segment 11-12. For this segment 2 possibilities are found with a cost of 10 (both lifts are possible and equivalent). The segmental probability of 11-12 is thus 1/2.

Referring to Figure 14 is illustrated the propagation for determining the minimum cost paths for segment 12-11. For this segment 2 possibilities are found with a cost of 10 (both lifts are possible and equivalent). The segmental probability of 12-11 is thus 1/2. With reference to FIG. 15 is illustrated the propagation for determining the minimum cost paths for the S-D segment. For this segment 2 possibilities are found with a cost of 13. The segmental probability of S-D is thus 1/2. These different segments are then concatenated to build the different routes. The probability of each route is calculated by the product of segmental probabilities divided by the number of possible routes: P (SD) = 1/10, P (S-I1-D) = P (S-I1) * P (I1 -D) / N = 1/91/31/5 = 1/135, P (S-I 2 -D) = P (S-I 2) * P (I 2 -D) / N = 1/91/41/5 = 1/180, P (S-I1-12-D) = P (S-I1) * P (I1-I2) * P (I2-D) / N = 1/91/21/41/5 = 1 / 360, P (S-12-I1-D) = P (S-I2) * P (I2-I1) * P (I1-D) / N = 1/91/21/31/5 = 1/270 . In order to determine a probability of crossing / crossing in the catchment area of the marketing target, the different routes are listed in order to determine their number and among this number, the number NF of those who cross said catchment area. From this census it appears that S-D, SI1-D and S-11-12-D have no possibility of passage through the catchment area. On the other hand, out of the 36 possible routes for S-I2-D, 12 cross said zone. Moreover, out of 54 possible routes for S-12-11-D, 18 cross said area.

A probability of passage in the catchment area can thus be determined P = P (route) * NF (route) = 12/180 + 18/270 = 1/8. If now, we want to add a temporal aspect to these two routes to cross the catchment area, it is appropriate to give values to the times. With an average travel time tD of l mn and a time of visit of a center of interest t, of 10 mn, it is obtained a crossing of the catchment area, for a time of entry in the environment to : - for the S-I2-D route: tf = to + 9 * tD = to + 9mn, - for the S-12-11-D route: tf = to + 9 * tD = to + 9mn. Here the two routes are equivalent. Such an estimation of the crossing time makes it possible to determine an ideal moment for sending at least one advertising message relating to the marketing target.

Thus, the invention makes it possible to locate a person geographically in any environment, and more advantageously in an environment in a shadow zone. It is particularly well adapted to environments in which the person evolves in three dimensions, in horizontal planes, but also vertically.

In particular, the increment numbers to pass from one mesh to another or by an elevator, an escalator, a staircase, a center of interest can be modified without departing from the scope of the invention. The invention is not limited to the previously described embodiments but extends to any embodiment within its spirit.

Claims (12)

REVENDICATIONS1. A method of geographically locating a person in an environment in a shadow zone, characterized in that it comprises the following steps - recovering a model (1) from the environment, - determining a source point ( 14), - determining at least one point of interest (8), 10 - determining at least one destination point (15), - calculating the possible routes (11) connecting the source point (14) to a point destination (15) passing through all, none, or part of said at least one center of interest (8), - determining a probability for each route (11). 15 2. Method according to claim 1, wherein the centers of interest (8) are weighted / ordered before the calculation of the routes. 3. A method according to claim 1 or 2, further comprising a step of estimating the position as a function of time, considering a mean travel time (tD), an average time of stopping at the points of interest (t ,), and a displacement along the possible route (s) (11), to determine one or more positions, respectively with a probability equal to the probability of the route (11). 25 4. Method according to any one of claims 1 to 3, further comprising, for at least one marketing target (9) defined by a catchment area (10), a step of determining a probability of passage in said area of catchment (10) of the marketing target (9). 5. Method according to claim 4, wherein a catchment area (10) is defined circular by a point / center stitch (9) and a radius. 6. Method according to any one of claims 1 to 5, wherein the model 35 (1) of the environment is three-dimensional, and rectilinear, comprising stages (2, 3, 4), and junctions (7) between floors elevator type (A), escalator or stair (E), each floor being meshed by a grid, each dot / mesh 30dudit grid can be free (12), occupied (13), input / output (5, 6) or junction (7). 7. Method according to any one of claims 1 to 6, wherein the source point (14) is determined by a last known position of a mobile terminal associated with the localized person determined by: - a last GPS position obtained before the entry into a shadow zone, - a last GPS position obtained before entering the darkened zone at the entry / exit (5, 6) of the nearest environment, 10 - the entry / exit (5, 6) of the most probable environment, - an entry / exit (5, 6) of the environment confirmed by a dialogue with the person, - a destination (15) obtained by a previous geographical location calculation, or a position determined by other means, such as a localizing interaction. The method of any one of claims 1 to 7, wherein said at least one destination point (15) is determined from the inputs / outputs (5, 6) of the environment. 9. The method of claim 8, wherein a center of interest (8) is a location selected on the basis of a user profile or a location for which the person has made an interaction among: 25 - purchase, - request for information , - subscription to a loyalty program / registration of a loyalty card, - interest / response for an offer, 30 - frequent / recent visit, - sponsorship. The method of any one of claims 1 to 9, wherein calculating the possible routes (11) connecting the source point (14) to a destination point (15) uses a point-to-point routing type propagation algorithm. -point used for printed circuit routing. A method according to any one of claims 1 to 10, wherein the calculation of the possible routes (11) connecting the source point (14) to a destination point (15) further comprises the definition of at least one segment per route , each route segment connecting the source point (14) to a point of interest (8) or a destination point (15), or a point of interest (8) to another point of interest (8) or at a destination point (15), the method comprising another step of calculating a travel cost. 12. The method of claim 11, wherein the calculation of a journey cost 10 adds: a progression cost of at least one increment of progression, preferably +1 per mesh, a junction crossing cost: at least one elevator increment, preferably +1, regardless of the number of floors for an elevator, at least one escalator increment, preferably +1 per floor + the distance horizontally traveled for an escalator, - at least one step increment, preferably +2 per floor + the horizontal distance traveled for a staircase. 18. The method of claim 12, wherein calculating the possible routes (11) connecting the source point (14) to a destination point (15) further comprises a confirmation step retaining for each route segment, only the one or more minimum cost path. The method of claim 13, wherein calculating the possible routes (11) connecting the source point (14) to a destination point (15) further comprises a route reconstruction step (11) recombining said cost path segments. minimum. The method of claim 13 or 14, wherein determining a probability for each route (11) comprises determining a segmental probability for each route segment equal to the inverse of the number of minimum cost runs. for said route segment. 21. A method according to any one of claims 13 to 15, wherein the determination of a probability for a route (11) is performed by multiplying the segmental probabilities of the different route segments constituting said route (11) and dividing by the number of possible combination of segments. A method according to any one of claims 1 to 16, wherein determining a probability of passage of the person located in a catchment area (10) of a marketing target (9) comprises a step of determining the routes passing through said catchment area (10) and a step of calculating the sum of the products of the probabilities of the routes (11) passing through said catchment area (10) by the number of routes passing through said catchment area (10). ). 18. A method according to any one of claims 1 to 17, further comprising a step of determining an entry date (tf) of the person located in a catchment area (10) of a marketing target (9). for a route (11) passing through the catchment area (10) comprising the following steps: - determining an entry date (to) into the environment, - adding a travel time (tD) per mesh crossing along the route (11) before the catchment area (10), - adding a visit time (t,) per point of interest crossed along the route (11) before the trade (10). 21. A method according to any one of claims 4 to 18, further comprising a step of transmitting an advertisement relating to a marketing target (9) to a portable terminal associated with the localized person, when the probability of passing through the catchment area (10) of said marketing target (9) is greater than a threshold. 22. The method of claim 19, wherein said message is transmitted substantially on the date of entry (tF) of the person located in said catchment area (10) of the marketing target (9), or on the date of entry. (tF) removed from a predefined time interval. 23. Computer program product capable of implementing the steps of the method according to any one of the preceding claims. 24. A system for geographically locating a person in an environment in a shadow zone, comprising means for storing a model of the environment and a processor configured to: determine a source point, determine at least one point of interest , determining at least one destination point, calculating possible routes connecting the source point to a destination point through all, none, or part of said at least one center of interest, determining a probability for each route.