Route recommendation based on usage level for soft modes of transport

Abstract

The invention relates to a route recommendation method for a user of soft modes of transport, in which method as many segments as possible making up the route are selected where the estimated usage level for the segment in question exceeds a threshold. The invention also relates to a terminal or to a server implementing the route recommendation method.

Description

Recommended route using active transportation based on usage levels

[0001] The present invention relates to the field of route research and recommendation for users of active transportation. It relates in particular to a route recommendation method and an associated terminal.

[0002] With the development of smart mobile devices such as smartphones and tablets, it is common for users of such devices to use applications that determine a route between a starting point and a destination and guide them along the chosen route. Numerous route planning systems are available, particularly for drivers seeking routes with the lightest possible traffic.

[0003] For pedestrians and cyclists, who can be considered active transportation users, the search criteria may differ. In urban areas, for example, active transportation users may want a route where they feel safe. To achieve this, they may prefer busy streets with high traffic rather than quieter ones. However, current route planning systems do not offer this type of recommendation. Some applications allow users to request routes along main streets rather than secondary streets, which might suggest that these streets are busier than others, but this is not always the case. Indeed, some secondary streets may be busier than main streets, and some main streets may not provide the sense of security the user seeks.

[0004] Therefore, there is a need to improve state-of-the-art methods in this type of situation.

[0005] The invention improves upon the existing state of the art.

[0006] To this end, the invention aims at a route recommendation method for a user in soft mobility in which the selection of a maximum of segments composing the route is carried out if an estimated frequency rate for the segment considered exceeds a threshold.

[0007] Thus the recommended route includes segments where the traffic volume is high enough that the user of soft mobility is not isolated during their journey and therefore feels safe.

[0008] Advantageously, the estimated attendance rate is an attendance rate of people using soft mobility.

[0009] Thus, the feeling of security is reinforced by the presence of people who move in the same way as the user and who can therefore be more easily accessible.

[0010] In one particular embodiment, the estimated usage rate for a route segment is calculated from a database containing network signaling data from mobile terminals present in that segment.

[0011] Network data, such as signaling data emitted by mobile terminals in the segment, makes it possible to estimate the presence of people carrying these terminals and their speed of movement, and thus to determine a traffic rate on this segment.

[0012] In a particular embodiment, a historical database is constructed from network signaling data collected on predefined portions of a geographical area, this historical data including data on the frequency of people using soft mobility per predefined portion, at given times.

[0013] This database therefore contains data that can be collected over a long period of time, for example over several months or several years in order to accurately and reliably estimate a rate of attendance at a given time.

[0014] In one embodiment, a predictive model of the usage rate of people using active mobility for a given portion is obtained from historical data and contextual data.

[0015] Thus, taking into account contextual data, such as weather data or data from particular events, makes it possible to define a prediction model adapted to different situations and thus to estimate the traffic rates of requested routes even more reliably.

[0016] In a suitable embodiment, the prediction model is obtained by training a neural network, an inference of the neural network giving an estimate of the attendance rate for a given portion from a route request and associated context data.

[0017] An architecture based on a neural network makes it possible, in effect, to determine a relevant prediction model and to use it for future queries.

[0018] In an advantageous embodiment, an estimate of the usage rate for a route segment is obtained from the estimate of the usage rate of the geographical portion including that segment and mapping data of the geographical portion.

[0019] Thus, the estimated attendance rate is obtained with greater geographical precision so as to be able to select the segments for which the attendance rate has a sufficient value.

[0020] In an advantageous embodiment, attendance rates are estimated based on a time range.

[0021] Therefore, depending on the time of day, traffic levels may vary, and recommended routes may differ accordingly. This allows users to choose a route and time of day when the estimated traffic level is higher.

[0022] In a particular embodiment, segments of the route are proposed if their estimated lighting rate also exceeds a brightness threshold and / or if their estimated crime rate is also below a crime threshold.

[0023] Therefore, other criteria that enhance the user's sense of security along their route can be taken into account. For example, well-lit areas improve visibility and reinforce this feeling of security for users of active transportation. Similarly, information on crime statistics for a given geographic area can strengthen the feeling of security in that area if these statistics are low.

[0024] In one embodiment, the estimated lighting rate for a route segment is obtained from data captured by environmental cameras.

[0025] Data captured by environmental cameras, particularly in the area under consideration, allows for the measurement of brightness levels through image analysis. These images also enable the estimation of lighting periods in these areas, leading to a more accurate assessment of brightness levels.

[0026] Advantageously, a step is performed to display the segments and their attendance rate.

[0027] Thus, the user can choose the route that best suits them according to the frequency of the segments they want.

[0028] The user can also select routes for which there are segments that do not have a sufficient rate of traffic but which may still be of interest to the user if, for example, the rate is not far from the threshold or if the segment in question is very short and / or if the rate of light is known and reassuring.

[0029] In a particular embodiment, where a segment of a recommended route includes at least one segment for which the estimated attendance rate is below a threshold, the following steps are implemented: - search for at least one person located in a segment close to the segment for which the estimated attendance rate is below the threshold and for which their contact details are recorded in a database of an assistance service; - sending a notification of a request for assistance to at least one person located; - in the case of a positive response to the request for assistance, sending an appointment notification to both the person located and the user.

[0030] Thus, the assistance service compensates for the existence of segments with insufficient usage in the recommended route. User safety is therefore enhanced by this assistance service, particularly for these segments.

[0031] The invention relates to a communication terminal comprising a processing circuit for implementing the recommendation process as described.

[0032] It also relates to a server comprising a processing circuit for implementing the recommendation process as described.

[0033] The terminal and the server have the same advantages as the process they implement.

[0034] The invention relates to a computer program comprising instructions for implementing the recommendation process as described above, when executed by a processor.

[0035] Finally, the invention relates to a processor-readable recording medium on which is recorded a computer program containing instructions for executing the recommendation process described above.

[0036] Other features and advantages of the invention will become clearer upon reading the following description of particular embodiments, given by way of simple illustrative and non-limiting examples, and the accompanying drawings, among which:

[0037] illustrates, schematically, a terminal and a server according to an example of an embodiment of the invention;

[0038] illustrates in the form of an organizational chart the main steps of the route recommendation process according to one embodiment of the invention;

[0039] illustrates in the form of a flowchart the steps of a particular method of implementing the recommendation process according to the invention;

[0040] illustrates in the form of a flowchart a method for constructing a historical database from collected network signaling data;

[0041] illustrates in the form of a flowchart the steps to obtain an estimated attendance rate for a given segment and a given time;

[0042] and illustrate examples of representation of a recommended route and the segments constituting it associated with an estimated traffic rate, for two different time periods;

[0043] illustrates a representation of a recommended route and its constituent segments, along with their respective usage rates; and

[0044] illustrates in the form of an organizational chart the steps of a particular implementation method when at least one of the segments of a recommended route is associated with an estimated insufficient attendance rate.

[0045] Lare represents on the one hand, a terminal T, for example a mobile terminal such as a mobile phone, for example of the "smartphone" type, a digital tablet, or a personal computer, on the other hand a server S.

[0046] Terminal T includes the following functional modules:

[0047] - a communication module (COM-T), input / output intended to communicate with a communication network R;

[0048] - a processing module (µP-T) comprising a processor and an operating system to control the interactions between the different modules of the terminal;

[0049] - a memory module (MEM-T) in which is stored an application or software module (APP) comprising program instructions for the implementation of the steps of the recommendation process according to an embodiment of the invention and as described for example with reference to the;

[0050] - an interface module (INT.) allowing the user carrying the terminal to interact;

[0051] - a location module (LOC.) to implement a terminal location function. This module is optional, as terminal location can be performed using network signaling data.

[0052] Terminal T can implement the recommendation process as described in a first embodiment. In a second embodiment, terminal T sends a route recommendation request via a software module or a mobile application to a server S of the communication network R. This request is received by this server S, which is capable of collecting data or obtaining data from a database DB1, DB2, or DB3. This data, which will be described later, can be network signaling data from mobile terminals in a defined area, video data captured in a specific area, or GPS (Global Positioning System) positioning data.

[0053] Mapping data for locations where terminal T is located or locations specified in the route recommendation request are also available in a database.

[0054] The S server and database can be integrated, for example, into a service platform of an operator or the provider of the mobile application APP.

[0055] This server S is illustrated and includes the following functional modules:

[0056] - an input / output communication module (COM-S) intended to communicate with the communication network R;

[0057] - a processing module (µP-S) including a processor to control the interactions between the different modules of the server;

[0058] - a memory module (MEM-S) in which is stored a computer program (Pg) comprising instructions for the execution of the steps of the route recommendation process according to an embodiment of the invention and as described for example with reference to the.

[0059] The diagram represents the steps of the route recommendation process when implemented, according to a first embodiment by a terminal, for example terminal T described in the diagram, or when implemented, according to a second embodiment, by a server S of the communication network, for example as described in the diagram.

[0060] In step E20, in the first embodiment, the terminal receives a route recommendation request via its user interface between a starting point A and a destination B. These starting and ending points are specified, for example, by an address, geographic coordinates, a point of interest, or any other location identifier. The starting point can also be deduced from the location of the terminal initiating the request. In the second embodiment, the server S receives this request, for example, via a communication from a user's terminal T.

[0061] This request can also include a desired time range for this route.

[0062] Other information may also be contained in this request, for example a desired attendance threshold, a maximum number of route segments with an attendance rate below the desired threshold, or a percentage of distance to be traveled in addition to the shortest route.

[0063] A segmentation is then performed in E21 to determine the possible route segments between point A and point B defined in the query. These segments are those that can be traveled by a user using active transportation, such as a pedestrian, cyclist, scooter rider, or other low-speed individual vehicle. These route segments are, for example, sections of streets or paths. These route segments or sections can be of equal length, for example, on the order of a few hundred meters.

[0064] Determining these segments can be done using existing route planning applications. This route calculation can, for example, use a Dijkstra-type algorithm to find the shortest path between two points, as well as digital maps to visualize roads and points of interest. For this purpose, the map of the geographical area in question, for example, an urban area, is represented as a graph where intersections are nodes and street segments are edges. Several sequences of segments can thus be determined for a journey between points A and B. The sequences may have different distances, but prioritizing the search for segments that provide the shortest possible route is essential.

[0065] In step E22, a usage rate of active transportation users is estimated for each of the segments determined in the previous step. This usage rate is determined from data collected and recorded in a database. In a first implementation example, the data consists of network signaling data collected and recorded (for example, in the BD1 database illustrated in Figure 1) anonymously over several time periods and for a given area. The advantage of this type of data is that it provides a usage estimate very close to reality, due to the large number of people carrying mobile devices who thus exchange signaling information.

[0066] In a second implementation example, this data consists of GPS data captured and recorded for several time periods and for a given area. This data can be stored in a database (for example, in the BD2 database shown in Figure 1). This data may not be sufficient since it is subject to the consent of the individuals carrying the devices. The estimated visitor numbers may therefore not be close enough to reality.

[0067] In a third implementation example, this data consists of video data captured by cameras installed in specific areas, such as surveillance cameras. This data can be stored in a database (for example, in the BD3 database shown in Figure 1). The availability of this data depends on the presence of cameras in the environment, which can vary from one urban area to another.

[0068] A method for estimating the attendance rate per segment is detailed with reference to figures 4a and 4b described later.

[0069] Using digital maps that list the various roads or paths a user of active transportation can take, it is possible to compile the data recorded for a given segment. This data is then analyzed according to a desired time period to estimate the number of people using active transportation within that segment.

[0070] Only terminals with a slow speed, for example less than 15 km / h, are considered. Thus, for a given time period, a day of the week, or even depending on the season or a calendar of events in that area, an estimated number of people present and using active transportation is calculated for the segment in question.

[0071] Contextual data, described later, further enriches the estimate of the attendance rate obtained.

[0072] An attendance rate for this segment can then be defined. This can be simply the estimated number of people in this segment, or the number of people normalized to the segment size if the segment sizes are different.

[0073] Thus, in step E23, the estimated attendance rate for the determined segments is compared to a threshold S which can be defined by the user himself, for example in his query or which can be defined by default.

[0074] This rate could, for example, be 10 people, in one implementation example. Of course, other threshold values ​​can be defined. Segments with a occupancy rate exceeding this threshold are retained. The routes with the most segments exceeding this threshold are then determined in step E24. The Dijkstra algorithm is used here, for example, to find the shortest path that maximizes the sum of the occupancy rates on the segments that make up the route. Indeed, there may be several routes with segments that have an occupancy rate exceeding the threshold S. The occupancy rate value can be higher or lower than the threshold, or it can be below the threshold but for a short segment.

[0075] The resulting routes may lengthen the travel distance compared to the shortest route between point A and point B. This is because the shortest route may include segments with low traffic levels that do not exceed the required threshold. Therefore, to maximize the number of segments with an estimated traffic level above the threshold S, the resulting route is very often longer. A compromise between travel length and sufficient traffic levels must then be found. A threshold between the shortest path and the recommended route is then taken into account.

[0076] For example, a maximum distance difference between the shortest route and the recommended route can be requested and given in the initial request.

[0077] These routes and an estimate of the traffic volume associated with each segment of the route are presented to the user on their mobile device interface for a given time period. They can then decide which route to take based on their preferences, the estimated traffic volumes, or the distances to be covered.

[0078] The same route can also be offered but for different time slots, with different estimated traffic volumes associated with each segment. The user can then decide which time slot is best for traveling the route.

[0079] La illustrates a particular implementation of the recommendation process in which other parameters are taken into account for selecting the segments composing the route. In one example implementation, step E23, as explained with reference to the, selects the segments (Si ) for which the estimated attendance rates are above a threshold S (F Si >S). A maximum number of segments meeting this criterion are thus selected to constitute the route to be proposed to the user. Multiple routes may be recommended. There may be some route recommendations in which one or more of the constituent segments do not have a sufficient estimated usage rate. In this case, it is possible to refer to another criterion, in step E26, for example the light level (L). Si) associated with the segment in question. This brightness level can be obtained via video data from cameras installed in an area encompassing the segment. This video data allows for the measurement of the brightness level. The user can specify a desired brightness level and therefore a threshold to be exceeded, or this threshold can be set as a default. Thus, if the brightness level for the segment in question is below this threshold S l , then the segment in question is not selected. Conversely, if the brightness level is greater than the threshold S l , then this segment can be selected, even if it has an estimated attendance rate lower than the desired minimum attendance rate.

[0080] This brightness criterion can also be an additional factor in segment selection. In this case, both the estimated traffic volume and the brightness level are taken into account when selecting the segment to be included in the recommended route.

[0081] Similarly, another selection criterion can be considered. In E27, crime or delinquency rate data by area can be obtained via a database that collects this type of information. In this case, for a given segment, if that segment is in an area with a crime rate (TC) Si ) is strong, for example greater than a threshold S CTherefore, this segment is not selected for the recommended route. This selection criterion, like the crime criterion, can be a condition in addition to the ridership rate, to be met, or a criterion to be evaluated only if the ridership rate for this segment is insufficient.

[0082] Figures 4a and 4b illustrate, in the form of a flowchart, an example of the implementation of step E22. In a first embodiment, the data used to estimate a usage rate for a segment are network signaling data from the terminals present in the segment under consideration.

[0083] A first phase, illustrated in the diagram, allows for the creation of a historical database. A geographical area is considered; in this example, this could be an urban area or a district within that urban area. In step E41, signaling data from terminals on the mobile network is collected over this geographical area, with terminals regularly connecting to the telecommunications antennas covering this area (Collect.Sign.(t0, …t1, …t1). p(Ants). This signaling data allows for the identification of an approximate position of terminals within this geographic area, anonymously. The data is collected for several time periods and different days of the week. Thus, by analyzing successive connections to the antennas, the time elapsed between these connections, and the distance between the antennas, it is possible to determine the speed at which terminals move within the geographic area.

[0084] A method such as that described for example in published patent application EP4325903 can for example be implemented for this determination.

[0085] On the other hand, the geographical area is divided into E42, into portions of predefined size (P (j…P) ), for example in a square with sides of 150m, thus creating an analysis grid.

[0086] At step E43, each segment thus obtained is analyzed by cross-referencing the position and speed of movement data of the terminals to estimate the number of people using active mobility (V <Vm) dans une telle portion géographique et à un instant donné. Un déplacement en mobilité douce est par exemple un déplacement à moins de 15 km / h (ici Vm=15m / h).

[0087] This data is continuously updated over a period of time that can range from a few months to even a few years, allowing for dynamic and precise collection of visitor numbers in each section at a specific date and time. This makes it possible to build a complete and reliable historical database (BD-HIST).

[0088] From this historical and contextual database (BD-CONTEXT), a model for estimating the attendance rate per geographical area is determined, as illustrated in the.

[0089] Context data can be of different types. It can be data such as:

[0090] - Weather: Indeed, rainy or freezing weather can reduce the number of people present in a given area compared to sunny weather with pleasant temperatures. This weather data is obtained via a weather forecasting application server, or via a communication network database;

[0091] - seasonality: seasons or months or days of the week;

[0092] - hour and minutes of the day;

[0093] - distance from places of interest: proximity to attractions, shops or public events;

[0094] - other contextual factors such as special events like concerts, demonstrations, public holidays, etc…

[0095] Predicted traffic on each segment P jFor example, it is based on a neural network (NS) architecture. Time is considered a dimension on which convolutions are performed to extract relationships between variables (contextual data) at different times. A UNET-type architecture can be used.

[0096] In an example implementation, for a short-term prediction (15 min to 2 h), the output time sequence is discretized at 15 min and has a length of 16 (4 hours total). The explanatory variables (inputs) are hybrid:

[0097] - time series of observations from the last few hours on other geographical portions at distances that allow us to gain explainability for the prediction of the target portion.

[0098] - exogenous variables such as the contextual data described above.

[0099] This same architecture can be adapted with larger time steps for longer-term predictions.

[0100] A learning (or training) phase in E45, using this architecture, allows us to define a predictive model for the attendance rate per geographical area (Mod-est-F / P). j ).

[0101] This model is then applied during an inference step of this neural network (E46), to the input data from step E21 which indicates route segments between a point A and a point B given in the user request for route recommendation and to the context data (D. Context.) which relates to the request and the requested time range.

[0102] An estimate of the traffic rate is then obtained for the geographical areas concerned by the route recommendation request (East-F / P j ).

[0103] At step E47, an estimate of the traffic volume per segment of the route(s) obtained at step E21 is determined. At this stage, a refinement is therefore made to bring the estimate down to the segment level rather than to the level of geographical portions as defined above.

[0104] Thus, for a given portion, the associated route segments are identified according to a given map of the portion considered.

[0105] If there is only one segment in this portion, the estimate of the number of people using active mobility, given for the portion considered, is therefore attributed to the only segment of this portion.

[0106] If there are several segments in this section, then we can apply a rule of proportionality: the number of people using soft mobility on the section considered corresponds to the number of people (frequency rate) on all the segments (fragments of road) on this section whose total length we calculate.

[0107] Let P1 be the portion under consideration, F P1 The estimate of the number of people (attendance rate) for this portion and L1 the length of the set of segments identified for this portion P1, we obtain the number of people on a segment i of length li, (F Si ) according to the following formula: F Si =li*F P1 / L1.

[0108] It is possible to refine this calculation by taking into account the type of road (segment), for example if the segment considered is a shopping street, a pedestrian street, an expressway, a cycle path, etc... In this case, a weighting can be applied according to the type of segment, a greater weighting could for example be given to pedestrian streets or cycle paths where the number of people in soft mobility is potentially greater than for expressways.

[0109] At the end of step E47, we obtain an estimate of the traffic rate on the segments of the routes given in step E21. These routes are then reviewed in step E23, as described in reference to take into account these estimated traffic rates and thus maximize the segments of the route to be recommended which have a traffic rate above a threshold.

[0110] In a second embodiment, the data collected in step E41 of the lasont method consists of GPS location data from mobile devices in the relevant geographic area. However, retrieving this data requires widespread consent from the individuals carrying these devices, including authorization to transmit and use this location information. An insufficient number of authorizations would make the method less reliable. The collected data then provides a precise position, allowing for a faster estimation of the movement speed of the device users.

[0111] In a third embodiment, the data collected in step E41 of the lasont are video data captured via cameras installed in certain areas, for example, surveillance cameras. Analysis of this data can be used to estimate the number of people present in a segment where a camera is actually planned; analysis at different times allows for the determination of the individuals' speed of movement.

[0112] However, this method of implementation requires a sufficient presence of cameras for all segments to be analyzed, which is not currently the case for all urban areas.

[0113] Figures 5a and 5b illustrate route representations and their constituent segments. Figure 5a illustrates a recommended route between point A and point B for a given date J0 and time, here 23h (H=23). This route comprises 15 segments, 11 of which have an estimated usage rate F Si greater than the threshold S defined either by default or by the user. In this example representation, the segments are of equal length. A risk rate can be easily calculated by considering the number of segments for which the frequency rate is below the threshold relative to the total number of segments in the route. This risk rate therefore has a value of R = 4 / 15 = 0.27.

[0114] This illustrates the same route for a different time period. The date here corresponds to D1 (the day after D0) for a time equal to 2:00 AM (H=2). For this time period, the number of segments for which the estimated attendance rate is below the threshold S has increased. This number is now 8. The risk rate is therefore equal to R=8 / 15=0.54. This risk rate is thus higher.

[0115] Therefore, a user planning to leave a party, for example, might prefer to leave at 11 p.m. rather than 2 a.m. to benefit from busier streets on their route from point A to point B and feel safer. This risk assessment can be calculated over several time periods, for example, hour by hour, to allow the user to have all the necessary information and make an informed decision about their departure time.

[0116] This illustrates another representation of a route recommendation that can be displayed on the user's terminal screen. In this representation, several routes are proposed with segments that are not all the same size. These segments can, for example, represent sections of streets without intersections.

[0117] Each segment is associated with an estimated traffic volume. Naturally, a representation illustrating the topology of the roads and paths can be provided for better visibility.

[0118] In this illustration, a first route IT1 comprises 6 segments (S1, S2, S3, S4, S5, and S6) of different lengths, for example, S1=200m, S2=50m, S3=300m, S4=300m, S5=100m, and S6=100m. The total distance is therefore 1050m. For each segment, an estimated usage rate of people using active transportation is associated, for example, using the method described with reference to Figures 4a and 4b. For this first route, the usage rates (F Si (with i from 1 to 6) are all greater than or equal to the threshold S with a value of 10. Thus, F S1 =12, F S2 =10, F S3 =15, F S4 =14, F S5 =11 and F S6 =10. An average usage rate can be determined for this route to provide a basis for comparison between several routes. The average for this route IT1 is therefore (12+10+15+14+11+10) / 6=12.

[0119] Thus, another route, IT2, is also proposed. This one comprises 5 segments (S1', S2', S3', S4', and S5') of different lengths, for example, S1'=100m, S2'=100m, S3'=300m, S4'=350m, and S5'=50m. The total distance is therefore 900m on this route, which is shorter than the IT1 route. For this route as well, each segment is associated with an estimated usage rate, for example, using the method described with reference to Figures 4a and 4b. The usage rates (F Si (with i from 1 to 4) are greater than or equal to the threshold S with a value of 10, only one segment has a rate below this threshold. Thus, we have F S1’ =14, F S2’ =12, F S3’ =30, F S4’ =15 and F S5’ =8. The average usage rate for this route is (14+12+30+15+8) / 5=15.8.

[0120] Thus, the user has the choice between a slightly longer route (+150m) but with an estimated sufficient attendance rate for all segments of the route and an average of this rate over the whole route slightly above the chosen threshold, here of 10, and a shorter route but one of whose segments has an estimated insufficient attendance rate and an average of this rate over the whole route, higher than that of route 1.

[0121] The user may therefore still be interested in choosing the second route since it offers an attractive average traffic rate, a shorter route and only a short segment for which the traffic rate is not satisfactory.

[0122] In one particular embodiment, an option to request assistance may be provided for segments whose estimated ridership is below the threshold S. The main steps of this embodiment are illustrated. Once a route has been proposed in E24 (described with reference to the), it is checked whether a segment S i of the route is associated with an estimated usage rate for a given time period, below the threshold S. In the case where the segment S i is less than the threshold S (O to E71), so we check in E72 that there are people present in the segment preceding this segment S i (S i-1) registered with a support service. This service offers support from at least one person registered as a support worker with that service. The support worker's contact information is stored in a dedicated database for this service, so they can be contacted when needed. Conditions of use for this contact information may be associated with it, such as a possible time slot, prohibited areas, or any other constraints. If at least one person exists in segment S i-1 (O at step E72), then a notification is sent to that person at E73. This notification is a request for assistance from a user wishing to travel along segment S iWithin a given time slot, if the person receiving this notification is present in a segment preceding or near segment Si and is available during the requested time slot, and if the person accepts this accompaniment (O at step E74), then a meeting place is notified to both the accompanying person and the person requesting the accompaniment in E75. This meeting place could, for example, be at an address in segment S. i-1 of the route taken by the user, this segment having a sufficient traffic rate for the user to feel safe.

[0123] If the person registered as an escort does not accept the proposed escort service, then E77 checks whether there are other people registered as escorts in an area near S. iIf this is the case (O at step E77), then the contact details of this new person are retrieved in E76 to send them a notification in E73. Otherwise, this support service cannot be offered.

[0124] With this service option, users wishing to travel from point A to point B can choose a route they feel safe on, based on information about the traffic levels of the segments they will be taking. If the route includes areas with lower traffic levels, insufficient for the user to feel safe, then other criteria can be considered to measure other parameters, as illustrated in [reference to relevant section], or the escort service described therein can be implemented.

[0125] Thus, the user can choose a safer route for themselves, according to the recommendations from the recommendation process as described.

Claims

Route recommendation method for a user in soft mobility in which the selection (E23) of a maximum of segments composing the route is carried out if an estimated frequency rate (E22) for the segment considered exceeds a threshold. Method according to claim 1, wherein the estimated attendance rate is an attendance rate of people using soft mobility. A method according to any one of the preceding claims, wherein the estimated usage rate for a route segment is calculated from a database containing network signaling data from mobile terminals present in that segment. A method according to claim 3, wherein a historical database is constructed from network signaling data collected on predefined portions of a geographical area, this historical data including data on the frequency of people using soft mobility per predefined portion, at given times. A method according to claim 4, wherein a predictive model of the attendance rate of people using active mobility for a given portion is obtained from historical data and contextual data. A method according to claim 5, wherein the prediction model is obtained by training a neural network, an inference of the neural network giving an estimate of the attendance rate for a given portion from a route request and associated context data. A method according to claim 6, wherein an estimate of the usage rate for a route segment is obtained from the estimate of the usage rate of the geographical portion including this segment and mapping data of the geographical portion. A method according to any one of the preceding claims, wherein attendance rates are estimated as a function of a time range. A method according to any one of the preceding claims, wherein the route segments are proposed if their estimated lighting rate further exceeds a brightness threshold and / or if their estimated crime rate is further below a crime threshold. A method according to claim 9, wherein the estimated lighting rate for a route segment is obtained from data captured by environmental cameras. A method according to any one of the preceding claims, wherein a step of displaying the segments and their attendance rate is carried out. A method according to any one of the preceding claims, wherein, in the case where a segment of a recommended route includes at least one segment for which the estimated attendance rate is less than a threshold, the following steps are implemented: - searching for at least one person located in a segment close to the segment for which the estimated attendance rate is less than the threshold and for which their contact details are recorded in a database of an assistance service; - sending a notification of a request for assistance to at least one person located; - in the case of a positive response to the request for assistance, sending an appointment notification to both the person located and the user. Communication terminal comprising a processing circuit for implementing the recommendation process according to any one of claims 1 to 12. Server comprising a processing circuit for implementing the route recommendation method according to any one of claims 1 to 12. A processor-readable recording medium on which is recorded a computer program containing instructions for carrying out the recommendation process according to any one of claims 1 to 12.

Images