FR3028651A1 - AUTOMATIC DETERMINATION OF SPEED LIMITATION ON A ROAD FROM A NAVIGATION SYSTEM - Google Patents

Abstract

The invention relates to a method for the automatic determination of a speed limit (SLs) in force on a road taken, or about to be borrowed, by a motor vehicle, in which it is established, by means of a first system (2), said navigation system including in particular a data receiver (20, 21) of a geographical positioning system embarked on said motor vehicle and mapping data (22), a probable speed limitation (SLNAV) . The method is characterized by an extraction step, from at least one attribute (Att) issued by the navigation system (2) and relating to a road context data, a road context in which the vehicle is located; and by a step of determining the speed limit in effect (SLs) on the road considered, from said likely speed limitation (SLNAV) and taking into consideration said road context extracted.

Description

The present invention relates to a method for determining speed limits on a road taken by a motor vehicle, and a system for setting up a speed limit on a road used by a motor vehicle. of this process. The main purpose of the invention is to propose a solution for automatically determining information relating to a speed limit that applies to a road taken, or about to be borrowed, by a motor vehicle. The information thus determined is then exploitable in different applications equipping the vehicle in question. In particular, but without limitation, the information relating to a speed limit is used in the context of a driving assistance system by explicitly indicating to the driver what is the maximum speed allowed, for example when this last is in excess of this maximum speed. The field of the invention is, in general, that of the driver assistance, which proposes a set of assistance systems for the driver, essentially intended to improve the conditions of traffic safety. Thus, for example, the following have recently been developed: night vision systems to help the driver to detect in advance obstacles which are difficult to perceive in night-time traffic conditions; Early warning systems for turning, to warn the driver as soon as possible of the presence of a turn; As part of the driver assistance systems, it is also now sought to provide the driver with a warning system of speed limits: such a system must automatically detect a speed limit in force on a borrowed road, or on a road that is about to be taken by a vehicle. In fact, actions concerning excessive speeds are necessary to reduce the number of accidents and the severity of their consequences. Many drivers do not respect the prescribed speeds: 40% of drivers do not respect them on motorways, 60% on national and departmental roads and 25% exceed the regulatory speed in the city by more than 10 km / h. Various solutions have been proposed to date to enable a vehicle to automatically determine a speed limit in force on a road. A first solution lies in the exploitation of information from a navigation system. Navigation systems are increasingly providing vehicles for guidance of the driver from a starting point (corresponding to the current position of his vehicle) to an end point (selected by the driver). With such a navigation system including a geographic positioning system data receiver and mapping data, it is known to establish a likely speed limitation. However, a number of inherent faults in this system limit its effectiveness: The current mapping is still very imprecise. It happens very often that in a given place, the information is missing. Indeed, there are 20 whole areas of the world that are not covered by the mapping databases; Sometimes the information provided by the navigation system is aberrant. For example, if the driver has planned to go to a place B which he has memorized in his navigation system and if, finally, on the way, he is led to go to a place C without following the indications given. by the navigation system, then the information given by the navigation system is inconsistent or even contradictory, with respect to the characteristics of the trajectory actually followed by the vehicle; Changes in the configuration of the road due to one-off events, for example the carrying out of work, entail a modification of the speed restrictions in force on the portion of road considered; these modifications are not then known to the navigation system; 3028651 3 Loss of GPS coverage is also possible, for example when traveling under a long tunnel. Finally, the navigation system is imprecise, the inaccuracy being of the order of ten meters. In order to overcome at least some of the various disadvantages mentioned above, another solution, described in particular in document EP 2 017 807 B1 in the name of the Applicant, consists in combining the data provided, on the one hand, by a navigation system and, on the other hand, by a system associating an on-board camera and image processing applications. More specifically, a method for automatically determining a speed limit in force on a road taken, or about to be borrowed, by a motor vehicle, comprises, according to the teaching of this document, the following steps: establishing, by means of a first system, said navigation system 15 notably involving a data receiver of a geographical positioning system and mapping data, a probable speed limitation associated with a first confidence index; constituting a first set of information comprising at least the likely speed limitation and the first confidence index; to establish, by means of a second system, an image processing system, notably involving a camera and image processing applications able to identify and interpret speed limit panels arranged in the vicinity of the road, a likely speed limitation associated with a second confidence index; - constitute a second set of information including at least the probable speed limit, and the second confidence index; and - determining the speed limit in force on the considered route, from the first set of information and the second set of information and taking into account the first confidence index and the second confidence index. Different calculation methods exist to determine the first confidence index, such as a weighted sum of several criteria that can involve the accuracy of the positioning of the vehicle by the navigation system 3028651, the level of information on the road, the functional class of the road, the type of road (highway, main road, secondary road), the environment (city, motorway exit ...), the guide mode selected or not by the driver ... 5 An example of calculation of the second confidence index will involve a weighted sum of different criteria, such as the shadow factor on the image considered, the symmetry index of the image considered, the coherence index of identification of speed limitation from one image to another ... Advantageously, the information relating to the speed limits provided by each of the two systems is extrapolated to consider other speed limits c as being likely to be in force on the route in question; each of these other speed limits is then associated with a weighting coefficient, called mass of belief, involved in the merging of all the information then available relating to the speed that may actually be in effect, merging information ensuring the final determination of the desired speed limit. In general, the determination of the confidence indices, and / or belief masses, and their intervention in the fusion of knowledge from the two systems depends on the merger strategy adopted. Advantageously, various methods extracted from the theory known as beliefs can be used in the data fusion. Notably, one of the methods known as the Dempster-Shafer conjunctival combination, associated with a so-called Dempster-Shafer equation, provides particularly convincing results. Other methods, based on Bayesian theories, or fuzzy logic set theory, can also be used in data fusion. Once the speed limit in force has been established, it can for example be displayed on a screen.

The confidence indices used for the merger are updated each time new data is provided by the navigation system or by the image processing system 102.

With the system described in the document EP 2 017 807, a determination of the speed limit in force is obtained that is much more reliable than with systems that only take into account data originating from a navigation system, or data from an image processing system. In particular, the risks of confusion due to erroneous information or misinterpretation of this information are avoided. We can also propose a degraded mode of operation, useful in the event of failure of one of the two systems, based on the other system, not defective.

Nevertheless, in practice, the Applicant has found that such a fusion system does not make it possible to satisfactorily manage all the driving situations that a user may be confronted with. In particular, the previous system does not work optimally when the camera on the vehicle detects a sign indicating a speed limit. Indeed, in this case, the fusion system will operate in a degraded mode that only takes into account the data provided by the navigation system. In this case, the problem is related to the inaccuracy of the geographical positioning, of the order of a hundred meters, of the navigation system.

It is an object of the present invention to overcome the foregoing limitations by providing a method for automatically determining a speed limit on a road from a likely speed limitation provided by an on-board navigation system that is more reliable. and more accurate.

More precisely, the subject of the present invention is a method of automatically determining a speed limit in force on a road taken, or about to be taken, by a motor vehicle, in which it is established, by means of a first system, said navigation system including in particular a data receiver of a geographical positioning system on board said motor vehicle and mapping data, a likely speed limitation, the method being characterized in that it comprises the following steps: extracting, from at least one attribute delivered by the navigation system and relating to a road context datum, a road context in which the vehicle is located; - determine the speed limit in force on the road in question, based on the said likely speed limitation and taking into consideration the said road context.

In addition to the main features which have just been mentioned in the preceding paragraph, the method according to the invention may have one or more additional characteristics among the following: said at least one attribute relating to a road context data makes it possible to distinguish the overall road conditions {city, highway, 15 exit, other}; the extraction step comprises a determination of said road context from a first attribute relating to the type of road, and a second attribute relating to the driving situation, the first and second attributes being delivered by the system of navigation; The determination step comprises, for each possible speed limit SLi taken in a predetermined set of regulatory speed limits, a calculation of a probability P (SLi) according to the following linear combination: ai CCONT CONT + C NAVa NAV 25 in which aiCONT is a confidence index function of each speed limit SLi possible and extracted context; ai NAV is a confidence index depending on each speed limitation SLi possible and the probable speed limitation SLNAV provided by the navigation system; and 3028651 7 and CNAV two predefined weighting coefficients related CCONT respectively to context and navigation; as a variant, the calculation of the probability P (SLi) is a function of the following linear combination: ## EQU1 ## in which oc iTRANS is a confidence index depending on each possible speed limitation SLi and a speed transition possible; and CTRANS is a predefined weighting coefficient related to the possible velocity transition; The confidence index (xTRANS is extracted from a predefined transition table, giving all the confidence index values for each possible speed limitation and for each possible speed transition, the confidence index aiCONT is extracted from a previously established context table, giving all the confidence index values for each possible speed limitation and for each context - the context table is updated by learning - the coefficient of the context-related CCONT weighting has a value greater than that of the navigation-associated CNAV weighting coefficient 20 - the CTRANS weighting factor associated with the possible speed transition to a value between the value of the associated weighting coefficient 3028651 8 CCONT the context and the value of the CNAV weighting coefficient associated with navigation - the confidence index aiNAV is set to 1 if the limit SLi possible speed is equal to the probable speed limitation SLNAV, and 0 otherwise; 5 - the speed limit in force on the considered road determined corresponds to the possible speed limitation associated with the highest probability P (SLi). - the predetermined set of regulatory speed limits is country specific.

The present invention also relates to a system for automatically determining a speed limit in force on a road taken or about to be taken by a motor vehicle, comprising a first system, called a navigation system, including geographical positioning system and mapping data for establishing a likely speed limitation, characterized in that it comprises: means for extracting, from at least one attribute issued by the navigation system and relating to a road context data, a road context in which the vehicle is located; Means for determining the speed limit in effect on the road considered, based on said likely speed limitation and taking into consideration said road context extracted. The invention will be better understood on reading the description which follows, with reference to the appended figures, in which: FIG. 1 schematically represents a system for the automatic determination of a speed limit in force on a road 3028651 borrowed, or about to be borrowed, by a motor vehicle in accordance with a possible implementation of the invention; FIG. 2 illustrates an example of a so-called context table established according to navigation contexts, in accordance with the principle of the present invention; FIG. 3 illustrates an example of calculation for the determination of a speed of limitation, using the context table of FIG. 2; FIG. 4 gives an example of a speed transition table that can be used in a variant of the method according to the invention; FIG. 5 illustrates another example of calculation involved in the determination of a speed of limitation, using the context table of FIG. 2 and the speed transition table of FIG. 4; - Figure 6 schematically shows another architecture of a system implementing the invention.

The various elements appearing in several figures will have kept, unless otherwise specified, the same reference. With reference to FIG. 1, a possible architecture for a system 1 for automatic determination of a speed limit in force on a road taken, or about to be borrowed, by a motor vehicle according to the invention will The system 1 uses, in accordance with the invention, information delivered by an on-board navigation system 2, in particular using a data receiver 25 of a geographical positioning system on board said motor vehicle associated with a vehicle. antenna 21 and mapping data from a database 22. The navigation system 2 may furthermore involve other sensors 23 (speed sensor, gyroscope, etc.) capable of delivering various information making it possible in particular to verify the coherence between the course actually followed by the vehicle, and the course provided by the navigation system 2. The navigation system 2 provides a likely speed limitation which will be noted in the SLNAV suite.

The navigation system 2 will also provide, in a known and non-detailed manner, a number of attributes, such as for example a first attribute relating to the type of road (motorway, national, departmental, municipal ...), and a second attribute relating to the driving situation (city, out of town, intersection ...). According to the invention, a processing module 10 of the system 1 will extract, from at least one attribute Att issued by the navigation system 2 and relating to a road context data, the road context in which the vehicle is traveling. find and then determine the effective speed limit SLs on the considered route, starting from the probable speed limitation SLNAV and taking into consideration said road context extracted. By way of nonlimiting example, the joint use of the first attribute relating to the type of road, and the second attribute relating to the driving situation makes it possible to distinguish the following four road contexts: - city, motorway, - exit, - other.

Once the road context has been extracted, the processing module 10 uses a prior context table 11, which groups, for each possible speed limit SLi corresponding to a regulatory speed limitation, aCONT confidence indices for each context. FIG. 2 gives an example of content of a table 11 giving values of the confidence index aCiONT for sixteen possible speed limits, denoted from SL1 to SL16, and in each of the four aforementioned road contexts (V = City; A = Highway, S = Exit, O = Other). The speed limits SL1 to SL14 correspond here to all the speed limit regulations (SLR in km / hour) that are found 30 depending on the country in which the vehicle travels, for example, for France, France. following set: 3028651 11; 10; 20; 30 ; 45; 50; 60; 70; 80; 90; 100; 110; 120, 998, 999} Two speed limits SL15 and SL16 are furthermore preferably used for unknown speed, for example 998 when the navigation is not able to deliver information on the speed of a portion of a road and implicit speed, for example 999, during an end of speed limit detection. The number in bold in each of the last four columns of the table 11 corresponds to the highest confidence index of the column, and thus makes it possible to identify the most likely speed limit for a given context. Thus, in the "City" context or in the "Exit" context, it is the speed limit SL6, ie 50 km / h in France, which has the highest probability of being met (confidence index equal to 0.936 or at 0.436) 15 whereas in the "Other" context, the most likely speed limitation is SL10, ie 90 km / h in France, with a confidence index of 0.76. The table 11 contexts, such as that shown in Figure 2, is obtained beforehand and stored in the system 1 on the vehicle. It can be advantageously updated by learning by performing statistics on the different contexts encountered when the vehicle is traveling. From this table 11 of contexts, the processing module 10 will compute, for each possible speed limitation, and according to the context extracted a probability P (SLi) which is a function of the following linear combination: C i CONT CONT + CNAVaNAV in which: aCONTis the confidence index function of each possible speed limit SLi and the extracted context, which is found in the context table 21; NAV is a confidence index according to each speed limitation SLi possible and the probable speed limitation SLNAV provided by the navigation system 2; and and CNAV two predefined weighting coefficients - CCONT 5 related respectively to context and navigation. For example, each probability P (SLi) is calculated by applying the following relation: ai P (SLi) = CCONTCONT + CNAVaNAV (1) 2 with ai = 1 if SLi NAV SLNAV and 10 aNAV = O otherwise The values of the weighting coefficients CCONT and CNAV depend on the importance that we want to give to the navigation system 2 on the one hand, and to the table 11 of the contexts on the other hand. The context-associated weighting coefficient CCONT preferably has a value greater than that of the navigation-associated CNAV weighting coefficient. For example, it is possible to choose: CCONT = 0.7 CNAV = 0.2 FIG. 3 gives an example of the probability calculations performed by the processing module 10 in a particular situation for which the navigation system 2 provides the value of 70. km / h as probable speed limitation SLNAV, while the context extracted by the processing module 10 corresponds to the "Other" road context. The first column of Figure 3 indicates all possible speed limits SLi. The second column gives the correspondence with the regulatory speed limits, according to the country, the third column gives the value used 5 for the weighting coefficient CCONT the fourth column gives the value of the confidence index ocicoNT for the context "Other" (value retrieved in the context table 11), the fifth column gives the value used for the weighting coefficient CNAV 'and the sixth column gives the value of the confidence index aiNAV' In this example, only the speed limit 10 SL8 (70 km / h) has a confidence index aNAV equal to 1, since the probable speed limitation SLNAV provided by the navigation system 2 is 70km / h. The last column of Figure 3 shows the result of the calculation of the probability P (SLi) with equation (1) above. In this column, the number in bold identifies the highest probability, here 0.266, which corresponds to the speed limit SL10 (ie 90 km / h). Here, the system 1 will therefore determine that the speed limit SLs in force on the road taken is the speed limit SL10 (ie 90 km / h), while the navigation system proposed 70 km / h. In a variant of the method according to the invention, it is possible to use, in addition to the table 11 of contexts described above, another table 12, called the speed transition table, grouping, for each possible speed limitation SLi corresponding to a Regulatory speed limitation, aTRANS confidence indices for each possible speed transition. FIG. 4 gives an example of a content of a table 12 giving values of the confidence indices aiTRANS for the sixteen possible speed limits SL1 to SL16. Each row of this table 12 gives the probability that there is a transition from a first possible speed limit to a second possible speed limit. For example, for the possible speed limit SL8 (here 70 km / h), the confidence index corresponding to the probability that one passes from the speed limit SL8 to any of the speed limits SL1 to SL5 is equal to 0, while the confidence index corresponding to the probability of passing from the SL8 speed limit to the speed limit SL6 (ie 50 km / h) is 0.4642. Here again, the most important value in each row has been bolded. The table 12 of speed transitions, such as that shown in Figure 3, is also obtained beforehand and stored in the system 1 on board the vehicle. It can be advantageously updated by learning by performing statistics on the situations encountered when the vehicle is traveling. In this variant, the probability P (SLi) then becomes a function of the following linear combination: CCONTcx CONT + CNAVaNAV + CTRANSa.TRANS in which aTiRANS is a confidence index provided by the table 12 of transitions of speeds; and is a predefined weighting factor related to the possible CTRANS transition. Here again, the value of the weighting coefficient CTRANS depends on the importance that is to be given to the table 12 of speed transitions. The weighting coefficient CTRANS preferably has a value between the value of the weighting coefficient CcoNT associated with the context and the value of the weighting coefficient CNAV associated with the navigation.

Each probability P (SLi) is for example calculated by applying the following relation: ai P (SLi) = CCONTCONT + CNAVaNAV + CTRANSaTRANS (2) 3 with aiNAV = 1 if SLi = SLNAVet 5 aiNAV = 0 otherwise One can choose for example: CCONT = 0.7 10 CNAV = 0.2 and CTRANS = 0.3 FIG. 5 gives an example of the calculations of probabilities performed by the processing module 10 according to this variant, always in a particular situation for which the navigation system 2 provides the value of 70 km / h as probable speed limitation SLNAV, while the context extracted by the processing module 10 corresponds to the "other" road context. It is also assumed that the speed limitation previously recommended by system 1 was 50 km / h. The first column of Figure 5 indicates all possible speed limits SLi. The second column gives the correspondence with the regulatory speed limits, according to the country, the third column gives the value used for the weighting coefficient CCONT 'the fourth column gives the value of 25 the confidence index aCONT for the context' Other (Value retrieved in context table 11), fifth column gives the value used 3028651 16 for CNAV weighting coefficient 'sixth column gives the value of the confidence index aiNAV the seventh column gives the value used' for the weighting coefficient C TRANS 'and the eighth column gives the value of the confidence index a iTRANS for a transition from the speed limit previously recommended (in our example 50 km / h) to any of the limitations of speeds possible. The eighth column thus corresponds to the line SL6 of the table 12 shown in FIG. 4. In this example, only the speed limit SL8 (70 km / h) has a confidence index 8, aNAV equal to 1, since the limitation of Probable speed SLNAV provided 10 by the navigation system 2 is 70km / h. The last column of Figure 5 shows the result of the calculation of the probability P (SLi) with equation (2) above. In this column, the number in bold identifies the strongest probability, here 0.207333, which corresponds to the speed limit SL10 (ie 90 km / h).

Here again, the system 1 will therefore determine that the speed limit SLs in effect on the road taken is the speed limit SL10 (ie 90 km / h), while the navigation system proposed 70 km / h. The method for determining the speed limit SLs is applicable when the data from a navigation system 2 is used, in this case the probable speed limitation SLNAV and the attributes making it possible to extract the context. road. As shown diagrammatically in FIG. 6, the system 1 as just described can also be a subsystem of a more complex system 3, such as that described in document EP 2 017 807, which merges in normal time a probable speed limitation data SLNAv provided by a navigation system 2, and a probable speed limitation data SLIMG provided by a system 4 for image processing including a camera 41 and processing applications 42 images capable of identifying and interpreting speed limiting panels disposed in the vicinity of the road. In this case, the subsystem 1 will only be activated in the situations where the system 3 operates in a degraded mode in which only a probable speed limitation SLNAV is available. Such a situation is for example encountered when the image processing system 4 detects a panel corresponding to an end of speed limitation, for example a city exit, or a panel dedicated to the end of limitation (generally a barred circular panel), and then delivers an EOSL signal (English initials set for End of Speed Signal Limitation). 10

Claims (15)

REVENDICATIONS1. A method for automatically determining a speed limit (SLs) in force on a road taken, or about to be taken, by a motor vehicle, in which, using a first system (2), said navigation system including a data receiver (20, 21) of a geographical positioning system embarked on said motor vehicle and mapping data (22), a probable speed limitation (SLNAV), the method being characterized in that it comprises the following steps: extracting, from at least one attribute (Att) delivered by the navigation system (2) and relating to a road context datum, a road context in which the vehicle is traveling. find ; - determine the speed limit in force (SLs) on the road considered, from the said probable speed limitation (SLNAV) and taking into consideration the said road context extracted. 2. Method according to claim 1, characterized in that said at least one attribute relating to a road context data makes it possible to distinguish the road contexts in the set {city, highway, exit, other}. 3. Method according to claim 2, characterized in that the extraction step comprises a determination of said road context from a first attribute relating to the type of road, and a second attribute relating to the driving situation. , the first and second attributes being delivered by the navigation system (2). A method as claimed in any one of the preceding claims, characterized in that the determining step comprises, for each possible speed limitation SLi taken in a predetermined set of regulatory speed limits, a calculation of a probability P ( SLi) function of the following linear combination: ai CCONT cont + CNAVa NAV 3028651 19 in which aCONT is a confidence index function of each possible speed limit SLi and the extracted context; NAV is a confidence index according to each possible speed limit SLi and the probable speed limitation SLNAV provided by the navigation system (2); and and CNAVs have two predefined weighting coefficients that are CCONT-related to context and navigation respectively. 5. Method according to claim 4, characterized in that the calculation of the probability P (SLi) is a function of the following linear combination: C ai a + C ii 10 C CONT CONT NAV + TRANS NAV TRANS where aTi RAN is a subscript confidence depending on each speed limitation SLi possible and a possible speed transition; and is a predefined weighting coefficient related to the possible CTRANS transition velocities. 6. Method according to claim 5, characterized in that the confidence index aTRANS is extracted from a predefined transition table (12) giving all the confidence index values for each possible speed limit and for every possible gear transition. 7. A method according to any one of claims 4 to 6, characterized in that the confidence index aCONT is extracted from a previously established table (11) of contexts, giving all the confidence index values for each speed limitation possible and for each context. 8. Method according to claim 7, characterized in that said context table (11) is updated by learning. 3028651 9. Method according to any one of claims 4 to 8, characterized in that the weighting coefficient CCONT associated with the context has a value greater than that of the weighting coefficient CNAV associated with the navigation. 10. Method according to claims 5 and 9, characterized in that the weighting coefficient CTRANS associated with the possible speed transition to a value between the value of the weighting coefficient CCONT associated with the context and the value of the weighting coefficient CNAV. associated with navigation. 11. Method according to any one of claims 4 to 10, characterized in that the confidence index aiNAv is set to 1 if the speed limit SLi possible is equal to the probable speed limitation SLNAV, and 0 otherwise . 12. A method according to any one of claims 4 to 11, characterized in that the current speed limit (SLs) on the considered route is the potential speed limitation associated with the highest probability P (SLi). Method according to any one of claims 4 to 12, characterized in that the predetermined set of regulatory speed limits is country dependent. 20 14. System (1) for automatically determining a speed limit (SLs) in force on a road taken, or about to be borrowed, by a motor vehicle, comprising a first system (2) navigation, including geographical positioning system (20, 21) and mapping data (22) for establishing a probable speed limitation (SLNAV), characterized in that it comprises: - means (10) capable of extracting, from at least one attribute issued by the navigation system (2) and relating to a road context data, a road context in which the vehicle is located; means (10, 11, 12) for determining the speed limitation 5 in effect (SLs) on the road considered, based on said probable speed limitation (SLNAV) and taking into consideration said extracted road context. 15. System (1) according to claim 14, characterized in that it constitutes a subsystem of a system (3) of fusion capable of automatically determining a speed limit in force on a borrowed road, or on the to be borrowed by a motor vehicle, on the one hand, from the probable speed limitation (SLNAV) provided by the navigation system, and on the other hand, from a probable speed limitation (SLIMG) provided by an image processing system (4) including a camera (41) and image processing applications (42) for identifying and interpreting speed limiting panels disposed in the vicinity of the road, said subsystem (1) being activated when the image processing system (4) delivers a signal (EOSL) representative of the detection of a speed limit termination panel.