FR3111225A1 - Computer-implemented method of road traffic management in a road area - Google Patents

Abstract

A computer-implemented method of road traffic management of a road zone (Z) divided into a plurality of sectors (S1,…,Si,…,Sn), in which, for each sector (Si), a centralized controller (20) calculates an avoidance probability (PSi,t) corresponding to the fraction of vehicles that must avoid said sector (Si), to make traffic more fluid in the road zone (Z) and establishes at least one centralized avoidance map ( 30) associating each of the avoidance probabilities (PSi,t) calculated with the corresponding sector (Si), andthe positioning and navigation aid system (13) of each of the vehicles (121) present in the road zone, transcribes the centralized avoidance map (30) into a personalized avoidance map (40), by randomly drawing from the avoidance probability (PSi,t), to calculate a final route of shorter duration (41 ) based on the custom avoidance map (40).Abstract figure: Figure 3

Description

Computer-implemented method of traffic management of a road area

The present invention relates to the field of road traffic management, particularly in urban areas. More specifically, the present invention relates to a method implemented by computer for managing road traffic in a road zone, to a data processing device allowing the implementation of this method, to a computer program comprising instructions which enable the method to be performed and a computer-readable recording medium comprising instructions which enable a computer to perform the method.

Cities are increasingly seeking to improve the management of their road traffic. Indeed, the improvement of road traffic management makes it possible to reduce the time spent by users in their vehicle, and therefore the pollution generated by road traffic.

The regulation of road traffic in cities is done mainly through traffic lights at crossroads. However, this is a very local traffic management and it does not take into account the entire road network. This management is therefore not optimized to manage road traffic at the scale of an agglomeration for example.

Moreover, when users choose their routes, they are completely free in their choice. No external device constrains them in their choice. Thus, to go from an initial position to a final position, each vehicle generally seeks to follow the shortest route in time proposed by its navigation aid system. This method of route calculation makes it possible to obtain an individual optimum. However, this individual optimum is known to be sub-optimal compared to a collective optimum aimed at minimizing the time lost by all the vehicles circulating on the road network. Studies have shown in particular that the total time spent by vehicles on an urban network can be reduced by 10% to 30% by seeking to reach the collective optimum compared to seeking to reach the individual optimum.

To achieve the collective optimum, it is best to distribute the vehicles over the entire road network, in order to better distribute the demand for vehicle travel in relation to the capacities available on the roads that make up the network.

We know suitable methods for managing road traffic taking into account the entire road network, and therefore allowing in particular to manage road traffic at the scale of an agglomeration.

A first known method is the perimeter control method. Perimeter control consists of defining an area in which one seeks to optimize traffic conditions by dynamically limiting the total number of vehicles entering the area. To dynamically limit the total number of vehicles that can enter the zone, the zone is equipped with traffic lights around its periphery, and the duration of the cycles of these traffic lights can be changed according to traffic conditions. Thus, the shorter the duration of the green light, the lower the number of vehicles that can enter the zone; and, conversely, the longer the duration of the green light, the greater the number of vehicles that can enter the zone. However, this method of perimeter control tends to create congestion outside the zone, and particularly at the periphery of the zone, because vehicles cannot anticipate the effect of the regulation.

A second known method is the optimal guidance method consisting in defining the optimal trajectory for all the vehicles. It makes it possible to achieve optimal operation of the road network thanks to the anticipation, by users, of the effects of regulation. However, it is difficult to transcribe global measurements to individual vehicles and it is impossible to control the trajectory of all the vehicles in the road network. Thus, optimal guidance as it is theoretically studied has no operational application because we do not know how to practically implement it.

The prior art is therefore devoid of methods, implemented by computer, for managing road traffic in a road zone limiting congestion, allowing collective management of road traffic, easy to implement, and making it possible to anticipate the effects of regulation.

Objectives of the invention

In this context, the invention aims to satisfy at least one of the following objectives:
- to provide a computer-implemented method of traffic management of a road area improved compared to the method of perimeter control
- providing a computer-implemented road traffic management method of a possible road area to be implemented;
- provide a computer-implemented process for managing road traffic in a road zone, making it possible to overcome the classic antagonism on road networks between individual optimum and collective optimum;
- to provide a computer-implemented method of road traffic management of a road area that is easy to implement in practice, in particular in terms of the computing power necessary for a real-time application, while respecting the principle of life private, that is to say without recording personal information relating to users;
- to provide a computer-implemented method of traffic management of a road area limiting congestion;
- to provide a computer-implemented method of road traffic management of a road area allowing collective management of road traffic, and in particular collective management of the routes taken by vehicles;
- to provide a computer-implemented method of road traffic management of a road area enabling the effects of road traffic regulation to be anticipated;
- provide a device for the implementation of the process, targeted by the above objectives and computer-implemented road traffic management of a road area;
- provide a computer program for the implementation of the process, targeted by the above objectives and computer-implemented road traffic management of a road area;
- provide a computer-readable medium comprising instructions for implementing the process, targeted by the above objectives and implemented by computer, road traffic management of a road area.

Summary

The present invention meets these objectives.

To this end, according to a first aspect, the invention provides a method implemented by computer, for managing the road traffic of a road zone, preferably of an urban zone, implementing at least one centralized controller comprising a memory, the method comprising the following steps:
a) the road zone, divided into a plurality of sectors, is stored in the memory of the centralized controller,
b) the centralized controller determines a traffic state for each sector at a time t,
c) for each sector, the centralized controller calculates an avoidance probability, depending on the traffic state determined in step b) and corresponding to the fraction of vehicles having to avoid said sector, during a time period Δt following l instant t, to improve traffic flow in the road zone,
d) the centralized controller establishes at least one centralized avoidance map associating each of the calculated avoidance probabilities with the corresponding sector,
e) the centralized controller broadcasts the centralized avoidance map established in step d) to a plurality of vehicles present in the road zone and each equipped with a positioning and navigation aid system, during the period Δt according to time t,
f) for each of the plurality of vehicles whose positioning and navigation aid system is activated to work out at least one route of shorter duration from an initial position to a final position in the road zone :
- (f1) the positioning and navigation aid system selects at least one sector from among the sectors,
- (f2) for each of the sectors selected, the positioning and navigation aid system performs a random draw, preferably following a Bernoulli test, from the probability of avoiding the selected sector, making it possible to define a on state or a blocking state for each of said selected sectors,
- (f3) the positioning and navigation aid system transcribes the centralized avoidance map into a personalized avoidance map, by associating the passing state or the blocking state defined for each of the selected sectors with the corresponding sector ,
- (f4) the positioning and navigation aid system calculates a final route of shorter duration from the initial position to the final position, according to the personalized avoidance map, avoiding the associated selected sectors in the blocking state,
g) once the time period Δt has elapsed, steps b) to f) are repeated.

Thus, in this method of road traffic management, the entire road area is taken into consideration. Traffic regulation is therefore well done at the scale of the road zone, and not at specific points of the road zone.

Furthermore, the information on the state of the traffic in the sectors of the road zone is contained in the avoidance probabilities associated with each of the sectors in the centralized avoidance map, and this information is transmitted to all the vehicles present in the road area and equipped with a positioning and navigation aid system. Each vehicle takes into account the probabilities of avoidance, and therefore the state of traffic in the sectors of the road zone, in establishing its final route of shorter duration. Traffic regulation is therefore carried out on all of these vehicles, and allows optimal collective management of traffic. In particular, this collective traffic management is based on collective management of the routes taken by each of these vehicles.

Taking avoidance probabilities into account also makes it possible to control the number of vehicles in the sectors in order to prevent too many vehicles from being in a sector at a given time. The method therefore makes it possible to prevent congestion.

In addition, the creation of the centralized avoidance map requires a limited number of calculations, and the interpretation of the centralized avoidance map by the positioning and navigation aid system to produce the personalized avoidance map is done using a simple random draw, and does not require specific modifications to the positioning and navigation aid systems currently fitted to vehicles. This method therefore has the advantage of being easily implemented.

Furthermore, this process allows vehicles present in the road zone and equipped with a positioning and navigation aid system to calculate their final route of shorter duration, taking into account the sectors to be avoided, before reach these areas. Vehicles can therefore be directed ahead of these sectors so as to avoid these sectors. Vehicles can therefore anticipate regulation. The creation of a congestion zone at the entrance to the sectors to be avoided is therefore limited.

According to a variant of the invention, the initial position is the current position of the vehicle.

According to a variant of the invention, step (f1) comprises the following sub-steps:
- the positioning and navigation aid system calculates an initial route of shorter duration from the initial position to the final position without sector avoidance conditions,
the positioning and navigation aid system selects the sector or sectors crossed by the initial route of shorter duration on the avoidance map and associated with a non-zero probability of avoidance.

Thus, the random draw is only carried out for the sectors likely to be crossed by the vehicle and in which the traffic needs to be regulated, which limits the number of calculations carried out by the positioning and assistance system to navigation. The final route of shorter duration is therefore different from the initial route of shorter duration when the initial route of shorter duration passes through sectors associated with a non-zero probability of avoidance and which are then associated with the blocking state.

According to a variant of the invention,
- in step a), an average reference speed defined for each sector and corresponding to the optimal functioning of the sector, is stored in the memory of the centralized controller,
- in step b), the centralized controller determines an average speed of each sector at time t, and
- in step c), for each sector, the centralized controller calculates the avoidance probability as a function of the average reference speed and the average speed of each sector.

Thus, the centralized controller can compare sequentially or continuously the average speed of the vehicles in each sector, on the one hand, and the reference average speed of the corresponding sector, on the other hand. This comparison makes it possible to establish the difference between these 2 speeds. This difference allows a correction of the calculation of the probability of avoidance.

According to a variant of the invention, the average reference speed, for each sector, is the average speed reached in the sector when a critical number of vehicles is circulating in the sector, the critical number of vehicles being the number of vehicles circulating in the sector, when the sum of the distances traveled by all the vehicles present in the sector per unit of time reaches a maximum value.

According to a variant of the invention, the road zone comprises a road network, and the method comprises, prior to step a), a step a0) in which the road zone is divided according to the road network and sectors of congestion identified from historical traffic data in the road area.

Thus, the road area is divided taking into account the traffic density and the possibilities of bypassing the roads and junctions constituting the road network, so that, for example, there is always an alternative route allowing to avoid a sector in the case where the latter would be associated with the blocking state.

According to a preferred embodiment of the invention, in step b), the centralized controller determines the traffic state of each sector using sensors placed in the road zone. The centralized controller can thus establish whether it is necessary to regulate traffic in certain sectors.

The sensors can for example be electromagnetic loops, speed sensors or tracking vehicles. Thus, the sensors can transmit to the centralized controller the average speed of the vehicles circulating in each sector.

By using the sensors placed in the road area, the centralized controller uses only global traffic data in the road area. No individual data on the vehicles present in the area is necessary. Users do not need to share their personal information like location, speed, route. Thus, the method makes it possible to respect the private data of its users. In this embodiment, the use of personal data is only made by the positioning and navigation aid system fitted to the vehicle. The positioning and navigation aid system does not transmit this data to the centralized controller. The centralized controller only sends general information to the positioning and navigation system, but the positioning and navigation system does not send information to the centralized controller.

According to another embodiment of the invention, the centralized controller is specific to a positioning and navigation aid system supplier and in step b). In this variant, the centralized controller determines the state of the traffic in each sector using information transmitted by the positioning and navigation aid system provided by said supplier of at least some of the vehicles of the plurality of vehicles present in the road area and equipped with said positioning and navigation aid system.

The provider can then use traffic condition information available to them to create the centralized avoidance map. The method can then be implemented using a simple update of the software allowing the operation of the positioning and navigation aid system, so that the centralized avoidance map is established from the data available to the supplier of the positioning and navigation aid system, then distributed to each positioning and navigation aid system of this supplier so that each vehicle equipped with the positioning and navigation aid system of this supplier can draw up a personalized avoidance card.

According to a variant of the invention, the avoidance probability is obtained:
- by means of a reactive controller of the Proportional Integral (PI) type and an "anti-runaway" type process,
- and/or by a predictive control approach;
- and/or by a reactive or predictive control approach including an automated learning step of the parameters on historical traffic data.

According to an advantageous variant of the invention, the sectors are adjacent.

Thus, the sectors cover the entire road area.

According to a variant of the invention, the road zone comprises roads, and the sectors are defined independently of the roads of the road zone.

Thus, the sectors are defined without being constrained by the topography of the traffic lanes of the road zone. They can thus be optimized, that is to say chosen in such a way as to allow alternative routes to exist and so that a sector does not group together all the roads allowing to go from a given first point to a second point. given, in particular taking into account historical traffic data in the road area.

According to a variant of the invention, the time period Δt is between 30 seconds and 10 minutes, preferably between 1 minute and 5 minutes.

Thus, the final route of shorter duration is adjusted regularly, which allows the process to be reactive and adapt to current traffic conditions, but not too often, which makes it possible to avoid modifying the route too often. final route of the vehicle. This time interval Δt therefore allows a good compromise between responsiveness and consistency of the final route of shorter duration.

According to another aspect of the invention, there is proposed a data processing device comprising a centralized controller comprising a memory and allowing the implementation of the method according to the invention.

According to another aspect of the invention, a computer program is proposed comprising instructions which, once they are loaded and implemented on a computer, allow the implementation of the method according to the invention.

According to another aspect of the invention, there is provided a computer-readable recording medium comprising instructions which, when executed by a computer, lead the latter to implement the method according to the invention.

Other characteristics, details and advantages will appear on reading the detailed description below, and on analyzing the appended drawings, in which:

Figure 1 schematically illustrates a road zone Z;

FIG. 2 shows the road zone Z of FIG. 1 divided into a plurality of sectors S1,…,Si,…,Sn;

FIG. 3 shows a diagram illustrating the road traffic management method according to the invention, and in particular the obtaining of a centralized avoidance map from the road zone divided into a plurality of sectors S1,…,Si, …,Sn of FIG. 2, and obtaining a final route of shorter duration from the centralized avoidance map;

FIG. 4 illustrates a final route of shorter duration calculated from a personalized avoidance map;

FIG. 5 shows a graph comparing the average speed of vehicles in the road zone over time when the method according to the invention is applied and when the method according to the invention is not applied;

FIG. 6 is an example of an algorithm that can be used to calculate the avoidance probability of each sector of the centralized avoidance map presented in FIG. 3;

Figure 7 shows a curve based on real observations and representing the sum of the distances traveled by the vehicles present in a sector per unit of time as a function of the number of vehicles present in this sector.

detailed description

Process

In the figures, the same references designate identical or similar elements.

FIG. 1 represents a road zone Z. This road zone Z comprises a road network. This road network comprises a plurality of roads 10 and a plurality of crossroads 11 marking the intersections between the roads 10. This road zone Z illustrates in particular an urban zone with a dense road network.

In this road zone Z a plurality of vehicles 12, 121 circulate. Among the plurality of vehicles 12, 121 circulating in the road zone Z, certain vehicles 121 are equipped with a positioning and navigation aid system 13. Each positioning and navigation aid system 13 makes it possible to position the vehicle 121 which it equips in the road zone Z. Thus, at each instant, the positioning and navigation aid system 13 knows the position of the vehicle 121 that he equips. Each positioning and navigation aid system 13 also makes it possible to draw up a route of shorter duration from an initial position P1 to a final position P2 in the road zone Z.

The road zone Z also comprises a plurality of sensors 14 making it possible to determine the state of the traffic in the road zone Z. In particular, the sensors 14 are speed sensors making it possible to evaluate the speed of the vehicles 12, 121 circulating in the road zone Z. They can, for example, be electromagnetic loops or radars.

A tracker vehicle 15 is also circulating in road zone Z. In FIG. 1, a single tracker vehicle 15 is illustrated, it being understood that there may be a plurality of tracker vehicles 15 in road zone Z. also to measure the speed of vehicles traveling in the road zone Z.

In order to be able to manage the road traffic in the road zone Z, the road zone Z is divided, during a step a0), into a plurality of sectors S ₁ ,…,S _i ,…,S _n as illustrated in figure 2. Sectors S ₁ ,…,S _i ,…,S _n are determined according to the road network and congestion sectors identified from historical traffic data in road zone Z.

In particular, the division of the road zone Z can be done experimentally.

For example, to define the sectors S ₁ ,…,S _i ,…,S _n , the roads 10 and the crossroads 11 which are the most used by the vehicles 12, 121 are identified from the historical traffic data in the area road Z. These roads 10 and these crossroads 11 are usual traffic congestion sectors. Thus, when they are present in a sector S _i , it must be ensured that it is possible to find another route for the vehicles 121 which would use them not passing through this sector S _i , or in other words passing only through the other sectors S ₁ ,…,S _{i -1} ,S _i+1 ,…,S _n . And, it is also necessary that this other route does not generate an excessive detour for these vehicles 121. Thus, for a given road network, the main rules which govern the definition of the sectors S ₁ ,…,S _i ,…,S _n are to distribute the most congested intersections and the main roads (urban boulevards, expressways, etc.) between the different sectors. This ensures that when an area needs to be avoided, other options are available. If a sector contains too many critical crossroads or main roads, it is sufficient to divide it into several sectors.

In another example, if the road zone Z is crossed by a watercourse and there are only two bridges in the road zone Z which allow to pass from one bank of the watercourse to the other bank of the watercourse, then it must be ensured that the two bridges are not in the same sector S _i . Because if the two bridges are in the same sector S _i then all the vehicles wishing to pass from one bank to the other will necessarily pass through this sector S _i , and no alternative route not passing through this sector S _i can be found . It will then be impossible to regulate the traffic in this sector S _i . This sector S _i then runs the risk of being regularly congested.

In FIG. 2, the sectors S ₁ ,…,S _i ,…,S _n are adjacent. The whole of the road zone Z is covered by the sectors S ₁ ,…,S _i ,…,S _n . Thus, traffic on all the roads 10 and intersections 12 constituting the road zone Z can be controlled. The sectors S ₁ ,…,S _i ,…,S _n illustrated in figure 2 are delimited by parallel horizontal lines and parallel vertical lines, but the sectors S ₁ ,…,S _i ,…,S _n could be defined by any other straight line or curve defining adjacent sectors.

In a variant not shown, the sectors S ₁ ,…,S _i ,…,S _n could also be separate. Thus, part of the road zone Z would then not be covered by any sector. This part of road zone Z would then not be regulated. This could, for example, be the case for a road, a crossroads or a bridge that it would not be possible to circumvent without causing an excessive detour for the vehicles.

The sectors S ₁ ,…,S _i ,…,S _n are moreover defined independently of the roads 10 of the road zone Z. In the example illustrated in FIG. 2, the peripheral edges of the sectors S ₁ ,…,S _i ,…,S _n do not overlap with any of the roads 10 of the road zone Z. Alternatively, provision could be made for certain ends of certain sectors S ₁ ,…,S _i ,…,S _n to overlap with roads 10 of the road zone Z. It could also be provided in another example that the ends of the sectors S ₁ ,…,S _i ,…,S _n are superimposed on the roads 10 of the road zone Z. Like the sectors S ₁ ,…, If _i ,…,S _n are defined independently of the roads 10 of the road zone Z, the sectors S ₁ ,…,S _i ,…,S _n are defined without being constrained by the roads 10 of the road zone Z. Thus, the shape given to sectors S ₁ ,…,S _i ,…,S _n is free.

Figure 3 illustrates a computer-implemented method of road traffic management of road zone Z. This method implements a centralized controller 20 comprising a memory 21 for recording and storing information.

In this method, the road zone Z divided into a plurality of sectors S ₁ ,…,S _i ,…,S _n , as illustrated in FIG. 2 is stored in the memory 21 of the centralized controller 20 during a step a). An average reference speed V _{S i,ref} defined for each sector S ₁ ,…,S _i ,…,S _n is also stored in the memory 21 of the centralized controller 20.

The average reference speed V _{S i,ref} corresponds to the average speed reached by the vehicles traveling in the sector S _i when a critical number of vehicles N _c are traveling in the sector S _i . The average reference speed V _Si,ref and the critical vehicle number N _c are shown in Figure 7.

FIG. 7 represents a curve in a frame, showing the evolution of the sum of the distances traveled by the vehicles present in the sector S _i per unit of time as a function of the number of vehicles circulating in the sector S _i . This curve was obtained from experimental observations. The curve reaches a maximum value at a point M. The ordinate of this point M gives the maximum value D of the sum of the distances traveled by the vehicles present in the sector S _i per unit time. The abscissa associated with point M corresponds to the critical number of vehicles N _c . In the example represented in FIG. 7, the maximum value D is equal to 15000 meters traveled by all the vehicles circulating in the sector S _i per second, and the critical number of vehicles N _c is equal to 1000 vehicles.

Thus, the critical number of vehicles N_vscorresponds to the number of vehicles circulating in the sector S_I when the sum of the distances traveled by all the vehicles present in the sector per unit of time reaches a maximum value D.

When a larger number than the critical number of vehicles N _c circulates in the sector, this forms congestion and the average speed of the vehicles circulating in the sector decreases. In FIG. 7, when there are more than 1000 vehicles circulating in the sector S _i , it is actually observed that the value of the sum of the distances traveled by the vehicles present in the sector S _i decreases.

When a smaller number than the critical number of vehicles N _c circulates in the sector, the average speed of the vehicles circulating in the sector increases but the sum of the distances traveled by all the vehicles present in the sector is smaller. In FIG. 7, when there are fewer than 1000 vehicles circulating in the sector S _i , it is actually observed that the value of the sum of the distances traveled by the vehicles present in the sector S _i decreases.

Thus, the critical number of vehicles N _c corresponds to the number of vehicles allowing optimal operation of the sector S _i , that is to say that it is the number of vehicles which makes it possible to maximize the sum of the distances traveled by the all the vehicles circulating in the sector S _i per unit of time.

The average reference speed V _Si,ref associated with the circulation of the critical number of vehicles N _c in the sector S _i corresponds to the average speed reached in the sector S _{i ,} when its operation is optimal. The value of the average reference speed V _Si,ref corresponds to the value of the slope of the straight line connecting the origin of the frame in which the curve is represented to the point M. The value of the average reference speed V _Si,ref is therefore equal to the maximum value D divided by the critical number of vehicles N _c . It is equal to 15 ms ^-1 in the example illustrated in figure 7.

Then, during a step b) the centralized controller 20 determines the traffic state for each sector S ₁ ,…,S _i ,…,S _n at time t. For this, the centralized controller 20 recovers the average speed V _{S i,t} of each sector S _i during a period of time Δt preceding the instant t. The average speed V _{S i,t} of each sector S _i corresponds to the average speed reached by the vehicles traveling in the sector S _i during the time period Δt preceding the instant t.

In a preferred embodiment of the invention, the centralized controller 20 is external to the positioning and navigation aid systems 13 equipping the vehicles 121 in the road zone Z. The centralized controller 20 is then common to the set of vehicles 121 present in road zone Z and equipped with a positioning and navigation aid system 13.

The centralized controller 20 retrieves the average speed V _Si,t of each sector S _i using sensors 14 placed on the road zone Z and/or using tracer vehicles 15 circulating in the road zone Z.

By using the sensors 14 arranged in the road zone Z, the centralized controller 20 uses only global data on traffic in the road zone Z. No individual data on the vehicles 12, 121 present in the sectors S ₁ ,..., If _i ,…,S _n is not necessary. Users don't need to share their personal information like location, speed, route. Thus, the method makes it possible to respect the private data of its users. The use of personal data is only made by the positioning and navigation aid system 13 which equips the vehicle 121 and never goes back to the centralized controller 20. The centralized controller 20 only sends general information to the positioning and navigation aid system 13 but the positioning and navigation aid system 13 does not send information to the centralized controller 20.

In another mode of implementation of the invention, the centralized controller 20 is specific to a supplier of positioning and navigation aid system 13, and the centralized controller 20 determines the state of the traffic of each sector S ₁ ,…,S _i ,…,S _{n ,} using information transmitted by the positioning and navigation aid system provided by said supplier of at least some of the vehicles 121 of the plurality of vehicles present in the road zone Z and equipped with said positioning and navigation aid system 13.

The centralized controller 20 can then obtain information on the average speed V _Si,t of each sector S _i only using the information transmitted by the positioning and navigation aid systems 13 provided by the supplier or it can obtaining information on the average speed V _Si,t of each sector S _i both with the aid of the sensors 14 arranged in the road zone Z and with the aid of the information transmitted by the positioning and driving assistance systems navigation 13 provided by the supplier.

In this case where the centralized controller 20 is specific to a positioning and navigation aid system supplier 13, the personal information of the users such as their position, their speed, their route are sent to the centralized controller 20. However, the The use of this information is preferably limited to the supplier of the positioning and navigation aid system 13.

From the average reference speed V _{S i,ref} and the average speed V _{S i,t} for each sector S _i , the centralized controller 20 calculates an avoidance probability P _{S i,t} for each sector S _i at during a step c). This avoidance probability P _Si,t corresponds to the fraction of vehicles having to avoid the sector S _i during the time period Δt following the instant t. The purpose of calculating the avoidance probability P _Si,t is to make it possible to improve the traffic conditions in the road zone Z, by limiting the formation of congestion. By limiting congestion, the average speed V _{S i,t} of each sector S _i is thus increased. The calculation of the avoidance probability P _Si,t thus makes it possible to smooth traffic in the road zone Z.

The avoidance probability P _{S i,t} can be calculated in different ways. The centralized controller 20 can for example comprise a reactive controller of Proportional Integral (PI) type using an “anti-runaway” (or “anti-Windup”) process. The Proportional Integral (PI) type reactive controller using an “anti-runaway” process can use the algorithm given in FIG. 6 to calculate the avoidance probability P _{S i,t} .

The algorithm of figure 6 uses several constants. ε is the minimum value of the observed error for which a traffic control action is triggered. It is for example equal to 0.1. k _p is the proportional gain of the proportional integral reactive controller. It is for example equal to 0.1. k _i is the integral gain of the proportional integral reactive controller. It is, for example, equal to 1/360. p _min is the minimum value that the avoidance probability P _{S i,t} can take. It is for example equal to 0. p _max is the maximum value that the avoidance probability P _{S i,t} can take. It is preferably chosen as being strictly less than 1, and it is for example equal to 0.95. Δt is the time period in seconds after which the algorithm is executed again. It is for example equal to 180.

The algorithm of figure 6 also uses several variables: V _{S i,ref} is the average reference speed for each sector S _i , V _{S i,t} is the average speed at time t for each sector S _i , P _{S i,t} is the avoidance probability of zone S _i at time t, integral _i is the value of the integral term of the reactive controller for each zone S _i and int_shut _i is a boolean that defines whether the integral term _i should be updated or not. It is this term that enables the “anti-runaway” process.

At the start of the algorithm, the average reference value V _{S i,ref} for each sector S _i is estimated. As explained above, this average reference value V _{S i,ref} is predefined and corresponds to the average speed reached by the vehicles traveling in the sector S _i when a critical number of vehicles N _c are traveling in the sector S _i . This average reference value V _{S i,ref} is stored in the memory 21 of the centralized controller 20.

During the time period Δt, for each sector S _i , the average speed V _{S i,t} over the time period Δt preceding the instant t is estimated. This estimation can be made using sensors on the road or using tracer vehicles for example. The difference between the average reference speed V _{S i,ref} and the average speed V _{S i,t} at time t is calculated. This difference corresponds to error . If this error difference is positive, that is to say if the average reference speed V _{S i,ref} is greater than the average speed V _{S i,t} at time t, then this means that it is in our interest to restrict the entry of vehicles into this sector S _i during the time period Δt.

Depending on the value of the calculated error, the integral integral term _i can be updated. Thus, when the absolute value of the error is greater than the minimum value of the error ε, and the boolean int_shut _i has the value "false", then the integral term _i is updated and receives the value of int_shut _i to which is added the value of the error multiplied by the time period Δt. It will be noted that this updating of the value of the integral term _i is not carried out when the error is small, that is to say when the error is less than the minimum value of the error ε, and when the anti-runaway process is activated, i.e. when the boolean int_shut _i has the value “true”.

Once the integral integral term _i has been updated or not, the output term receives the value recommended by the centralized controller 20 for the avoidance probability. This recommended value is the sum of the multiplication between the proportional gain k _p and the error and of the multiplication between the integral gain k _i and the integral integral term _i .

It is then verified that the recommended value output by the reactive controller for the avoidance probability is between the minimum value p _min and the maximum value p _max defined as constants. If the output recommended value is not in this interval, then the anti-runaway process is set up, and the recommended value is brought back to the minimum value p _min if it was lower than p _min or to the maximum value p _max if it were greater than p _max . The value thus obtained is the value of the avoidance probability P _{S i,t} . As p _max is strictly less than 1, access to a sector S _i is never completely prevented.

The boolean int_shut _i is reset to "false".

Then the anti-runaway process is executed. If the output recommended value has been changed to match the maximum value p _max or the minimum value p _min , then the integral term _i is no longer updated. The boolean int_shut _i evaluates to "true", which prevents further updating of the integral term integral _i . Moreover, the integral integral term _i is fixed at the last known value.

The value of the avoidance probability P _{S i,t} is then applied to the corresponding sector S _i . The algorithm is then repeated for the next time period Δt.

According to another example, the avoidance probability P _{S i,t} could be obtained by a predictive control approach (“Model Predictive Control”). It could also be obtained by a reactive or predictive control approach including an automated learning step of the parameters on historical traffic data, for example by using Markov processes.

Once the avoidance probability P _{S i,t} has been calculated, the centralized controller 20 establishes a centralized avoidance map 30 during a step d). On this centralized avoidance map 30, each sector S _i is associated with the corresponding avoidance probability P _{S i,t} . Thus, the centralized avoidance map 30 contains information on the state of the traffic in the sectors S ₁ ,…,S _i ,…,S _n of the road zone Z.

During a step e), the centralized avoidance map 30 is broadcast to the vehicles 121 present in the road zone Z and equipped with a positioning and navigation aid system 13. The centralized avoidance map 30 is broadcast during the time period Δt following the calculation of the avoidance probability P _{S i,t} . The positioning and navigation aid system 13 of each of the vehicles 121 present in the road zone Z then receives the centralized avoidance map 30.

During a step f), for each of the vehicles 121 present in the road zone Z and equipped with a positioning and navigation aid system 13 which is activated to work out a route of shorter duration of a initial position P1 to a final position P2 in the road zone Z, the positioning and navigation aid system 13 performs a sub-step (f1) consisting in selecting at least one sector S _j from among the set of sectors S ₁ ,…,S _i ,…,S _n dividing the road zone Z.

In a first variant, the sectors S _j can, for example, be selected when they are located at a distance less than a predetermined distance from the position of the vehicle. The set of sectors S ₁ ,…,S _i ,…,S _n dividing the road zone Z can also all be selected.

In a second variant illustrated in FIG. 4, the positioning and navigation aid system 13 calculates an initial route of shorter duration 31 between the initial position P1 and the final position P2 without condition of avoiding sectors S ₁ ,…,S _i ,…,S _n . In other words, this calculation is carried out independently of the state of traffic in the road zone Z. The initial route of shortest duration 31 corresponds to the shortest route in time to go from the initial position P1 to the final position P2. The positioning and navigation aid system 13 then selects from among the sectors S ₁ ,…,S _i ,…,S _n , the sectors S _j crossed by this initial route of shorter duration 31, or even selects the sectors S _j traversed by this initial route of shorter duration 31 and associated with a non-zero avoidance probability P _{S i,t} . The positioning and navigation aid system 13 then selects the sectors S _j likely to be crossed by the vehicles and access to which must be limited to guarantee good circulation of the sector S _j . The number of calculations performed by the positioning and navigation aid system 13 is thus limited.

During a sub-step (f2), the positioning and navigation aid system 13 performs a random draw by following a Bernoulli test from the avoidance probability P _{S i,t} for each sector S _j selected. This random draw results in two possible values: 0 or 1. Value 0 is associated with a passing state 43 and value 1 is associated with a blocking state 42.

During a sub-step (f3), the positioning and navigation aid system 13 then transcribes the centralized avoidance map 30 into a personalized avoidance map 40 on which the blocking state 42 or l The on state 43 of each selected sector S _j is associated with the corresponding sector S _j . Thus, each vehicle 12 present in the road zone Z and equipped with a positioning and navigation aid system 13 then has a personalized avoidance map 40 which is specific to it since the passing 43 or blocking 42 states of each selected sectors S _j depend on the random draw made by the positioning and navigation aid system 13 of each vehicle 121.

The positioning and navigation aid system 13 then calculates a final route of shorter duration 41 making it possible to go from the initial position P1 to the final position P2 according to the personalized avoidance map 41, during a sub-step (f4). This final route of shorter duration 41 corresponds to the shortest route in time allowing to go from the initial position P1 to the final position P2 while avoiding the selected sectors S _j associated with the blocking state 42 on the map d custom avoidance 40.

As can be seen in FIGS. 3 and 4, the final route of shorter duration 41 does not pass through the selected sectors S _j associated with the blocking state. In particular, in FIG. 4, since the initial route of shorter duration 31 passes through sectors associated with the blocking state 42, the positioning and navigation aid system has modified this initial route of shorter duration 31 to propose the final route of shorter duration 41 which does not pass through the sectors associated with the blocking state 42. Traffic in these sectors associated with the blocking state 42 is thus limited. It can be noted that the final route of shorter duration 41 passes through the selected sector S _j associated with the on state 43.

Furthermore, as the blocking 42 or passing 43 state of each sector is calculated from the avoidance probability P _{S i,t} , the quantity of vehicles 121 for which the sector S _i is associated with the passing state, and the quantity of vehicles 121 for which the sector S _i is associated with the blocking state are adapted to this avoidance probability P _{S i,t} . Thus, the number of vehicles able to enter the sector S _i is controlled so that the number of vehicles in the sector S _i exceeds the number of critical vehicles as little as possible. Traffic in road zone Z can therefore be regulated.

The regulation is made to reach a collective optimum, that is to say to minimize the time lost by all the vehicles circulating on the road network. Thus, on average, the vehicles 12 circulating in the road zone Z have the shortest route in time. It is possible that some vehicles have a slightly longer journey time than if the positioning and navigation aid system did not take account of the centralized avoidance card 30, that is to say if the positioning and navigation aid system 13 sought to achieve the individual optimum. However, this collective management of road traffic, based on the collective management of the routes taken by the vehicles present in the road zone, allows a majority of the vehicles present in the road zone to have a shorter journey time.

This collective traffic management makes it possible to limit congestion. Vehicles traveling in road zone Z can thus have a higher average speed, and therefore a shorter journey time.

The difference in average speed in the road zone Z over time when the method according to the invention is used and when the method according to the invention is not used is presented in the graph of Figure 5.

On this graph, the ordinate represents the average speed in meters per second of vehicles traveling in road zone Z and the abscissa represents the time in hours.

The curve 45 represents the average speed of the vehicles 12, 121 traveling in the road zone Z when the method according to the invention is not used, that is to say when the traffic is not regulated. Note that between 2 am and 4 am, the average speed decreases sharply and notably falls below 5 ms ^-1 .

The curve 46 represents the average speed of the vehicles traveling in the road zone Z when the method according to the invention is used, that is to say when the traffic is regulated. Note that the speed is generally maintained between 5 ms ^-1 and 8 ms ^-1 , and in particular, between 2 a.m. and 4 a.m., the average speed of vehicles traveling in the road zone is higher than when the traffic is not regulated.

The graph clearly shows the efficiency of the traffic management process. The method makes it possible to limit the congestion zones in the road zone Z, and therefore to maintain an average speed of the vehicles traveling in the road zone higher than in the absence of regulation.

Device

The method according to the invention can be implemented by a data processing device comprising a centralized controller 20.

computer program

The invention also relates to a computer program comprising instructions which, once they are loaded and implemented on a computer, allow the implementation of the method according to the invention.

Recording medium

The invention also relates to a computer-readable recording medium comprising instructions which, when executed by a computer, lead the latter to implement the method according to the invention.

Claims (14)

A computer-implemented method of road traffic management of a road zone (Z), preferably an urban zone, implementing at least one centralized controller (20) comprising a memory (21), the method comprising the following steps :
a) the road zone (Z), divided into a plurality of sectors (S ₁ ,…,S _i ,…,S _n ), is stored in the memory (21) of the centralized controller (20),
b) the centralized controller (20) determines a traffic state for each sector (S _i ) at a time t,
c) for each sector (S _i ), the centralized controller (20) calculates an avoidance probability (P _{S i,t} ), depending on the traffic state determined in step b) and corresponding to the fraction of vehicles having to avoid said sector (S _i ), during a period of time Δt following time t, to smooth traffic in the road zone (Z),
d) the centralized controller (20) establishes at least one centralized avoidance map (30) associating each of the avoidance probabilities (P _{S i,t} ) calculated with the corresponding sector (S _i ),
e) the centralized controller (20) broadcasts the centralized avoidance map (30) established in step d) to a plurality of vehicles (121) present in the road zone (Z) and each equipped with a positioning system and navigation aid (13), during the period Δt following the instant t,
f) for each of the vehicles (121) of the plurality of vehicles whose positioning and navigation aid system (13) is activated to work out at least one route of shorter duration from an initial position (P1) to 'at a final position (P2) in the road zone (Z):
- (f1) the positioning and navigation aid system (13) selects at least one sector (S _j ) from among the sectors (S _i ),
- (f2) for each of the sectors (S _j ) selected, the positioning and navigation aid system (13) performs a random draw, preferably following a Bernoulli test, from the avoidance probability ( P _{S i,t} ) of the selected sector, making it possible to define an on state (43) or a blocking state (42) for each of the sectors (S _j ) selected,
- (f3) the positioning and navigation aid system (13) transcribes the centralized avoidance map (30) into a personalized avoidance map (40), by associating the on state (43) or the 'blocking state (42) defined for each of said sectors (S _j ) selected to the corresponding sector (S _i ),
- (f4) the positioning and navigation aid system (13) calculates a final route of shorter duration (41) from the initial position (P1) to the final position (P2), depending on the personalized avoidance map (40), avoiding the selected sectors (S _j ) associated with the blocking state (42),
g) once the time period Δt has elapsed, steps b) to f) are repeated. Method according to claim 1, wherein step (f1) comprises the following sub-steps:
- the positioning and navigation aid system (13) calculates an initial route of shorter duration (31) from the initial position (P1) to the final position (P2) without sector avoidance conditions,
- the positioning and navigation aid system (13) selects the sector or sectors (S _j ) crossed by the initial route of shorter duration (31) on the avoidance map (30) and associated with a probability of avoidance (P _{S i,t} ) not zero. Process according to claim 1 or 2, in which:
- in step a), an average reference speed (V _{S i,ref} ) defined for each sector (S _i ) and corresponding to the optimal functioning of the sector (S _i ) is stored in the memory (21) of the centralized controller (20),
- in step b), the centralized controller (20) retrieves an average speed (V _{S i,t} ) of each sector (S _i ) at time t, and
- in step c), for each sector (S _i ), the centralized controller (20) calculates the avoidance probability (P _{S i,t} ) as a function of the average reference speed (V _{S i,ref} ) and the average speed (V _{S i,t} ). Method according to claim 3, in which the average reference speed (V _Si,ref ), for each sector (S _i ), is the average speed reached in the sector (S _i ) when a critical number of vehicles (N _c ) circulates in the sector (S _i ), the critical number of vehicles (N _c ) being the number of vehicles circulating in the sector (S _i ) when the sum of the distances traveled by all the vehicles present in the sector (S _i ) per unit time reaches a maximum value. Method according to one of Claims 1 to 4, the road zone Z comprising a road network, comprising, prior to step a), a step a0) in which the road zone Z is divided according to the road network and sectors identified from historical traffic data in road zone Z. Method according to one of Claims 1 to 5, in which in step b), the centralized controller (20) determines the traffic state of each sector (S ₁ ,…,S _i ,…,S _n ) at using sensors (14) arranged in the road zone Z. Method according to one of Claims 1 to 6, in which the centralized controller (20) is specific to a positioning and navigation aid system supplier (13) and in step b), the centralized controller ( 20) determines the traffic status of each sector (S ₁ ,…,S _i ,…,S _n ) using information transmitted by the positioning and navigation aid system (13) supplied by said supplier of at least some of the vehicles of the plurality of vehicles (121) present in the road zone (Z) and equipped with said positioning and navigation aid system (13). Method according to one of Claims 1 to 7, in which the avoidance probability (P _Si,t ) is obtained:
- by means of a reactive controller of the Proportional Integral (PI) type and an "anti-runaway" type process,
- and/or by a predictive control approach;
- and/or by a reactive or predictive control approach including an automated learning step of the parameters on historical traffic data. Method according to one of Claims 1 to 8, in which the sectors (S ₁ ,…,S _i ,…,S _n ) are adjacent. Method according to one of Claims 1 to 9, the road zone (Z) comprising roads (10), in which the sectors (S ₁ ,…,S _i ,…,S _n ) are defined independently of the roads of the zone road. Method according to one of Claims 1 to 10, in which the time period Δt is between 30 seconds and 10 minutes, preferably between 1 minute and 5 minutes. Data processing device comprising a centralized controller (20) comprising a memory (21) and allowing the implementation of the method according to one of Claims 1 to 11. Computer program comprising instructions which, once they are loaded and implemented on a computer, allow the implementation of the method according to at least one of Claims 1 to 11. Computer-readable recording medium comprising instructions which, when executed by a computer, lead the latter to implement the method according to at least one of Claims 1 to 11.