Server, system and method for determining the position of a traffic jam end

Abstract

Server (100) for determining a position (x2) of a traffic jam end, comprising: - a computing unit (10); - a memory (20); - a receiving unit (30) for receiving a plurality of measurement data (80), each with at least one position data (x) of a vehicle (71); wherein the server (100) is configured to calculate the position (x2) of the traffic jam end using at least one sigmoid function and the received measurement data (80).

Description

[0001] The invention relates to a server, a system and a method for determining the position of a traffic jam end.

[0002] A server for determining the location of a traffic jam is known from the prior art. Google Traffic can determine the geographical area where a traffic jam has occurred. The server does this by, for example, evaluating the speed of smartphones in vehicles. However, this traffic jam data only indicates the location of the jam with too great an inaccuracy. The type of traffic jam, its future development, or its most recent dynamic evolution cannot be determined using Google Traffic's approach.

[0003] Another way to determine the location of a traffic jam is to install stationary sensors, such as cameras or induction loops, along a relevant section of road. These stationary sensors evaluate the traffic conditions, particularly the traffic flow and density. They measure the speed and distances between individual vehicles and calculate the traffic flow and density from this data.

[0004] One disadvantage of this type of location tracking is that the calculation of the traffic jam's end position can only be performed in the section of road where the sensors are installed. Installing stationary sensors, such as cameras or induction loops, is very expensive and therefore not used across the board.

[0005] EP 1 235 195 A2 describes a method for determining traffic jam data. In this method, a first vehicle transmits its current position, linked to a time, to a central control unit. The central control unit stores this information in a database and uses it to determine a route for the first vehicle. Using the routes of other vehicles in the immediate vicinity, which are also stored in the database, the central control unit generates a route forecast for the first vehicle. This forecast indicates how the speed of the first vehicle is likely to change on the upcoming section of the route. However, this approach cannot determine the exact location, behavior of the traffic jam, or the position of its end.

[0006] US Patent 2007 / 0005231 A1 describes a system and a method for determining the position of the end of a traffic jam. A vehicle belonging to the system includes a controller that analyzes the vehicle's speed. As the vehicle, traveling at a constant speed, approaches the end of the traffic jam and thus reduces its speed, the controller determines the location of the end of the jam at the point where the vehicle's speed is approximately zero or remains constant at a very low speed. This has the disadvantage that the position of the end of the traffic jam can only be determined very imprecisely. For example, if the end of the traffic jam is very gradual, where the speed decreases steadily but does not reach zero or a very low and constant level, this system and method cannot provide an exact position for determining the location of the end of the traffic jam.

[0007] Based on this state of the art, the task is to provide a server, a system, and a method that addresses the aforementioned disadvantages. In particular, a server should be created that can determine the exact position of the end of a traffic jam and, if applicable, its development, comprehensively and regardless of location. This should also enable precise position determination of the end of a traffic jam, even in cases of "soft" traffic jams, where vehicles gradually enter at progressively slower speeds. A further task is to provide a server for determining the position of the end of a traffic jam that is capable of identifying the type of traffic jam. Is it a "hard" traffic jam, requiring sharp braking upon entering, or a "soft" traffic jam, where a gradual reduction in speed can be assumed?

[0008] This task is solved by a server according to claim 1, a system according to claim 7 and a method according to claim 13.

[0009] In particular, the task is solved by a server for determining the position of a traffic jam end, which includes – a computer unit; – a storage facility; – a receiving unit for receiving a large number of measurement data, each with at least one position data entry for a vehicle.

[0010] The server is preferably designed to calculate the position of the end of the traffic jam using at least one sigmoid function and the received measurement data.

[0011] The at least one sigmoid function sig(x) used to locate and characterize the position of the end of the traffic jam can, for example, have the following formula:

[0012] As shown, this can be defined using four parameters [a1, a2, a3, a4].

[0013] The sigmoid functions can be determined, for example in a first iteration cycle, by randomly selecting parameter values. The measurement data can then be used to select at least one sigmoid function that accurately models the actual course of the traffic jam and thus also its end point. The selected, and therefore high-quality, sigmoid function can then be used to calculate the position of the end point of the traffic jam.

[0014] Alternatively or additionally, parameter values ​​of the sigmoid functions can be determined or calculated based on at least some of the measurement data. The measurement data is transmitted, for example, from a vehicle to the server via a radio network, preferably a mobile network, which then stores it in its memory. The measurement data can include a vehicle's position data, which the processing unit uses to calculate the vehicle's speed based on the transmission time of the position data.

[0015] An advantage of the server according to the invention is that it can determine the position of the end of the traffic jam using the sigmoid function, regardless of the position from which the vehicle transmitted its measurement data to the server. The sending vehicle may be located before or shortly after the end of the traffic jam. The sigmoid function is therefore suitable for making statements about the end of the traffic jam even with measurement data from any position, for example, within the traffic jam itself.

[0016] Another advantage is that the shape of the sigmoid function can be used to determine how the vehicle's speed changes over time. This can be used to characterize the end of a traffic jam. If the sigmoid function shows a rapid and sharp drop in speed, it indicates a hard end to the jam, where vehicles traveling freely encounter a cluster of stationary vehicles. Conversely, if the sigmoid function shows a slow and gradual drop in speed, it suggests that the vehicle will enter the jam with surrounding vehicles at a gradual reduction in speed, resulting in a soft end to the jam. The server can be configured to transmit this information to participants, such as vehicles that have subscribed to the service. The server can also use this information to assess the danger level of the traffic jam.For example, several hazard categories (e.g., high, low, minor) can be defined, with the server classifying each traffic jam into one of these categories.

[0017] Preferably the measurement data are data tuples and include: – Traffic information data; and / or – Speed ​​data that specifies at least one speed of the respective vehicle; and / or – Distance data that specifies at least a distance between the respective vehicle and a vehicle in front of it; and / or – Braking frequency data, which indicates the braking frequency of the respective vehicle.

[0018] This allows the server to use the measurement data to determine the vehicle's surroundings, such as traffic density. For example, by transmitting the distance to a vehicle ahead, the server can calculate the traffic density. Traffic density can be calculated using the following formula: ρ = 1 r + s

[0019] Traffic density ρ depends on the distance r between two vehicles and the length s of the vehicle behind. The same calculation can also be performed using traffic information data, which, for example, provides information on the number of vehicles in the immediate vicinity of the vehicle, lane-changing behavior, or other traffic-related data. Similarly, traffic density can also be calculated based on the vehicle's braking frequency. It is also possible to determine traffic density based on the entire detected area surrounding the measuring vehicle. Thus, it is also possible to calculate the position of the end of a traffic jam by observing changes in the traffic density profile.

[0020] A further advantage of the invention lies in the fact that the sigmoid function can be used to continuously model traffic density over the road or route, preferably also over time. This makes it possible to determine the position of the end of a traffic jam with just a few measurement points – regardless of where the measurement points or data were recorded. It is also possible to determine the type of traffic jam. If the traffic density increases rapidly and sharply, it is a hard traffic jam. If the sigmoid function shows a slow and gradual increase, it is a soft traffic jam. The server can be configured to transmit relevant information to participants, such as vehicles that have subscribed to this service. The vehicles can process this information and use it to issue warning signals to the driver or other road users.Furthermore, this information can be used to influence the functioning of a driver assistance system. The driver assistance system may then reduce the vehicle's speed.

[0021] The measurement data could advantageously include hazard warning light data. This data could indicate the use of the vehicle's hazard warning lights and / or the use of hazard warning lights by a vehicle in the immediate vicinity, detected by a sensor and / or camera. This hazard warning light data could contribute to a more precise determination of the position and / or characteristics of the end of the traffic jam.

[0022] In a further embodiment of the invention, the server is configured to determine a plurality of parameter sets to calculate the position of the traffic jam end. Each parameter set defines a first sigmoid function and a second sigmoid function. The first sigmoid function of the parameter set models a speed profile, and the second sigmoid function of the parameter set models a traffic density profile. A parameter set can be defined by eight parameters [v1, v2, v3, v4, ρ1, ρ2, ρ3, ρ4], wherein four parameters [v1, v2, v3, v4] represent the speed profile and four parameters [ρ1, ρ2, ρ3, ρ4] represent the traffic density profile. The plurality of parameter sets can preferably be greater than 10, more preferably greater than 100 or greater than 1000.By defining two sigmoid functions in a parameter set, where the first sigmoid function represents the vehicle's speed profile as a function of location and the second sigmoid function represents the traffic density profile as a function of location, the advantages of both speed and traffic density profiles are combined.

[0023] Another advantage is that lane-specific localization allows for the precise determination of the traffic jam's end characteristics. Speed ​​profiles and traffic density profiles can lead to different results regarding the traffic jam's end. The server can output the position determined by the modeled speed profile or the position determined by the modeled traffic density profile. It is also possible to calculate the final position of the traffic jam's end by determining an average between the two calculated positions, for example, the position exactly between the two determined positions. The parameter sets can be used to make precise statements about the traffic jam's end characteristics. When determining the type of traffic jam end, both changes in speed and traffic density are taken into account, allowing for more accurate predictions.Advantageously, the sigmoid function can also be used to model acceleration and deceleration profiles, and analogously, the acceleration profile can be determined as a function of position to pinpoint the location of the traffic jam's end. By modeling the acceleration profile, it is also possible to precisely calculate the position of the traffic jam's end and to determine its characteristics.

[0024] In an advantageous embodiment, the server can include an evaluation unit. The evaluation unit assesses the quality of at least a selection of a plurality of sigmoid functions with different parameters calculated by the server, using at least the measurement data. The measurement data are compared with the calculated sigmoid functions. The closer the curve of the sigmoid function is to the value of the measurement data, the higher the quality of the sigmoid function and the better it is rated. Such an evaluation of the sigmoid functions can be achieved, for example, by classifying the sigmoid functions into a system of 10 classes, with class 10 containing the highest quality sigmoid functions.By selecting preferred sigmoid functions, for example, higher-class sigmoid functions such as classes 9 and 10, or the top 5, especially the top 50 or 500, determining the position of the congestion end can be simplified and the number of correctly detected positions increased. The evaluation of the sigmoid functions based on the measurement data can also be achieved, for example, by determining the least-squares fitting residual between the sigmoid function and the measurement data and evaluating the respective sigmoid functions based on the size of this residual.

[0025] In one embodiment, the evaluation unit for calculating the sigmoid functions can utilize a particle filter and / or a support vector machine (SVM) and / or linear discriminant analysis (LDA). The particle filter generates continuous updates of the sigmoid functions based on new measurement data. Here, the particle filter approximates the a posterior distribution of the state probabilities of the sigmoid functions using a finite set of parameters [v1, v2, v3, v4, ρ1, ρ2, ρ3, ρ4]. The probability density function is approximated over the sigmoid functions using a sample set, the particles. Unlike alternative approaches, particle filters, due to their non-parametric form, can approximate arbitrary distributions. Similarly, a calculated velocity profile and / or traffic density profile can be interpolated using a polynomial of degree m, where m = 3 is an advantageous choice.The coefficients of this polynomial, together with other properties of the signal, such as the gradient of velocity over time or the gradient of traffic density over time, can be interpreted as a point in an n-dimensional hyperspace. A previously trained SVM or LDA is then able to assess how well the calculated sigmoid functions correspond to the vehicle's measurement data. The advantages lie in the fast and reliable evaluation as well as the compact representation of the evaluation rules.

[0026] In a further embodiment of the invention, the server can receive traffic jam data from another server. This data indicates an area where a traffic jam has occurred. Using this data, the sigmoid function is calculated. This calculation can be performed by the server making a parameter preselection based on the traffic jam data. This allows the calculation of the sigmoid functions to be controlled effectively in advance. The traffic jam data enables a parameter preselection that models only those sigmoid function profiles that exhibit higher quality from the outset than sigmoid functions calculated by randomly selecting parameters. This has the advantage of optimizing the calculation of the sigmoid functions and resulting in improved and faster determination of the traffic jam's end position.

[0027] Furthermore, the task is solved by a system comprising a server, as described in the preceding sections, and vehicles, the vehicles being equipped to transmit measurement data to the server. This results in similar or identical advantages to those already described in connection with the server.

[0028] In a preferred embodiment, at least one vehicle can be configured to transmit measurement data at regular intervals. In another preferred system, at least one vehicle can transmit measurement data when a corresponding request is received from the server. A combination of regular measurement data transmission and measurement data transmission on request is also possible. Furthermore, the at least one vehicle can itself transmit measurement data to the server via a trigger. This trigger can function as a feature of various database management systems, particularly large relational database management systems, and, upon a specific type of data change, call a stored program that allows or prevents this change and / or performs other actions, such as transmitting selected measurement data to the server.This ensures optimal and efficient transmission of measurement data within the system.

[0029] In a further advantageous embodiment, the server can be configured to select at least one vehicle from a list of vehicles, particularly using traffic jam data, and request the selected vehicle to transmit measurement data. The server can request all vehicles in the list to regularly transmit at least position data, which is additionally assigned to the vehicles in the list. Based on this position information, the server selects vehicles located in the vicinity of the traffic jam known from the traffic jam data and requests them to send measurement data. Another possibility is that the server only initiates the request for measurement data from vehicles once it has information about a traffic jam. In this case, the server can request all measurement data from the vehicles listed.Alternatively, the server can first initiate a position query for all listed vehicles and save only the positions of the vehicles in the list. Based on the traffic jam data, a selection of vehicles would then be made, specifically those located within the area of ​​the traffic jam. If no measurement data is yet available for these vehicles, the server can request it in a second step to calculate the position of the end of the traffic jam. All these methods have the advantage of simplifying, optimizing, and ensuring the efficient transmission of measurement data between the vehicles and the server.

[0030] In another preferred configuration, the server can be trained to: a) to determine, based on traffic congestion data, the preliminary position of the end of the traffic jam and / or a traffic jam center and / or a traffic jam start, in addition to a traffic direction; b) to determine the vehicle position and direction of travel for a large number of vehicles; c) using the vehicle position and the vehicle direction of travel, select at least one vehicle that is located in front of the provisional position of the end of the traffic jam and / or the center of the traffic jam, preferably in front of the provisional position of the beginning of the traffic jam, and is moving towards the end of the traffic jam.

[0031] By selecting vehicles that are located ahead of the end of the traffic jam and are approaching it, only measurement data from vehicles directly related to the calculated position of the traffic jam's end is used. This further optimizes and reduces the transmission of measurement data.

[0032] In a further embodiment, the at least one vehicle can include at least one distance measuring unit. This distance measuring unit can be configured to measure the distance between the vehicle and a vehicle traveling ahead of it. This distance can be used to determine and / or transmit traffic information data. Such a distance measuring unit could, for example, be the front radar for ACC (Adaptive Cruise Control), a laser, a camera, or any other unit suitable for measuring the distance to a vehicle ahead. An advantage of such a distance measuring unit is that the distance values ​​can be used to model traffic density profiles. Depending on the speed and the distance to the vehicle ahead, the computer unit or the vehicle itself can determine the traffic density in the vicinity of the measuring vehicle.

[0033] In a preferred embodiment, the server can be configured to transmit the calculated position of the traffic jam's end to vehicles. This allows the position of the traffic jam's end to be displayed in the vehicle. This informs the driver, for example via their navigation system, about the exact position and / or characteristics of the traffic jam's end. If the traffic jam's end is located behind a blind curve or is a hard-edged jam, the driver can be warned in time, thus reducing the risk of an accident.

[0034] Furthermore, the problem is solved by a method for determining the position of a traffic jam end, in particular by means of a server as described in the preceding explanations, and / or within a system as described in the preceding explanations, comprising the steps: – Determining a multitude of parameter sets, wherein each parameter set defines a first sigmoid function and a second sigmoid function, where the first sigmoid function of the parameter set models a speed profile and the second sigmoid function of the parameter set models a traffic density profile; – Receipt of measurement data from at least one vehicle; – Evaluation of the quality of at least some of the sigmoid functions defined by the parameter sets based on the received measurement data; – Selection of at least one parameter set based on the evaluation; – Calculation of the position of the end of the traffic jam based on at least one selected parameter set; – Sending the position of the end of the traffic jam to a / the vehicle.

[0035] The advantages are similar or identical to those already described in connection with the server and the system.

[0036] Another preferred method includes the following steps: – Generating, preferably randomly generating, further sets of parameters based on the at least one selected set of parameters, in particular within predefined variation intervals; – Receipt of further measurement data from at least the vehicle or another vehicle; – Evaluation of the quality of at least some of the sigmoid functions defined by the further parameter sets based on the received second measurement data; – Selection of at least one additional parameter set based on the evaluation; – Calculation of the position of the end of the traffic jam based on at least one additional selected parameter set; – Sending the position of the end of the traffic jam to one / the other vehicle.

[0037] New parameter sets can be generated by randomly and slightly varying the eight parameters within each set, introducing a certain amount of noise. This process allows for the creation of numerous different parameter sets. Based on the previously selected parameter set, these new sets represent the traffic situation in an improved and more refined way compared to the initial sets. By re-evaluating and selecting these parameter sets, the position of the traffic jam end, calculated in the first step, can be further specified and determined more precisely. The process of generating new parameter sets, comparing them, and evaluating them against new measurement data can be repeated as often as desired.This allows the position and characteristics of the end of the traffic jam to be determined more and more precisely, and simultaneously adapted to the current changing conditions.

[0038] The problem according to the invention is further solved by a computer-readable storage medium that contains executable instructions which, when executed, cause a computer to implement the method already described. This results in similar or identical advantages to those already described in connection with the server, the system, and the method.

[0039] The invention is described below by means of several exemplary embodiments, which are explained in more detail with reference to the figures. These show:

[0040] Fig. 1. A schematic representation of a server 100 ;

[0041] Fig. 2 a schematic representation of two communicating servers 100 and 101 ;

[0042] Fig. 3 a schematic representation of a system;

[0043] Fig. 4 A schematic top view of two vehicles driving one behind the other 71 and 72 ;

[0044] Fig. 5 a sigmoid function that represents a velocity profile 50 modeled;

[0045] Fig. 6 a sigmoid function that represents a traffic density profile 60 modeled;

[0046] Fig. 7 a sigmoid function from Fig. 5 to determine the position x2 of the end of the traffic jam;

[0047] Fig. 8 a sigmoid function from Fig. 6 to determine the position x2 of the end of the traffic jam;

[0048] Fig. 9 a schematic representation for determining the property of a jam end;

[0049] Fig. 10 a schematic flow diagram for determining the position x2 of the end of the traffic jam;

[0050] Fig. 11 another schematic flowchart made up Fig. 10 to determine the position x2 of the end of the jam; and

[0051] Fig. 12 an execution cycle of the probabilistic evaluation of the parameter sets 40 and 42 based on measurement data 80 , 81 , 82 and 83 the vehicles 71 , 72 , 73 and 74 .

[0052] In the following description, the same reference numbers are used for identical and equivalently functioning parts.

[0053] The server's goal 100 It is to calculate the position of the end of a traffic jam.

[0054] In the following, the term "end of a traffic jam" will be understood as the position at which a vehicle is forced to reduce its speed and / or distance to a vehicle in front due to external influences, such as a traffic accident, increased traffic volume or environmental influences.

[0055] Fig. Figure 1 shows a schematic representation of a server. 100 , which is a computer unit 10 , a storage 20 , a receiving unit 30 and a unit of assessment 90 includes.

[0056] As in Fig. As shown in 2, the server receives 100 from another server 101 Traffic data 21 The traffic data 21 indicate an area on a road where a traffic jam has occurred.

[0057] The receiving unit 30 is trained to process a large number of measurement data 80 ,81 , 82 , 83 to receive, as it is in Fig. Figure 3 shows the measurement data. 80 , 81 , 82 , 83 Each includes at least one position data entry x of vehicles 71 , 72 , 73 , 74 , whereby the vehicles 71 , 72 , 73 , 74 a fleet of vehicles 70 represent the vehicle fleet 70 is defined by the fact that they are vehicles 71 , 72 , 73 , 74 It concerns vehicles that are in close proximity to each other and are all traveling in the same direction.

[0058] The server 100 is designed to use at least one sigmoid function, defined for example by four parameters [a1, a2, a3, a4], and the received measurement data 80 , 81 , 82 , 83to calculate the position x2 of the end of the traffic jam. The sigmoid functions are determined, for example in a first iteration cycle, by randomly chosen parameter values. The parameters [a1, a2, a3, a4] of the sigmoid function can also be determined using the traffic jam data. 21 will be calculated.

[0059] In the Fig. Figure 3 also shows a schematic representation of a system. The system includes the server. 100 and the vehicles 71 , 72 , 73 , 74 , whereby the vehicles 71 , 72 , 73 , 74 are trained to process measurement data 80 , 81 , 82 , 83 to the server 100 to transmit. The measurement data can be used in this process. 80 , 81 , 82 , 83 automatically from the vehicles at regular intervals 71 , 72 , 73 , 74 to the server 100be transmitted or the transmission of the measurement data 80 , 81 , 82 , 83 This only occurs upon request from the server. 100 A combination of the two transmission methods is also conceivable. This ensures optimal and efficient transmission of measurement data within the system.

[0060] Furthermore, the server selects 100 from a list of vehicles 71 , 72 , 73 , 74 at least one vehicle 71 , especially using traffic data 21 , off. The selected vehicle 71 will be asked to provide measurement data 80 to the server 100 to transmit. The selection of the vehicle 71 This can be done in different ways: In one possibility, the server requests 100 all vehicles 71 , 72 , 73 , 74those on the list, to regularly transmit at least position data x to him, which additionally includes the vehicles 71 , 72 , 73 , 74 are assigned to the list. Based on this position information x, the server selects 100 vehicles 71 , 72 , 73 , 74 from those located in the immediate vicinity of the area affected by the traffic data 21 known traffic jams and requests them to provide him with measurement data. 80 , 81 , 82 , 83 to send, with the help of which the server 100 the position of the congested x2 is determined.

[0061] Another possibility is that the server 100 the querying of measurement data from vehicles 71 , 72 , 73 , 74 It only starts once it receives information. 21 There may be a traffic jam. The server may be experiencing this. 100 on the one hand, all measurement data 80 ,81 , 82 , 83 the vehicles 71 , 72 , 73 , 74 request those listed. On the other hand, the server can 100 As a first step, a position query of all listed vehicles 71 , 72 , 73 , 74 start and only position x of the vehicles 71 , 72 , 73 , 74 Save to the list. Based on traffic data. 21 A selection of vehicles will then be made 71 , 72 , 73 , 74 to take place, whereby vehicles 71 , 72 , 73 , 74 Vehicles located within the traffic jam area will be selected. Are any of these vehicles lying down? 71 , 72 , 73 , 74 No measurement data yet 80 , 81 , 82 , 83 Before, the server can 100In a second step, request this information to calculate the position of the end of the traffic jam x2.

[0062] The server is also 100 trained to use traffic data 21 To determine the preliminary position of the end of the traffic jam x2 and / or a traffic jam center and / or a traffic jam start, as well as the direction of traffic in which the traffic jam occurred. Same determination of the server. 100 This also applies to a large number of vehicles. 71 , 72 , 73 , 74 , whereby their vehicle position x and vehicle direction of travel are determined. Using the vehicle position x and the vehicle direction of travel, the server selects 100 at least one vehicle 71 from, which is located in front of the provisional position of the end of the traffic jam x2 and / or the center of the traffic jam, preferably in front of the provisional position of the beginning of the traffic jam and is moving towards the end of the traffic jam x2. By selecting vehicles 71 ,72 , 73 , 74 , which are located at position x1 in front of the position of the end of the traffic jam x2 and are moving towards it, only such measurement data will be recorded. 80 , 81 . 82 , 83 of vehicles 71 , 72 , 73 , 74 It uses data that is also directly related to the position of the traffic jam end to be calculated. This further optimizes and reduces the transmission of measurement data.

[0063] Once the server 100 Once it has calculated the position of the end of the traffic jam x2, it transmits this position x2 to the vehicles. 71 , 72 , 73 , 74 This allows the position of the end of the traffic jam x2 to be displayed in the vehicles. 71 , 72 , 73 , 74This allows the driver to be informed, for example via their navigation system, about the exact position x2 and / or the characteristics of the traffic jam's end. If the end of the traffic jam x2 is located behind a blind curve or is a hard-edged jam, the driver of the vehicle can 71 , 72 , 73 , 74 They will be warned in time, thus reducing the risk of accidents.

[0064] The vehicles 71 , 72 , 73 , 74 transmit measurement data 80 , 81 , 82 , 83 These are data tuples. These data tuples include traffic information data, speed data, and the speed v of the respective vehicle. 71 Specify and provide distance data that define a distance r between the respective vehicles. 71 and one of the respective vehicles 71 preceding vehicle 72 Specify. As in Fig. As shown in section 4, the vehicle includes 71 a transmitter unit 76 , to obtain the measurement data 80 to the server 100 to transmit. Furthermore, the vehicle includes 71 a distance measuring unit 75 , to determine the distance r to a vehicle in front 72 to eat.

[0065] Using the measurement data 80 , 81 , 82 , 83 The server determines 100 the vehicle's surroundings 71 , such as the traffic density ρ, where the traffic density ρ depends on the measured distance r and the vehicle length s of the measuring vehicle 71 is.

[0066] Fig. Figure 5 shows a sigmoid function that represents a velocity profile. 50 The model is defined by four parameters [v1, v2, v3, v4]. It illustrates how the speed v of a vehicle changes. 71 changes via location x. Fig. Figure 6 shows a sigmoid function representing a traffic density profile. 60 The model is defined by four parameters [ρ1, ρ2, ρ3, ρ4]. It illustrates how the traffic density ρ changes in the immediate vicinity of a vehicle. 71 The speed changes at location x. Each speed profile is a separate profile. 50 and a traffic density profile 60 represent a parameter set 40 The server 100 is trained to handle a large number of parameter sets 40 , 42 to determine the position of the end of the traffic jam x2. A parameter set 40 , 42 This can be determined by eight parameters [v1, v2, v3, v4, ρ1, ρ2, ρ3, ρ4]. This allows for precise localization of the vehicles. 71 , 72 , 73 , 74 Thus, the position of the end of the traffic jam x2 and the properties of the end of the traffic jam are determined with lane-level accuracy.

[0067] Fig. Figure 7 shows the speed profile.50 and Fig. 8 based on the traffic density profile 60 , how the position of the end of the traffic jam x2 is determined. This is done using the speed profile. 50 A tangent line 51 to the constant velocity v is drawn at a position x1 before the end of the traffic jam x2. A second straight line at a position x3 inside the traffic jam represents the slope. 52 The decrease in speed after the end of the traffic jam is represented by x2. This is due to the intersection of the tangent line 51 and the slope. 52 The position of the end of the traffic jam x2, and thus the beginning of the jam entry, is determined. The same procedure is used in Fig. 8 using the traffic density profile 60 This method is also used to determine the position of the end of the traffic jam x2. Here, a tangent 61 to the constant traffic density ρ is drawn at a position x1 before the end of the traffic jam x2. A second straight line at a position x3 inside the traffic jam represents the slope. 62The increase in traffic density after the end of the traffic jam is represented by x2. This is due to the intersection of tangent 61 and the slope. 62 The position of the end of the traffic jam x2 and thus the beginning of the traffic jam entrance is determined.

[0068] Does the slope increase 52 of the speed profile 50 quickly down and the incline 62 of the traffic density profile 60 quickly, it is a hard traffic jam where vehicles 71 , 72 , 73 , 74 from a free-running vehicle to encounter a group of, for example, stationary vehicles. As the incline increases... 52 of the speed profile 50 slowly down and the incline 62 of the traffic density profile 60 slowly approaching, it is a soft traffic jam end into which the vehicles 71 , 72 , 73 , 74 drive in steadily at an ever-decreasing speed v.

[0069] Another way to determine the property of the end of the congestion is in Fig. Figure 9 illustrates this. If the gradient dv of the speed decrease over time has a large negative value and the gradient dρ of the traffic density increase over time has a large positive value, this indicates a hard end to a traffic jam. Conversely, if the speed gradient over time has a small negative value and the traffic density gradient over time has a small positive value, this indicates a soft end to a traffic jam.

[0070] By transmitting the position of the traffic jam end x2 as well as the property of the traffic jam end by the server 100 to the vehicles 71 , 72 , 73 , 74 , this information is processed and used to issue warning signals to the driver or other road users.

[0071] Fig. Figure 10 shows a flowchart of a procedure used to determine the position of the end of the traffic jam x2. Here, the server is... 100 trained to perform the following steps: – Determining a large number of parameter sets 40 , where each parameter set 40 defines a first sigmoid function and a second sigmoid function, where the first sigmoid function represents a velocity profile. 50 and the second sigmoid function a traffic density profile 60 modeled; – Receiving measurement data 80 at least one vehicle 71 ; – Evaluation of the quality of at least some of the parameters defined by the parameter sets 40 defined sigmoid functions using a unit of evaluation 90 based on the received measurement data 80 ; – Selection of at least one parameter set 41 based on the assessment; – Calculation of the position x2 of the end of the traffic jam based on at least one selected parameter set 41 ; – Sending the position x2 of the end of the traffic jam to a / the vehicle 71 .

[0072] The assessment unit 90 is trained to handle the parameter sets 40 to be evaluated using a particle filter. The particle filter enables continuous updates of the sigmoid functions. 50 , 60 through new measurement data 80 , 81 , 82 , 83 generated. Here, the particle filter approximates the a posteriori distribution of the state probabilities of the sigmoid functions. 50 , 60 by a finite set of parameters [v1, v2, v3, v4, ρ1, ρ2, ρ3, ρ4]. The probability density function is then determined by a sample set, the particles, using the sigmoid functions. 50 , 60approximates. In contrast to alternative approaches, particle filters, due to their non-parametric form, can approximate arbitrary distributions.

[0073] For an even more precise determination of the position of the end of the traffic jam x2, or for an update of the position of the end of the traffic jam x2, shows Fig. 11. Another flowchart, which determines the position of the end of the traffic jam x2. Here, the server 100 trained to perform the following steps: – Generating, preferably randomly generating, further sets of parameters 42 based on at least one selected parameter set 41 , especially within given variation intervals; – Receipt of further measurement data, at least from the vehicle 71 and / or another vehicle 72 ; – Evaluation of the quality of at least some of the parameters determined by the other parameter sets 40defined sigmoid functions using a unit of evaluation 90 based on the further measurement data 81 ; – Selection of at least one additional parameter set 43 based on the assessment; – Calculation of the position x2 of the end of the traffic jam based on at least one other selected parameter set 43 ; – Sending the position x2 of the end of the traffic jam to a / the vehicle 71 or another vehicle 72 .

[0074] The generation of new parameter sets 42 This can be achieved by defining the eight parameters [v1, v2, v3, v4, ρ1, ρ2, ρ3, ρ4] per parameter set 41 Each parameter set is slightly and randomly altered with a certain amount of noise. This measure allows for a wide variety of different parameter sets. 42 generated by the previously selected parameter set. 41 represent the new parameter sets 42compared to the first parameter sets 40 The traffic situation in an improved and adapted form. By re-evaluating and selecting the parameter sets. 42 The position of the traffic jam end x2, which was calculated in a first step, can be further specified and determined more precisely through this second step. This involves generating new parameter sets. 42 , comparing and evaluating these new parameter sets 42 with ever new measurement data 81 This process can be repeated as often as desired. Therefore, the position x2 and the properties of the congestion end can not only be determined more and more precisely, but are also continuously adapted to the current, changing conditions.

[0075] In Fig. Figure 12 illustrates yet another way in which the position of the traffic jam end x2 can be determined. Estimating the most probable parameter set. 41This occurs cyclically. At the beginning, it is done randomly or with the help of traffic data. 21 about more likely parameterizations – a large number of parameter sets 40 each with eight parameters [v1, v2, v3, v4, ρ1, ρ2, ρ3, ρ4]. In a next step, the sigmoid functions of the velocity profile are generated. 50 and the traffic density profile 60 , which are uniquely defined by eight parameters [v1, v2, v3, v4, ρ1, ρ2, ρ3, ρ4], using measurement 1000 in the evaluation step 2000 evaluated. Parameter sets 40 , which the measurement 1000 Those that correspond better or are closer to the actual measured case will therefore receive a higher rating. In the selection step 3000 The parameter sets will be 41 It is determined which parameters should be pursued further. Then, the eight parameters [v1, v2, v3, v4, ρ1, ρ2, ρ3, ρ4] are determined for each selected parameter set. 41Each parameter is slightly altered randomly with a certain amount of noise. Now, a multitude of different parameter sets are available. 42 before. Through the measurements 1000 represents the multitude of parameter sets 40 The traffic situation is now better than before the assessment step. 2000 Will further measurement data be available at a later date? 81 or multiple synchronous / asynchronous measurements 1000 , the parameter sets 41 The prediction from the last time step to the respective new time point is made. This can be done, for example, using macroscopic traffic models described by partial differential equations. The prediction step can also be... 4000 can also be omitted completely if the noise in the parameter sets is too high. 42 , which before or after the prediction 4000It can be applied if it is large enough to capture the dynamics of the position of the traffic jam end x2. This sequence of steps is evaluated. 2000 , based on the measurement 1000 , Selection 3000 as well as prediction 4000 This process occurs cyclically and as often as desired. Reference symbol list 10 computer units 20 storage 21 traffic data 30 receiver units 40 parameter sets 41 Parameter set 42 additional parameter sets 43 additional parameter sets 50 Speed ​​profile 51 Tangent to the speed before the end of the traffic jam 52. Gradient of the speed decrease 60 Traffic density profile 61 Tangent to traffic density before the end of the traffic jam 62. Increase in traffic density 70 vehicle fleet 71 vehicles 72 more vehicles 73 vehicles 74 vehicles 75 Distance measuring unit 76 transmitting units 80 measurement data 81 more measurement data 82 measurement data 83 measurement data 90 rating units 100 servers 101 more servers 1000 measurements 2000 rating 3000 choices 4000 prediction x Position data entry x1 Position before the end of the traffic jam x2 Position of the end of the traffic jam x3 Position after the end of the traffic jam v speed dv Gradient of the velocity decrease over time ρ traffic density dρ Gradient of traffic density increase over time r paragraph between two vehicles s length of the vehicle QUOTES INCLUDED IN THE DESCRIPTION

[0076] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0077] EP 1235195 A2

[0005] US 2007 / 0005231 A1

[0006]

Claims

[1] Server ( 100 ) to determine a position (x2) of a traffic jam end, comprising – a computing unit ( 10 ); – a storage ( 20 ); – a receiving unit ( 30 ) to receive a large number of measurement data ( 80 ), each with at least one position data entry (x) of a vehicle ( 71 ); where the server ( 100 ) is formed, using at least one sigmoid function and the received measurement data ( 80 ) to calculate the position (x2) of the end of the traffic jam. [2] Server ( 100 ) according to claim 1, characterized by that the measurement data ( 80 Data tuples are and include: – Traffic information data; and / or – Speed ​​data that includes at least one speed (v) of the respective vehicle ( 71 ) specify; and / or – Distance data that specify at least a distance (r) between the respective vehicle ( 71 ) and one corresponding to the respective vehicle ( 71 ) indicate the vehicle ahead; and / or – Braking frequency data, which shows the braking frequency of the respective vehicle ( 71 ) indicate. [3] Server ( 100 ) according to any of the preceding claims, characterized by that the server ( 100 ) is trained to handle a variety of parameter sets ( 40 ) to determine, where each parameter set ( 40 ) defines a first sigmoid function and a second sigmoid function, where the first sigmoid function represents a velocity profile ( 50 ) and the second sigmoid function a traffic density profile ( 60 ) modeled. [4] Server ( 100 ) according to one of the preceding claims, characterized by a valuation unit ( 90), which is trained to ensure the quality of at least a selection of items from the server ( 100 ) calculated multitude of sigmoid functions using at least some of the measurement data ( 80 ) to evaluate. [5] Server ( 100 ) according to one of the preceding claims, in particular according to claim 4, characterized by that the unit of assessment ( 90 ) is trained to evaluate the sigmoid functions using a particle filter and / or a support vector machine and / or linear discriminant analysis. [6] Server ( 100 ) according to any of the preceding claims, characterized by that the server ( 100 ) from another server ( 101 ) Traffic data ( 21 ) receives the traffic data ( 21 ) specify an area where a traffic jam has occurred, and calculate the sigmoid function using the traffic jam data ( 21 ) is calculated. [7] System that includes a server (100 ) according to one of the preceding claims and vehicles ( 71 , 72 , 73 , 74 ) includes, where the vehicles ( 71 , 72 , 73 , 74 are trained to process measurement data ( 80 ) to the server ( 100 to be transmitted. [8] System according to claim 7, characterized by that at least one vehicle ( 71 , 72 , 73 , 74 ) is trained to do this: – measurement data at regular intervals ( 80 ) to transmit; and / or – upon request from the server ( 100 ) Measurement data ( 80 to be transmitted. [9] System according to any one of the preceding claims, characterized by that the server ( 100 ) is trained to select from a list of vehicles ( 71 , 72 , 73 , 74 ) at least one vehicle ( 71), especially using traffic data ( 21 ) select and the selected vehicle ( 71 ) to request measurement data ( 80 to be transmitted. [10] System according to claim 9, characterized by that the server ( 100 ) is trained to do this: a) to determine, based on the traffic jam data, a preliminary position of the end of the traffic jam (x2) and / or of a traffic jam center and / or of a traffic jam start, in addition to a traffic direction; b) for a large number of vehicles ( 71 , 72 , 73 , 74 ) To determine vehicle position (x) and vehicle direction of travel; c) using the vehicle position (x) and the vehicle direction of travel at least one vehicle ( 71 ) to select which is located in front of the provisional position of the end of the jam (x2) and / or the center of the jam, preferably in front of the provisional position of the beginning of the jam and is moving towards the end of the jam (x2). [11] System according to any one of the preceding claims, characterized by that at least one vehicle ( 71 ) at least one distance measuring unit ( 75 ) includes, which is designed to determine the distance (r) between the vehicle ( 71 ) and one of the vehicle ( 71 ) to measure the distance (r) of the vehicle ahead and the distance is used to determine and / or transmit traffic information data. [12] System according to any one of the preceding claims, characterized by that the server ( 100 ) is designed to display the calculated position (x2) of the end of the traffic jam to vehicles ( 71 , 72 , 73 , 74 to be transmitted. [13] Method for determining a position (x2) of a traffic jam end, in particular by means of a server ( 100 ) according to any one of claims 1 to 6 and / or within a system according to any one of claims 7 to 12, comprising the steps: – Determining a large number of parameter sets ( 40 ), where each parameter set ( 40 ) defines a first sigmoid function and a second sigmoid function, where the first sigmoid function represents a velocity profile ( 50 ) and the second sigmoid function a traffic density profile ( 60 ) modeled; – Receiving measurement data ( 80 ) at least one vehicle ( 71 ); – Evaluation of the quality of at least some of the parameters defined by the parameter sets ( 40 ) defined sigmoid functions based on the received measurement data ( 80 ); – Selection of at least one parameter set ( 41 ) based on the rating; – Calculation of the position (x2) of the end of the traffic jam based on at least one selected parameter set ( 41 ); – Sending the position (x2) of the end of the traffic jam to a / the vehicle ( 71 ). [14] The method of claim 13, comprising the steps: – Generating, preferably randomly generating, further sets of parameters ( 42 ) based on at least one selected parameter set ( 41 ), especially within predetermined variation intervals; – Receipt of further measurement data ( 81 ) at least of the vehicle ( 71 ) and / or another vehicle ( 72 ); – Evaluation of the quality of at least some of the parameters determined by the other parameter sets ( 40 ) defined sigmoid functions based on the further measurement data ( 81 ); – Selection of at least one additional parameter set ( 43 ) based on the rating; – Calculation of the position (x2) of the end of the traffic jam based on at least one other selected parameter set ( 43 ); – Sending the position (x2) of the end of the traffic jam to a / the vehicle ( 71) or another vehicle ( 72 ). [15] Computer-readable storage medium comprising executable instructions which cause a computer to implement the method of claim 13 or 14 when the instructions are executed.

Images