Method for predicting the spatial and temporal extent of a traffic jam and information technology system

Abstract

The invention relates to a method for predicting the spatial and temporal extent of a traffic jam (1) on a road section (2), wherein vehicles of a vehicle fleet continuously transmit position information to a central computing unit.The method according to the invention is characterized in that a spatio-temporal congestion endpoint (P2), a spatio-temporal congestion release startpoint (P3) and a spatio-temporal congestion endpoint (P4) are determined, wherein the envelope (5) between the congestion release startpoint (P1), congestion endpoint (P2), congestion release startpoint (P3) and congestion endpoint (P4) describes the spatio-temporal extent of the congestion (1), and wherein the central computing device continuously re-estimates the envelope (5) during the duration of the congestion (1), taking into account the data points received from the vehicles during this time and the resulting actual position of the congestion endpoint (P2), congestion release startpoint (P3) and congestion endpoint (P4).

Description

[0001] The invention relates to a method for predicting the spatial and temporal extent of a traffic jam on a road section according to the type defined in more detail in the preamble of claim 1, and to an information technology system for carrying out the method.

[0002] Navigation systems offer a significant advantage for drivers. They make it easier to navigate the road network, allowing drivers to orient themselves more effectively, especially in unfamiliar areas. A destination can be entered, and a route is calculated from the vehicle's current location to the destination and displayed on a map. The driver can then follow this route to their destination.

[0003] On the way to the destination, events can occur that extend the travel time or suggest taking an alternative route. For example, there might be traffic jams, slow-moving traffic, road closures, or similar issues. Navigation systems can take such events into account when estimating travel time and planning routes. To enable the driver to make the best decision about whether to continue on the original route or choose an alternative route, the estimated travel time and, if applicable, the suggested alternative route should be determined by estimating the extent of the traffic jam as realistically as possible. If the extent and duration of the traffic jam are estimated incorrectly, the travel time displayed may be too long or too short, which could lead the driver to make an incorrect decision.If the estimated duration or extent of a traffic jam is too long, drivers are more likely to choose an alternative route; conversely, if the estimated duration or extent of the traffic jam is too short, they are more likely to remain on the original route. However, staying on the original route or taking the alternative route could actually result in a shorter overall travel time.

[0004] Current methods calculate the spatial and temporal extent of traffic jams, and thus the estimated travel time, solely based on traffic volume and average vehicle speed. This approach provides an inadequate prediction of the spatial and temporal extent of traffic jams, thereby increasing the risk of drivers making incorrect route choices.

[0005] This creates a need to specify means with which the spatial and temporal extent of traffic jams can be predicted more accurately.

[0006] The analysis of traffic jams is described in the following paper, also published by the inventor: L. Querfurth et al., Crash localization and traffic impact assessment via spatio-temporal analysis of connected vehicle data, Accident Analysis and Prevention 214 (2025) 10795, February 17, 2025, https: / / doi.org / 10.1016 / j.aap.2025.107956. In this system, vehicles in a fleet act as sensors for evaluating traffic conditions. At regular intervals, the vehicles transmit their current position, determined by a vehicle-integrated navigation system, to a server. Each position data entry is accompanied by a timestamp. The server can thus determine for each vehicle individually which part of the road network it is on at any given time and, consequently, its speed.When an unexpected event occurs that leads to a traffic jam, such as an accident, characteristic patterns can be identified in the corresponding traffic data. Initially, a few vehicles brake sharply, prompting the following vehicles to also brake hard. This front of vehicles braking to a standstill spreads out at a characteristic speed in the opposite direction of travel. The vehicles pass through characteristic speed phases or traffic phases, which are categorized according to Kerner: "free-flowing traffic," "synchronously flowing traffic," and "congested traffic." An accident-related traffic jam is characterized by the fact that the "free-flowing traffic" phase transitions to "congested traffic" relatively quickly, without a pronounced "synchronously flowing traffic" phase. Relevant traffic data can be analyzed in real time.This can be used to identify accidents on roads without requiring the vehicles involved to transmit data to an external agency. Based on characteristic patterns in the vehicles' braking behavior, different types of accidents or traffic jams can be classified. This classification focuses particularly on the impact of the resulting traffic jam on the subsequent flow of traffic. Predicting the spatial and temporal development of traffic jams is not part of this paper.

[0007] The entire content of this paper is considered included in this application by reference.

[0008] The present invention is based on the objective of providing an improved method for predicting the spatial and temporal extent of a traffic jam on a road section, which is characterized by a particularly realistic estimation of the extent in real time or near real time.

[0009] According to the invention, this problem is solved by a method for predicting the spatial and temporal extent of a traffic jam with the features of claim 1. Advantageous embodiments and further developments, as well as an information technology system for carrying out the method, are described in the dependent claims.

[0010] A generic method for predicting the spatiotemporal extent of a traffic jam on a road segment, wherein vehicles of a fleet continuously transmit position information, describing their current location at a given time of recording, to a central computing unit, and the central computing unit determines, by establishing a relationship between the respective position information and the times of recording, which vehicle is moving at what speed and at what time on the road segment, wherein the spatiotemporal starting point of the traffic jam is determined by identifying at least one vehicle that reduces its speed from essentially the applicable speed limit to essentially standstill following a defined pattern, is further developed according to the invention by...that, assuming a fixed backpropagation speed of an upstream traffic jam front against the direction of travel and a fixed duration of the traffic jam, a spatial-temporal endpoint of the traffic jam is determined, defining the place and time on the road segment at which the traffic jam essentially no longer increases; Assuming a substantially stationary downstream congestion front and a defined congestion dissipation start time, a spatial-temporal congestion dissipation start point is determined, defining the location and time on the road segment at which the congestion begins to dissipate; Assuming a fixed backpropagation speed of the downstream congestion front against the direction of travel from the congestion dissipation point until the upstream congestion front reaches the position defined by the congestion endpoint, a spatio-temporal congestion endpoint is determined, wherein the envelope between the congestion startpoint, congestion endpoint, congestion dissipation point and congestion endpoint describes the spatio-temporal extent of the congestion, and wherein During the duration of the traffic jam, the central computing unit continuously re-estimates the envelope, taking into account the data points received from the vehicles during this time and the resulting actual location of the end point of the traffic jam, the beginning point of the traffic jam clearance, and the end point of the traffic jam.

[0011] The inventive method enables a particularly realistic prediction of the spatial and temporal extent of traffic jams in real time. For this purpose, vehicles in a fleet continuously transmit their current location to the central computing unit. The central computing unit can, for example, be contactable via the internet and thus also be referred to as a cloud server. Vehicles can use integrated navigation systems to determine their position. It is also conceivable that mobile devices, such as smartphones, could function as vehicles, i.e., transmit the corresponding position data. The mobile device does not necessarily have to be connected to the vehicle.Taking into account the position of the respective vehicle on the road section and the time difference between successive recording times, the central computing unit can determine the distance traveled per unit of time and thus also the speed of movement of the vehicle.

[0012] If an unexpected event occurs on a section of road leading to a traffic jam, particularly an accident, one or more vehicles will initially brake according to characteristic patterns. This allows the point where the traffic jam begins to be identified. Specifically, the three traffic phases defined by Kerner are traversed. The braking vehicles pass through the phases of "free-flowing," "synchronous flow," and "congested" in a relatively short time, without a pronounced synchronous flow phase. The speed of the vehicles at the downstream end of the traffic jam is almost zero. The point where the traffic jam begins can therefore be identified by the fact that the vehicles fall below a certain speed threshold with a characteristic temporal behavior. This temporal behavior is also referred to as the "time constraint."The precise parameters for the speed threshold and the timing depend on the data quality of the data points transmitted by the vehicles (i.e., their position information and recording times), applicable legal requirements, and the driving behavior of a typical driver in the respective geographic region under consideration. For details, please refer to the paper published by the inventor mentioned at the beginning.

[0013] The starting point of a traffic jam describes the location and time on the road segment where the jam begins. Such a traffic jam is characterized by an upstream and a downstream front. The downstream front is formed by the vehicles at the front in the direction of travel. The vehicles approaching the jam form the upstream front. The upstream front continues to grow in the opposite direction of travel because more vehicles enter the jam than exit it at the downstream front. In particular, no further travel is possible for a certain period of time, so the downstream front is essentially stationary.

[0014] The upstream traffic jam front propagates at a typical speed in the opposite direction of travel for a characteristic duration. This speed is represented by the defined backpropagation speed. To make an initial estimate of the spatial and temporal extent of the traffic jam, the central computing unit uses a defined backpropagation speed of the upstream traffic jam front. The value of this defined backpropagation speed can be determined, for example, by a developer. Furthermore, the upstream traffic jam front is assumed to propagate with a defined duration. This duration corresponds to a fixed time period, which is also determined by a developer, for example, based on experience. This allows for an initial estimate of the location of the endpoint of the traffic jam on the road segment, as well as the time at which the endpoint will occur.The endpoint of a traffic jam can also be called the "Upstream Flow Starvation Point." This point is created by the rerouting of vehicles approaching the jam upstream through appropriate diversion systems. Such diversion systems primarily involve navigation systems in the respective vehicles that have detected the traffic jam. Diversion based on traffic jam information displayed on stationary traffic signs or broadcast over the radio is also conceivable. This leads to a reduction in the traffic flow upstream of the jam and thus a proportional decrease in the backpropagation speed of the upstream jam front. At a certain point, it may happen that no further vehicles approach the jam, so that the upstream jam front essentially becomes stationary.

[0015] It can also happen that a fixed congestion duration is assumed to initially estimate the spatial and temporal extent of the congestion, but the congestion continues to grow after the fixed duration is exceeded, without the endpoint being observed. In this case, the location of the endpoint can be estimated again by assuming another fixed duration. This second fixed duration can be the same as the first or a different duration. This process continues until the endpoint of the congestion can actually be determined from the data points transmitted by the vehicles.

[0016] The location of the point where the traffic jam begins to clear on the road segment is essentially determined by the position where the downstream congestion front initially formed. A fixed duration for the beginning of the traffic jam is assumed to determine when the vehicles resume their journey. This duration corresponds, for example, to a time period also determined by a developer based on experience. As soon as the vehicles at the front of the traffic jam start moving again, the congestion begins to clear. Thus, the vehicles further back in the traffic jam gradually start moving. The downstream congestion front then propagates at a specific, characteristic backpropagation speed against the direction of travel towards the upstream congestion front. This fixed backpropagation speed of the downstream congestion front can also be determined by a developer based on experience.The point of maximum congestion is reached when the downstream congestion front and the upstream congestion front meet.

[0017] The method involves continuously re-estimating the starting point, end point of the traffic jam, the beginning point of its dissipation, and the end point throughout the duration of the traffic jam, i.e., from the detection of the starting point until the actual observation of the end point. Initially, the assumptions described above are made. Over the course of the traffic jam observation, and thus based on the data points transmitted by the vehicles, the actual occurrence of these points can be observed gradually. This is used to progressively adjust the envelope between the starting point, end point of the traffic jam, the beginning point of its dissipation, and the end point, thereby approximating these points to the actual spatial and temporal extent of the traffic jam. This allows for a particularly realistic prediction of the traffic jam's spatial and temporal extent.

[0018] After detecting the starting point of the traffic jam, a predetermined backpropagation speed is initially assumed for the upstream traffic jam front to propagate against the direction of travel. As time progresses, the actual backpropagation speed of the upstream traffic jam front can be determined from the data points transmitted by the vehicles. To further estimate the endpoint of the traffic jam after the initial estimate, the backpropagation speed actually observed from the data points is then used instead of the predetermined backpropagation speed. This allows for an even more precise and therefore more realistic determination of the location of the spatiotemporal endpoint of the traffic jam.

[0019] The same applies to the magnitude of the backpropagation speed of the downstream congestion front, which is taken into account in the estimation after the start of the congestion dissipation. This determines the speed at which the vehicles actually start moving after the congestion dissipation begins and the resulting speed at which the downstream congestion front moves in the direction of the upstream congestion front.

[0020] An advantageous refinement of the method involves determining the specified backpropagation speed of the upstream and downstream congestion fronts, as well as the specified duration of the congestion and the specified duration of the congestion's resolution, based on empirically determined historical data points from past congestion events. This allows for a particularly realistic initial estimate of the respective backpropagation speeds and durations, such that the initial estimate of the congestion's spatial and temporal extent will already show a high degree of agreement with the actual extent. For this purpose, congestion information aggregated by the central computing unit over an observation period, such as the last month, the last few months, the last year, several past years, and so on, can be taken into account.A local reference can also be established, allowing for different characteristic backpropagation speeds and durations to be used for different geographical regions, such as countries. These speeds and durations can also depend on the specific type of road, for example, a rural road, expressway, motorway, highway, or the like. The number of lanes can also be considered as an additional parameter.

[0021] It is particularly preferred that, depending on the pattern with which the at least one vehicle reduces its speed to detect the starting point of the traffic jam, different values ​​are determined for the defined backpropagation speeds of the upstream and downstream traffic jam fronts, as well as the defined duration of the traffic jam and the defined duration of the traffic jam's resolution. As already mentioned at the outset, the patterns with which individual vehicles brake at the downstream traffic jam front allow conclusions to be drawn about the nature of the traffic jam. In particular, if a large number of vehicles in close proximity brake particularly sharply, this is an indication of a critical accident, which can lead to a complete closure of the roadway. This thus has a particularly strong influence on the backflow of traffic. For corresponding classifications of the traffic flow, or...The traffic jam is also referred to in the paper submitted by the inventor.

[0022] By taking into account characteristic backpropagation velocities and durations for the different types of congestion, even more realistic initial estimates of the spatial and temporal extent of the congestion can be determined.

[0023] A further advantageous embodiment of the method according to the invention further provides that the magnitude of the fixed backpropagation velocity of the upstream congestion front is -12.0 km / h; and / or The height of the specified backpropagation velocity of the downstream congestion front is -15.0 km / h.

[0024] Since each traffic jam front moves against the direction of travel, it is assigned a negative velocity. Typically, the upstream traffic jam front moves with a backpropagation velocity of -12.0 km / h. The downstream traffic jam front, on the other hand, typically moves with a backpropagation velocity of -15.0 km / h. Taking these two values ​​into account allows for a particularly good approximation of the defined backpropagation velocities, enabling an even more precise initial estimate of the spatiotemporal position of the traffic jam's endpoint and the point of congestion.

[0025] According to a further advantageous embodiment of the method according to the invention, the central computing unit for detecting the actual location of the endpoint of the traffic jam takes into account whether the instantaneous rate of change of the backpropagation velocity of the upstream traffic jam front exceeds a defined threshold. The level of this threshold can also be determined by a developer based on their experience. Alternatively, the level of the defined threshold can be determined by considering historical traffic jam data from the data points aggregated by the central computing unit during past traffic jams. Different thresholds can also be defined depending on the geographical region, characteristic time periods such as during the holiday season, or even the type of road.

[0026] As soon as vehicles upstream of the traffic jam are informed about it, they are rerouted. This reduces the number of vehicles approaching the end of the jam, thereby changing the backpropagation speed of the upstream traffic jam front. Specifically, the backpropagation speed, or its magnitude, decreases. For example, the backpropagation speed drops from -12.0 km / h to -4.0 km / h. A particularly preferred threshold is set at a rate where the backpropagation speed of the upstream traffic jam front is halved or reduced by more than half.

[0027] A further advantageous embodiment of the method according to the invention provides that the central computing unit for detecting the actual position of the endpoint of the congestion takes into account whether the instantaneous backpropagation speed of the upstream congestion front falls below a defined threshold. Thus, the position of the endpoint of the congestion can be determined not only from the rate of change of the backpropagation speed, but also from the magnitude of the backpropagation speed itself. In particular, if the upstream congestion front is propagating only very slowly against the direction of travel, this is an indication that the endpoint of the congestion has been reached. A threshold value from the defined interval of -1 km / h to +10 km / h is particularly preferred. It is therefore also conceivable that the upstream congestion front is already moving in the direction of travel.Determining the location of the damming endpoint as a function of the propagation speed of the upstream damming front allows the location of the damming endpoint to be determined when this is not possible solely by considering the instantaneous rate of change of the backpropagation speed.

[0028] According to a further advantageous embodiment of the method according to the invention, the central computing unit for detecting the actual position of the starting point of the traffic jam takes into account whether the instantaneous backpropagation speed of the downstream traffic jam front falls below a defined threshold, in particular a threshold within the defined interval of -30.0 km / h to -2.0 km / h. The traffic jam begins to clear when the vehicles located at the front of the traffic jam start to move. This movement of the vehicles propagates against the direction of travel. This typically occurs at a speed within the interval of -30.0 km / h to -2.0 km / h.If it is detected that the downstream congestion front is starting to move, in particular with a backpropagation speed from said closed interval, and especially preferably with a speed of essentially -15.0 km / h, then the starting point of the congestion resolution is detected.

[0029] A further advantageous embodiment of the method according to the invention provides that the central computing unit for predicting the location of the congestion endpoint forecasts when and where the upstream congestion front, propagating from the congestion endpoint at its backpropagation velocity, and the downstream congestion front, propagating from the congestion dissipation point at its backpropagation velocity, will meet. As already mentioned at the outset, the location of the congestion endpoint is initially estimated assuming a fixed backpropagation velocity of the downstream congestion front until it reaches the assumed fixed location of the upstream congestion front. Gradually, the actual location of the congestion endpoint and the congestion dissipation point, as well as the respective backpropagation velocities, are determined. Taking these values ​​into account, the actual location of the congestion endpoint can be estimated even more accurately.Once the endpoint of the traffic jam is actually determined from the transmitted data points, the end of the traffic jam is detected. Performing a traffic jam forecast after the traffic jam has cleared is then unnecessary.

[0030] The respective backpropagation speeds of the upstream and downstream congestion fronts can be zero, positive, or negative during different phases of the congestion. In particular, after the congestion begins to dissipate, the downstream congestion front propagates at a backpropagation speed of -15.0 km / h in the opposite direction of travel, while the upstream congestion front remains essentially stationary from the point where the congestion ends, or moves relatively slowly in the opposite direction, thus causing the congestion to continue to grow.

[0031] According to a further advantageous embodiment of the method according to the invention, a navigation system used for route guidance utilizes the current spatiotemporal extent of the traffic jam, as predicted by the central computing unit, to forecast the estimated travel time for a vehicle on a navigation route passing through the traffic jam. Since the spatiotemporal extent of the traffic jam predicted by the central computing unit using the method according to the invention is particularly realistic, a realistic forecast of an extended travel time due to the traffic jam is also possible. This makes it easier for drivers to assess early on whether to proceed through the traffic jam or, if possible, to choose an alternative route.Furthermore, the user experience can be improved by displaying a particularly realistic travel time, thus preventing the navigation system from already showing a departure when the vehicle is actually still stuck in traffic or from displaying an excessively long travel time even though the vehicle can already start moving.

[0032] The navigation system could be a vehicle-integrated system. However, it is also conceivable that the navigation system is integrated into a mobile device, such as a smartphone.

[0033] A further advantageous embodiment of the method according to the invention provides that a navigation system used for route guidance takes into account the current spatiotemporal extent of the traffic jam, as predicted by the central computing unit, when calculating the navigation route. Thus, the spatiotemporal extent of the traffic jam predicted by the central computing unit using the described method according to the invention can also be used for calculating navigation routes. Since this extent of the traffic jam is particularly realistic, it is also possible to calculate suitable navigation routes. In this way, alternative routes can be determined that actually lead to a shorter travel time compared to the originally chosen route. This prevents a driver from being guided through the traffic jam even though the actual travel time through the traffic jam is longer than using an alternative route, and vice versa.

[0034] According to the invention, a generic information technology system is further developed by a central computing unit, configured as described above, for predicting the spatial and temporal extent of a traffic jam. The central computing unit thus has read access to a computer-readable storage medium containing machine-interpretable instructions. When executed by a processor of the central computing unit, these instructions cause the unit to analyze the data points transmitted by the vehicles and determine the starting point, end point, and resolution point of the traffic jam, thereby predicting its spatial and temporal extent. The central computing unit is capable of receiving position information and timestamps from the vehicles or the mobile devices carried in the respective vehicles.The recording time can correspond to the time at which the respective position information was recorded in the vehicle or by the mobile device. Alternatively or additionally, the recording time can include the time at which the central computing unit received the respective position information or when it was transmitted by the vehicle or mobile device.

[0035] Accordingly, the central computing unit is also capable of transmitting the prediction of the spatial and temporal extent of the traffic jam back to the respective vehicles or mobile devices. For this purpose, the vehicles or mobile devices are preferably connected to the central computing unit via mobile network over the internet.

[0036] Further advantageous embodiments of the inventive method for predicting the spatial and temporal extent of a traffic jam and of the inventive information technology system also result from the exemplary embodiments, which are described in more detail below with reference to the figures.

[0037] This shows: Fig. 1. A schematic representation of a traffic jam at two different times; Fig. 2 a schematic representation of the spatial and temporal extent of the traffic jam in a trajectory map; Fig. 3 a schematic representation of a trajectory map in which parameters characterizing the spatial and temporal extent of the traffic jam are highlighted; Fig. 4 a schematic representation of a prediction of the spatial and temporal extent of the traffic jam using a method according to the invention at different prediction times; and Fig. 5 a flowchart of the method according to the invention.

[0038] Fig. Figure 1 shows a top view of a road segment 2 at two different times. The extent of the road segment 2 is illustrated by the distance d. For example, the distance d could be the respective kilometer segment of a highway. Due to an unforeseen event, such as an accident at the location marked with the large X, a traffic jam 1 forms. Fig. In section 1A, traffic jam 1 is in the build-up phase. Vehicles approaching traffic jam 1 are indicated in the diagram by a pictogram of a vehicle with an arrow.

[0039] A traffic jam is characterized by a downstream congestion front 4, which is initially represented by the first vehicles that brake and come to a standstill due to the event. As time progresses, more and more vehicles collide with these vehicles, so that an upstream congestion front 3 propagates along against the direction of travel F at a characteristic backpropagation speed.

[0040] Fig. Figure 1B shows the phase of the traffic jam in which the traffic jam 1 dissipates. Vehicles starting to move are also indicated here by a pictogram with an arrow. In the illustrated embodiment, the upstream traffic jam front 3 is essentially stationary, whereas the downstream traffic jam front 4 moves towards the upstream traffic jam front 3 at a characteristic backpropagation speed, contrary to the direction of travel F. The vehicles further ahead in the traffic jam 1 can gradually start moving, while the vehicles further back can only start moving at a later time.

[0041] The spatial and temporal extent of such a traffic jam 1 is described in the following figures using a trajectory map. On the abscissa of the in Fig. In diagram 2, time t is plotted on the ordinate, and distance d on road segment 2 is plotted on the ordinate. Each line drawn in the diagram represents the trajectory of an individual vehicle. The trajectories are inclined from the bottom left to the top right, meaning that as time t progresses, the respective vehicle travels the corresponding distance d on road segment 2. The flatter a trajectory curve, the slower the respective vehicle moves, until it comes to a standstill, which is indicated by a horizontal line. The steeper a trajectory line, the faster the vehicle travels.

[0042] How Fig. Figure 2 shows that the travel speed of the respective vehicles drops rapidly in the section of the trajectory map represented by the traffic jam (1) or the corresponding dotted lines. At the end of the traffic jam, the vehicles can accelerate again and continue their journey at the usual applicable speed limit. When the road is clear, all vehicles essentially assume the applicable speed limit, so the trajectories have the same slope in the diagram, meaning they run parallel to each other.

[0043] How Fig. Furthermore, as Kerner shows, a traffic jam is often followed by another area with a slower speed of movement, because the vehicles, which often accelerate rapidly, have to brake, thus leading to a renewed build-up of traffic. According to Kerner, this movement often occurs in a largely synchronous, flowing manner.

[0044] Furthermore, it shows Fig. 2 on the right-hand side shows the course of road section 2, including exits. Once a sufficient amount of time has passed since the traffic jam began, vehicles on road section 2 are usually informed about the upcoming traffic jam 1, allowing them to use the appropriate exit for a detour. This prevents the traffic jam from worsening. The relevant exit is located in Fig. 2 is indicated by a dashed line running horizontally through the diagram.

[0045] An inventive method for predicting the spatial and temporal extent of a traffic jam 1 on road section 2 provides that the vehicles of a fleet continuously transmit position information to a central computing unit. This unit determines, from the changes in position information at successive recording times, where each vehicle is moving on road section 2 and at what speed. As soon as one or more vehicles decelerate from essentially the applicable speed limit to a near standstill with a characteristic pattern, a traffic jam starting point P1 is detected. Fig. Figure 3 illustrates the procedure of the inventive method for predicting the spatial-temporal extent.

[0046] Solbad, once a corresponding starting point P1 of the impoundment is detected at time t1, the central computing unit begins predicting its spatial and temporal extent. The exact course of the impoundment area is not yet known. Therefore, initially defined parameters are used to estimate the corresponding area in the trajectory map. Assuming a fixed backpropagation velocity of the upstream impoundment front 3 against the direction of travel F and a fixed impoundment duration T12, a spatial and temporal impoundment endpoint P2 is determined. The impoundment endpoint P2 defines the location and time t2 at which the impoundment 1 essentially ceases to grow. The backpropagation velocity of the upstream impoundment front 3 is indicated in the trajectory diagram by the slope of the connecting line between the impoundment starting point P1 and the impoundment endpoint P2.The steeper this edge, the faster the upstream pressure front 3 propagates backwards against the direction of travel F. Therefore, the corresponding backpropagation speed is greater.

[0047] For the initial generation of the diagram, a fixed backpropagation velocity is assumed, specifically -12.0 km / h. It is assumed that the upstream impoundment front 3 propagates over the duration of the impoundment T12.

[0048] Furthermore, assuming a substantially stationary downstream congestion front 4 and a fixed congestion dissipation duration T13, a spatiotemporal congestion dissipation start point P3 is determined. The congestion dissipation start point P3 defines the location and time t3 at which the congestion 1 begins to dissipate. The congestion dissipation duration T13 is also fixed. Observations of historical congestion 1 are used in particular to determine the corresponding parameters.

[0049] Assuming a fixed backpropagation speed of the downstream congestion front 4 against the direction of travel F, a spatio-temporal congestion endpoint P4 is determined from the congestion dissipation point P3 until the upstream congestion front 3 reaches the position defined by the congestion endpoint P2. The congestion ends when the downstream congestion front 4 and the upstream congestion front 3 meet. Here, too, the corresponding parameters are fixed. Specifically, the corresponding backpropagation speed of the downstream congestion front 4 is -15.0 km / h. In this way, it is already possible at the time of the congestion start t1 to determine the Fig. The representation shown in Figure 3 can be generated. This provides a first approximation for the spatial and temporal extent of traffic jam 1.

[0050] Based on this representation, the data points from the vehicles received as the traffic jam progresses are further analyzed so that the actual backpropagation speed of the upstream traffic jam front 3 can be determined. Accordingly, the slope of the respective edge in the diagram is adjusted for later predictions. As soon as a characteristic kink in this line is detected, the presence of the traffic jam endpoint P2 is recognized. In other words, the instantaneous rate of change of the backpropagation speed of the upstream traffic jam front 3 exceeds a defined threshold. In particular, the backpropagation speed halves or slows down even further. Additionally or alternatively, the traffic jam endpoint P2 can also be recognized by the fact that the observed backpropagation speed of the upstream traffic jam front 3 falls below a certain threshold. In the Fig. In the embodiment shown in Figure 3, the backpropagation speed of the upstream impoundment front 3 from the impoundment endpoint P2 is 0 km / h, i.e., it is stationary.

[0051] As the traffic jam progresses, the prediction of its spatial and temporal extent becomes increasingly accurate. Eventually, the true backpropagation speed of the downstream jam front 4 can be determined, and a rapid increase in this speed allows the beginning of the jam's dissipation, P3, to be identified. By comparing the two jam fronts 3 and 4 and their respective backpropagation speeds (which, in this phase, is essentially 0 km / h for the upstream jam front 3), their intersection point can be found, thus predicting the end of the jam, P4. Once the two jam fronts 3 and 4 meet, the traffic jam 1 ends.

[0052] Fig. Figure 4 shows the prediction of the extent of the traffic jam 1, as made by the central computer, superimposed on the actual extent for an example case. The actual spatial and temporal extent of the traffic jam 1 is indicated by hatching. An envelope 5 between the starting point of the traffic jam P1, the end point of the traffic jam P2, the starting point of the traffic jam dissipation P3, and the end point of the traffic jam P4 defines the spatial and temporal extent of the traffic jam 1.

[0053] Fig. Figure 4A shows a superposition of the true extent of the reservoir with the initial approximation made for the extent of reservoir 1 at the reservoir's inception point P1. The actual spatiotemporal position of the reservoir's inception point P1 is known. The positions of reservoir's endpoint P2, reservoir's discharge point P3, and reservoir's endpoint P4 are estimated.

[0054] Fig. Figure 4B shows a superposition of the true extent of the reservoir with the approximation used for the reservoir endpoint P2. The actual spatiotemporal position of the reservoir start point P1 and the reservoir endpoint P2 is known. The position of the reservoir release start point P3 and the reservoir endpoint P4 is estimated.

[0055] Fig. Figure 4C shows a superimposition of the true extent of the reservoir with the approximation made to the reservoir release point P3. The actual spatiotemporal positions of the reservoir release point P1, the reservoir end point P2, and the reservoir release point P3 are known, while the reservoir end point P4 is estimated. At this stage, a very realistic estimate of the reservoir extent is already possible.

[0056] As can be seen, the later predictions of the extent of traffic jam 1 increasingly match the true extent.

[0057] In Fig. In 4D, all predictions are displayed in a superimposed representation.

[0058] Fig. Figure 5 illustrates the process of the method according to the invention once again using a flowchart. In step 501, the central computing unit recognizes the starting point of the traffic jam P1 from the data points transmitted by the vehicles, i.e., their respective position information and the associated recording times. In step 502, the central computing unit generates the following based on the defined parameters: Fig. 4A shows the initial approximation of the spatial and temporal extent of the traffic jam 1.

[0059] In step 503, the central computing unit checks whether an upstream starvation has been detected, i.e., whether the damming endpoint P2 has already occurred. If this is not the case, the location of the damming endpoint P2 is calculated in step 504 based on the aforementioned defined parameters: the defined damming duration T12 and the defined backpropagation velocity of the upstream damming front 3. Depending on the extent of the damming 1, the actual backpropagation velocity of the upstream damming front 3, derived from the transmitted data points, can be used instead of the defined backpropagation velocity. If the damming endpoint P2 can be identified from the underlying data, this occurs in step 505.

[0060] In step 506, the central computing unit checks whether the congestion resolution starting point P3 can be detected. If this is not the case, the position of the congestion resolution starting point P3 is estimated in step 507. For this purpose, the defined congestion resolution start time T13 is taken into account, after which a backpropagation of the downstream congestion front 4 with a defined backpropagation speed is assumed. If, on the other hand, a movement of the downstream congestion front 4 with a characteristic speed, in particular in the interval from -30.0 km / h to -2.0 km / h, most preferably -15.0 km / h, can be detected from the underlying data, the corresponding congestion resolution starting point P3 is detected in step 508.

[0061] Taking into account the specified backpropagation velocity or the actual backpropagation velocity of the downstream impoundment front 4, the intersection point with the upstream impoundment front 3 is calculated in step 509. This allows the location of the impoundment endpoint P4 to be estimated.

[0062] If points P1 to P4 are available (in estimated or observed form), the central computing unit calculates the envelope 5 in step 510 and can thus output the spatiotemporal extent of the traffic jam 1. The corresponding result is provided in step 511.

[0063] In step 512, the estimated spatial and temporal extent of traffic jam 1 can be used, for example by a navigation system to predict the travel time through traffic jam 1 or to provide an early alternative route around traffic jam 1.

[0064] As indicated by an arrow pointing back to step 503, 506 or 509, the central computing unit continuously re-estimates the extent of the traffic jam 1 during the duration of the traffic jam in order to track or realistically estimate the spatial and temporal extent of reality. QUOTES INCLUDED IN THE DESCRIPTION

[0000] 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 non-patent literature

[0000] L. Querfurth et al., Crash localization and traffic impact assessment via spatio-temporal analysis of connected vehicle data, Accident Analysis and Prevention 214 (2025) 10795, 17. Februar 2025, https: / / doi.org / 10.1016 / j.aap.2025.107956

[0006]

Claims

[1] Method for predicting the spatiotemporal extent of a traffic jam (1) on a road segment (2), wherein vehicles of a fleet continuously transmit position information describing their current location at a given time of acquisition to a central computing unit, and the central computing unit determines, by establishing a relationship between the respective position information and the times of acquisition, which vehicle is moving at which location at which speed on the road segment (2) at which time (t), wherein the spatiotemporal starting point (P1) of the traffic jam is determined by identifying at least one vehicle which reduces its speed from essentially the applicable speed limit to essentially standstill following a defined pattern, characterized by , that Assuming a fixed backpropagation speed of an upstream congestion front (3) against the direction of travel (F) and a fixed congestion duration (T12), a spatial-temporal congestion endpoint (P2) is determined, defining the location and time (t2) on the road section (2) at which the congestion (1) essentially no longer increases; Assuming a substantially stationary downstream congestion front (4) and a defined congestion dissolution start duration (T13), a spatial-temporal congestion dissolution start point (P3) is determined, defining the location and time (t3) on the road section (2) at which the congestion (1) begins to dissipate; Assuming a fixed backpropagation speed of the downstream congestion front (4) against the direction of travel (F) from the congestion dissipation starting point (P3) until reaching the position of the upstream congestion front (3) defined by the congestion end point (P2), a spatial-temporal congestion end point (P4) is determined, whereby the envelope (5) between the starting point of the congestion (P1), the end point of the congestion (P2), the starting point of the congestion dissipation (P3) and the end point of the congestion (P4) describes the spatial and temporal extent of the congestion (1), and wherein the central computing unit continuously re-estimates the envelope (5) during the duration of the traffic jam (1) taking into account the data points received from the vehicles during this time and the resulting actual location of the end point of the traffic jam (P2), the start point of the traffic jam (P3) and the end point of the traffic jam (P4). [2] Method according to claim 1, characterized by, that the magnitude of the specified backpropagation velocities of the upstream congestion front (3) and the downstream congestion front (4) as well as the specified congestion duration (T12) and the specified congestion dissipation start duration (T13) are determined based on empirically determined historical data points of past congestions (1). [3] Method according to claim 2, characterized by , that depending on the pattern with which at least one vehicle reduces its speed to detect the starting point of the traffic jam, different characteristics are determined for the specified backpropagation speeds of the upstream traffic jam front (3) and the downstream traffic jam front (4) as well as the specified traffic jam duration (T12) and the specified traffic jam resolution start duration (T13). [4] Method according to any one of claims 1 to 3, characterized by , that the magnitude of the specified backpropagation velocity of the upstream impoundment front (3) is -12.0 km / h; and / or the magnitude of the specified backpropagation velocity of the downstream storm front (4) is -15.0 km / h. [5] Method according to any one of claims 1 to 4, characterized by , that the central computing unit for detecting the actual location of the damming endpoint (P2) takes into account whether an instantaneous rate of change of the backpropagation velocity of the upstream damming front (3) exceeds a specified threshold. [6] Method according to any one of claims 1 to 5, characterized by , that the central computing device for detecting the actual location of the damming endpoint (P2) takes into account whether the instantaneous backpropagation velocity of the upstream damming front (3) falls below a specified threshold. [7] Method according to any one of claims 1 to 6, characterized by, that the central computing unit for detecting the actual location of the congestion release point (P3) takes into account whether the instantaneous backpropagation speed of the downstream congestion front (4) falls below a specified threshold, in particular a threshold within the closed interval of -30.0 km / h to -2.0 km / h. [8] Method according to any one of claims 1 to 7, characterized by , that the central computing unit for predicting the location of the damming endpoint (P4) predicts when and where the upstream damming front (3) propagating from the damming endpoint (P2) with its backpropagation speed and the downstream damming front (4) propagating from the damming dissipation startpoint (P3) with its backpropagation speed will meet. [9] Method according to any one of claims 1 to 8, characterized by, that a navigation system used for route guidance uses the currently predicted spatial and temporal extent of the traffic jam (1) by the central computing unit to predict the estimated travel time of a navigation route passing through the traffic jam (1) for a vehicle. [10] Method according to any one of claims 1 to 9, characterized by , that a navigation system used for navigation route guidance takes into account the current spatial and temporal extent of the traffic jam (1) predicted by the central computing facility for the calculation of the navigation route. [11] Information technology system, characterized by a central computing device set up for predicting the spatial and temporal extent of a traffic jam (1) using a method according to one of claims 1 to 10.

Images