Method for determining an indicator value for statistical characterization of event descriptions detected by vehicles in location-related manner

Abstract

The invention relates to a method for determining an indicator value (Z) for statistical characterization of event descriptions detected by at least one vehicle in location-related manner, wherein, on the basis of a road or navigation map, those inflow- and outflow-free road sections are identified for which the section length thereof is greater than or equal to a predetermined minimum section length. Identified mutually adjacent road sections are in each case connected to form a section chain (10) with a start point (11) and an end point (12). Event descriptions, comprising a geoposition (P1, P2) and at least one parameter of a vehicle environment model, are detected by vehicles and transferred to an analysis system. The analysis system in each case assigns a nearest road network location (O1, O2) of a section chain (10) to a geoposition (P1, P2) of an event description and describes it by a longitudinal distance (L, L1, L2) relative to the start point (11) of the section chain (10). For at least one section chain (10), a profile of a density value (p) of event descriptions related to the longitudinal distance (L, L1, L2) to the start point (11) is determined using a kernel density estimator of predetermined bandwidth. The profile of the density value (p) is transformed using local, bandwidth-related, Z-standardisation into a Z profile (Z1, Z2) of a Z value (Z) which is mean-free in relation to bandwidth.

Description

[0001] Method for determining an indicator value for the statistical characterization of location-related event descriptions recorded by vehicles

[0002] The invention relates to a method for determining an indicator value for the statistical characterization of event descriptions recorded in a location-related manner by at least one vehicle according to the preamble of claim 1.

[0003] Document DE 102022 105 919 A1 describes a system and method for the early detection of structural hazards in road traffic using a digital traffic network image or map. For this purpose, accident data, user input data, and sensor data are assigned to map segments using a computer system according to their georeferencing. According to one aspect, the computer system performs an evaluation, comprising at least a frequency determination and / or a data comparison.For the purpose of early detection, a georeferenced segment to which no accident data is assigned is identified as a potential hazard location if sensor data or user input data have been assigned with a predetermined frequency, and / or if a data comparison, in particular a comparison with an artificial intelligence (AI)-based model, of user input data and / or sensor data shows that a predefined degree of feature match or feature correlation with critical patterns exists. This enables early detection of hazard locations. According to a further aspect, it is proposed to determine a hazard score for identified hazard locations.

[0004] Document US 2023 / 0245560 A1 describes a system for determining and sorting location-related hazards based on vehicle-recorded events and / or an accident database. The system comprises an interface and a processor. The interface is configured to receive accident data and / or event data. The processor is configured to determine groups of incidents, with each group of incidents being assigned a set of accidents from the accident database and / or a set of events from the event data, grouped according to their geographical proximity. The processor is further configured to determine a set of road traffic estimates that are related to a group of incidents, and to determine an order of the groups of incidents based on the set of accidents, the set of events, and the set of road traffic estimates.

[0005] Document DE 102019215 587 A1 describes a method for calculating relative positive or negative accelerations (RPA, RNA) using crowdsourcing, comprising the following steps: a) generating a digital map in which different route sections are connected by nodes; b) calculating a relative positive or negative acceleration (RPA, RNA) of a vehicle traveling in a route section for this route section (A, B, C,...)., M), wherein the relative positive or negative acceleration (RPA, RNA) is calculated in the vehicle and sent to a backend, or wherein the vehicle sends data comprising current geocoordinates, current vehicle speeds, and / or current vehicle accelerations to a backend, which calculates therefrom the relative positive or negative acceleration (RPA, RNA) of the vehicle for this route section; and c) calculating an average or a distribution of the calculated relative positive or negative accelerations (RPA, RNA) of several vehicles for this route section in the backend.

[0006] Document DE 102019000630 A1 discloses a method for assigning a common geoposition to a plurality of events caused by a common root cause and recorded by at least one vehicle. A single geoposition is recorded for each individual event. The recorded events are indexed in a lexicographically sortable manner. Geopositions to which a plurality of recorded events are assigned are referred to as hotspots. A method is proposed for determining maximum hotspots by combining spatially adjacent and sufficiently close hotspots. Hotspots assigned a set of events that is a true subset of the set of events assigned to another hotspot are eliminated. Hotspots merged into a maximum hotspot in this way are assigned a common geoposition.Document WO 2021191051 A1 discloses a method for identifying potential hazards in road traffic using connected vehicles. An event indicating a potential hazard is recorded and transmitted to a central computer unit along with its geoposition. The transmitted event is entered into a digital map as a hotspot if a large number of similar events with the same geoposition are recorded. Context information is added to the transmitted event. The hotspot is analyzed to identify a potential hazard. A current hotspot is compared with confirmed hotspots. The hotspots are visualized on a platform. The geoposition of a hotspot of certain traffic-critical events is transmitted to vehicles located near this hotspot.

[0007] A hotspot, as known from the state of the art, can thus be understood as a point-like (locally limited) cluster of events assigned to the same or a very similar geolocation. Such a cluster (or hotspot) is characterized by an absolute frequency and a maximum radius. The absolute frequency indicates how often (within a predetermined period of time) an event was recorded within the maximum radius around the hotspot.

[0008] This well-known form of describing a hotspot has the disadvantage that traffic-relevant events typically do not occur and are recorded in a circular area around a geoposition, but rather along traffic routes.

[0009] Furthermore, the analysis of such hotspots is made more difficult by the fact that it is often not the absolute but the relative frequency of an event (related to the number of vehicle passages at the geoposition) that is relevant for an assessment, because it estimates the probability of the event occurring when passing through this geoposition.

[0010] However, since the number of vehicle passages for the geoposition of a hotspot is typically not known or only known inaccurately for technical reasons and due to restrictions imposed by data protection, the relative frequency of an event cannot be determined or can only be determined roughly.

[0011] Therefore, there is a need for an improved method for the detection, classification and analysis of hotspots characterized by the conspicuousness of an indicator value that characterizes event descriptions recorded at such hotspots.

[0012] The publication E. Schubert, A. Zimek, H.-P. Kriegei: Generalized Outlier Detection with Flexible Kernel Density Estimates. Proceedings of the 14th SIAM International Conference on Data Mining (SDM), Philadelphia, PA, 2014, pp. 542–550, https: / / doi.org / 10.1137 / 1.9781611973440.63 describes a method for analyzing the relationship between the estimation of a density function and the identification of outliers based on such a density function.

[0013] The invention is based on the object of providing an improved method for determining an indicator value for the statistical characterization of event descriptions recorded in a location-related manner by at least one vehicle.

[0014] The object is achieved according to the invention by a method having the features of claim 1.

[0015] Advantageous embodiments of the invention are the subject of the subclaims.

[0016] To detect hotspots, parameters of the vehicle's environment are continuously recorded in at least one vehicle. For example, a dynamic vehicle environment can be recorded by measuring distances (relative to the vehicle's own position), speeds, and trajectories of surrounding vehicles, people, and / or other road users using a stereo camera and / or a radar and / or lidar device or similar measuring methods. Additionally or alternatively, a static vehicle environment can be recorded using the vehicle sensors, for example, the arrangement and properties of lane markings and / or traffic signs and / or traffic guidance devices. For example, lane markings can be recorded and measured with regard to their spacing, lane marking types, lane widths, curvatures, and similar parameters.

[0017] In parallel with the recording of such vehicle environment parameters, the vehicle's position, i.e., the geoposition recorded using a global navigation satellite system (GNSS) or a similar geopositioning system, is continuously determined. The objects and their properties detected in the static and / or dynamic vehicle environment using various sensors are merged into an environment model, which can also include map material from a road or navigation map and, optionally, additional dynamic, geoposition-related live data, such as current traffic volume.

[0018] Triggers are defined on this complex environmental model that describe events of note (particularly traffic-relevant) for driving the vehicle. Such events can be, for example, warnings or interventions from an assistance system that monitors the vehicle's surroundings. When such a trigger is triggered, an event description is generated and sent by the respective vehicle to a backend of an analysis system. The event description includes the respective geoposition at which the trigger was triggered, as well as the sensor data and parameters recorded by the vehicle's sensors that are relevant to describing the event.

[0019] The analysis system can, for example, be implemented on a server or in a cloud and provide as a backend a set of predetermined services and / or functions in the form of an application programmer interface (API).

[0020] The analysis system generates an analysis map from the event descriptions transmitted by one or more vehicles to the backend based on a road or navigation map or similar map material. In such map material, traffic routes, especially roads, are assigned geopositions, which are referred to below as road network locations. In other words, while geopositions can generally be assigned to any topographical feature, road network locations refer to those geopositions that are accessible via a road.

[0021] First, the longest possible, continuous road sections that are not interrupted by intersections are identified. In particular, road sections whose length reaches or exceeds a predetermined minimum length can be selected for the analysis map. Typically, such road sections are assigned to highways or expressways that have few intersections. Road sections shorter than the predetermined minimum length are filtered out and not included in the analysis map.

[0022] Adjacent road sections are then linked (like a linked list) to form section chains. In other words, a topologically one-dimensional (i.e., only linearly extended) section chain is formed from an ordered sequence of road sections, which has a single starting point and a single end point along its linear extent. Each section chain is assigned a unique chain identifier and a starting point (relative to the analysis map). Thus, every location along a section chain can be uniquely determined by the chain identifier and a distance from the starting point of the section chain.

[0023] In a method for determining an indicator value for the statistical characterization of location-specific event descriptions recorded by at least one vehicle, road sections without inflow and outflow are identified based on a road or navigation map or similar map material in which traffic routes are mapped with reference to geopositions. Along such road sections, the number of passing vehicles can be assumed to be approximately the same everywhere. In this way, longer traffic routes are divided into road sections that are connected to at least one other road section at intersections, junctions, exits, entrances, and similar junctions.

[0024] According to the invention, among the road sections thus determined, those whose section length is equal to or greater than a predetermined minimum section length are identified. This makes it possible to select road sections exclusively, or at least essentially only, along certain types of traffic routes for which the assumption of a substantially equal vehicle throughput (along the respective traffic route) applies to a good approximation. In particular, the section length can be selected such that the selected road sections are assigned to motorways or expressways.

[0025] In a subsequent step, among the identified road sections, those are determined which are connected to one another along a traffic route via an intersection, junction, exit, entrance or similar junction and for which the geoposition of an end of a first identified road section therefore corresponds to the geoposition of an end of a second identified road section.

[0026] A chain of such road sections, connected in the manner of a linked list, is called a section chain. A section chain extends linearly from a starting point to an end point and therefore has no branches along which a traffic flow could be distributed. Therefore, the assumption of a uniform traffic flow applies to such section chains with a particularly good approximation.

[0027] At least one event description is recorded from at least one vehicle and transmitted to an analysis system. Such an event description includes at least one parameter of a vehicle environment model, which includes the static and / or dynamic environment of the vehicle as well as operating and / or control parameters of the vehicle. Furthermore, such an event description includes the geolocation at which these parameters were recorded.

[0028] Event descriptions are recorded and transmitted in particular when the vehicle environment model describes a condition or a change in condition that is particularly important for the assessment of traffic flow and / or vehicle condition.

[0029] The analysis system evaluates incoming and / or stored event descriptions. The analysis system assigns a geoposition of an event description, which typically does not exactly correspond to a road network location due to tolerances and inaccuracies in the recording by a geopositioning system, to the nearest road network location in a section chain, typically by (perpendicular) projection onto the nearest section chain.

[0030] This nearest road network location is described by a unique identifier of the section chain and by a longitudinal distance that specifies the distance (in predetermined length units chosen the same for all section chains) from the starting point of the section chain. Furthermore, the analysis system uses a kernel density estimator to determine a progression of a density of event descriptions for at least one section chain based on the longitudinal distance to the starting point. For this purpose, densities (density values) are determined at incremental, preferably equidistant, longitudinal distances from the starting point of the section chain. These densities estimate the frequency of event descriptions within an increment of the longitudinal distance.By using a kernel density estimator, the scatter of these determined values ​​can be reduced by taking into account frequencies of event descriptions from neighboring increments of the longitudinal distance when determining the density in an increment (in the manner of a weighted, moving average).

[0031] The density is determined at a specific location, determined by the longitudinal distance from the starting point of the segment chain, by subjecting the longitudinal distances (L, L1, L2) assigned to the event descriptions by projection of this segment chain to a kernel density estimation using a kernel density estimator with a predetermined bandwidth. In this process, a sequence of Dirac pulses arranged at the assigned longitudinal distances is convolved with a non-negative, preferably symmetric kernel weighting function.

[0032] Methods for selecting the kernel weighting function, in particular for selecting the bandwidth of such a kernel weighting function, are known from the prior art, for example from the publication Wand, MP; Jones, MC (1995). Kernel Smoothing. London: Chapman & Hall / CRC. ISBN 978-0-412-55270-0.

[0033] Each density of event descriptions determined along a section chain is then transformed into a Z-value using a local Z-standardization.

[0034] A local Z-standardization is understood here and in the following as the transformation of a density value pt, which was determined for one or for a set of road network locations, taking into account a totality of density values ​​Pi,p2< ■■■Pi - PN>, which were determined within a predetermined area, referred to as a bandwidth, along the respective section chain around the road network location of the density value pt. For this totality of density values, a (local) mean and a (local) standard deviation <7® is determined. The local Z-standardization transforms the density value pt into a Z-value Z^ according to the rule

[0035] In other words, with local Z-standardization, the density value is converted into a Z-value, which represents the deviation of the density from the mean relative to the standard deviation in relation to the surrounding area, known as the bandwidth.

[0036] The higher the local Z-value, the more conspicuous the observed point is compared to the surrounding area, measured by the variability of the surrounding area, regardless of the absolute density. A higher Z-threshold can reduce the result set to more conspicuous locations.

[0037] This makes it possible to characterize and identify road network locations as conspicuous based on the Z-value, especially those road network locations where the probability of detecting / reporting an event description is significantly higher than at road network locations in a predetermined vicinity along the section chain. An advantage of the method over the state of the art is that it enables a reliable characterization of such anomalies, regardless of the absolute (total) number of vehicles passing a road network location.

[0038] This advantage is achieved, on the one hand, by selecting road network locations with comparatively homogeneous vehicle density through the inventive formation of section chains. On the other hand, this advantage is achieved by transforming density values ​​into an indicator using local Z-standardization that is largely insensitive to fluctuations in the dispersion of density values ​​along a section chain.

[0039] In one embodiment of the method, the predetermined minimum length for road sections to be used as the basis for forming section chains is selected such that the at least one section chain is formed exclusively or predominantly from road sections assigned to a motorway or expressway. This achieves particularly good homogeneity of traffic flow along the formed section chains. In other words, due to the comparatively low inflows and outflows, density values ​​recorded along such section chains are particularly easy to compare; in particular, increased density values ​​of event descriptions particularly reliably indicate an exceptional traffic situation.

[0040] In one embodiment of the method, the vehicle environment model merges a static and a dynamic vehicle environment. The term "static vehicle environment" refers to the vehicle environment that is essentially constant over time, particularly one that is unaffected by the activities of other road users, such as the arrangement and properties of lane markings and / or traffic signs and / or traffic guidance devices. A static vehicle environment can be captured, for example, using surround-view cameras, distance sensors, radar, and lidar sensors.

[0041] "Dynamic vehicle environment" refers to the vehicle environment that changes over time, in particular the vehicle environment that changes due to the movement of other road users. For example, a dynamic vehicle environment can be captured by measuring distances (relative to the vehicle's own position), speeds, and trajectories of surrounding vehicles, people, and / or other road users using a stereo camera and / or a radar and / or lidar system or similar measurement methods.

[0042] In one embodiment of the method, a peak region is identified as an area in which, along a chain of sections between a peak start and a peak end, connected road network locations have been assigned a Z-value that exceeds a predetermined threshold. This allows areas with conspicuous traffic situations, such as accident blackspots, to be identified with particularly high reliability.

[0043] Preferably, the predetermined threshold value is selected to be one. Based on statistics of the density values ​​that approximately follow a normal distribution, those areas are identified as peak regions that have a higher density of event descriptions than approximately 85 percent of the total road network locations in a section chain. This enables particularly reliable determination of peak regions with conspicuous traffic situations. In one embodiment of the method, a peak region is assigned an identifier of the respective section chain (which is uniquely assigned, for example, as a number or character string), a longitudinal distance of the peak start relative to the start point of the respective section chain, a longitudinal distance of the peak end relative to the end point of the respective section chain, a maximum value of all Z-values ​​within the peak region, and a description of the geometry of the peak region.A description of the geometry of the peak region can, for example, be specified as a set of geolocations assigned to all or some (e.g., equidistantly selected) road network locations of the peak region. Using such features, the evaluation of peak regions is particularly easy and reliable.

[0044] In one embodiment of the method, the vehicle environment model comprises at least one object and / or one parameter detected by vehicle sensors. Based on the vehicle environment model, a trigger is determined that identifies and / or marks an event worth considering, for example, an event relevant to the control and guidance of the vehicle. Based on such a trigger, an event description is created and transmitted to the analysis system.

[0045] With such a trigger, event descriptions can be created that are interesting and relevant for subsequent analysis.

[0046] Preferably, a trigger marks a warning issued by a driver assistance system and / or an intervention by such a driver assistance system. This allows driving situations in which the driver's behavior potentially requires correction to be captured in event descriptions. Determining a statistical indicator for such driving situations based on road network locations (i.e., specific locations in road section chains) enables a particularly meaningful assessment of potentially dangerous locations and the detection of corresponding peak regions, for example, the identification of potential accident blackspots. This allows for the identification of clues that contribute to improving road safety.

[0047] In one embodiment of the method, event classes are defined and recorded in a list of event classes. Event classes can be (by way of example only) a braking maneuver initiated by a driver assistance system, a vehicle parameter indicating reduced vehicle stability, or a potential collision with an object in the vehicle's surroundings detected by a driver assistance system. At least one such event class is assigned to each event description. To determine the density of event descriptions, only those (filtered) event descriptions whose event classes satisfy a specific filter criterion are used. For example, only event descriptions from a specific event class or from a subset of event classes are used. This enables a particularly specific evaluation of density values ​​and / or a particularly specific determination of peak regions.

[0048] In one embodiment, an event description comprises a direction of travel, which refers to a direction of a section chain pointing from its starting point to its end point. To determine the density of event descriptions, only those event descriptions are used for which the direction of travel is the same, i.e., for which the vehicle was moved in the same direction (along the section chain). This makes it possible to particularly specifically detect direction-related anomalies (for example, differentiated between a downhill and an uphill journey on a road section with a steep gradient, or differentiated according to other hazards that vary depending on the direction of travel, for example, due to obscured views, on-ramps, off-ramps, or curve radii).

[0049] In one embodiment of the method, Z-values ​​determined for road network locations along at least one section chain are visualized in a map display. The visualization can be achieved, for example, by color representation by assigning different colors to different ranges of Z-values. Additionally or alternatively, Z-values ​​can be visualized by markers applied perpendicular to the respective section chain, wherein the length of a marker applied at a road network location is selected depending on the Z-value determined for this road network location. The length of such a marker is preferably selected proportional to the Z-value.

[0050] This enables a particularly clear representation of gradual differences in abnormalities, especially hazards, along a chain of sections, without the need for arbitrary thresholding and segmentation of peak regions. Exemplary embodiments of the invention are explained in more detail below with reference to drawings.

[0051] Showing:

[0052] Fig. 1 schematically shows an analysis card with a section chain,

[0053] Fig. 2 schematically shows the determination of a Z-standardized Z-value for a density of event descriptions determined by means of a kernel density estimation and

[0054] Fig. 3 shows a schematic representation of Z-values ​​along a section chain depending on the direction of travel.

[0055] Corresponding parts are provided with the same reference numerals in all figures.

[0056] Figure 1 schematically shows an analysis map with a section chain 10. The section chain 10 was formed from the concatenation of intersection-free road sections (not shown in detail), whose length reaches or exceeds a predetermined minimum length. Each of these road sections (not specifically designated in Figure 1) comprises at least one, typically a plurality of, road network locations 01, 02, which can, in principle, be located as densely as desired (limited only by the resolution of the underlying map material). For improved clarity, only a first and second road network location 01, 02 are shown as examples in Figure 1.

[0057] The section chain 10 can be obtained, for example, by analyzing a road or navigation map. It is shown in Figure 1 in its position and extent relative to a geocoordinate system, which is represented in a highly simplified manner and schematically comprises a longitudinal coordinate (geographical longitude), represented here as the x-direction x, and a latitudal coordinate (geographical latitude), represented here as the y-coordinate y.

[0058] The section chain 10 has a starting point 11 at a first end and an end point 12 at the opposite second end. It is identified by an identifier, which is chosen here only as an example as the character string "#2." At least the starting point 11 and the end point 12 can each be assigned to both a road network location (not specified in this case) and a geoposition (also not specified in this case).

[0059] Each point along the section chain 10 can be uniquely identified by a longitudinal distance L from the starting point 11 (specified in arbitrarily selectable distance units that are the same for all section chains 10) and by the identifier of the section chain 10.

[0060] Each point of a section chain 10 can thus be assigned a road network location 01, 02 on a road section of the underlying road or navigation map according to its distance from the starting point 11. For example, the sequence of the assembled road sections of the section chain 10 can be traced, starting from the starting point 11, until the longitudinal distance L corresponding to the respective point from the starting point 11 is reached. Thus, by referring to the geoposition of the starting point 11, each point of a section chain 10 can also be assigned coordinates in the geocoordinate system.

[0061] For clarity, Figure 1 shows two geopositions P1, P2 determined by one or more vehicles using a geopositioning system and transmitted with an event description. For example, a braking action could have been triggered by a driver assistance system at each of these geopositions P1, P2. Due to inaccuracies in the geopositioning system, the geopositions P1, P2 reported by different vehicles typically deviate from the course of the section chain 10 (i.e., the set of road network locations 01, 02 of this section chain 10 in the geocoordinate system).

[0062] Each geoposition P1, P2 is assigned a nearest road network location 01, 02 along the section chain 10. Typically, the nearest location can be uniquely determined by vertical projection onto the section chain 10. If there are multiple nearest road network locations 01, 02 that are equally distant, a road network location 01, 02 can be randomly selected.

[0063] By way of example, Figure 1 shows a first geoposition P1, which is described by a first pair of geocoordinates x1 (along the x-direction x) and y1 (along the y-direction y). A first road network location 01 on the section chain 10 is assigned to the first geoposition P1, which is located at a first longitudinal distance L1 from the starting point 11. For example, the first longitudinal distance L1 is assumed to be 5.2 length units. The first geoposition P1 can then be described by the identifier "#2" of the section chain 10 and by the longitudinal distance L1 "5.2 length units" from the starting point 11.

[0064] In the analysis, an event description reported for the first geoposition P1 is assigned the first road network location 01 and the lateral transverse distance d1 of the first geoposition P1 from the nearest first road network location 01, which in this case is "1.0 length units." For example, the entirety of the information on the identifier of the assigned section chain 10 as well as on the location and lateral distance of the nearest road network location 01 can be summarized in a character string "#2, 5.2, +1," which is assigned to the first geoposition P1. Generally, this assignment will not be unique; that is, there may be event descriptions at other (different) geopositions that are assigned the same character string.

[0065] In an analogous manner, the second geoposition P1 is described by a second pair of geocoordinates x2 (along the x-direction x) and y2 (along the y-direction y). The second geoposition P2 is assigned the second road network location 02 as the nearest point of the section chain 10. Furthermore, the second geoposition P2 is assigned a descriptive character string "#2, 4.1, +1" corresponding to the second longitudinal distance L2 of the nearest road network location 02 from the starting point 11 (in this case, "4.1 length units") and the lateral transverse distance d2 between the road network location 02 and the second geoposition P2 (in this case, "1.0 length units").

[0066] With this assignment rule, explained here only as an example, each geoposition P1, P2 can be assigned a point in a generally curvilinear coordinate system defined by the family of section chains 10. Thus, the location determination, relative to a nearest section chain 10, can be reduced to specifying the longitudinal distance L relative to its starting point 11.

[0067] By performing all further evaluations per section chain 10, the complexity of the acquired data can be reduced from two dimensions (coordinates along the x-direction x and the y-direction y) to only one dimension (longitudinal distance from the starting point 11 of the respective section chain 10).

[0068] A density distribution is calculated from the event descriptions recorded along a section chain 10 by applying methods known from the prior art as kernel density estimation (KDE) based on the longitudinal distances L between the road network locations 01, 02 assigned to the respective geoposition P1, P2. For this purpose, a density p of event descriptions is determined for a distance chain 10 at regular intervals, for example, starting from the starting point 11 at longitudinal intervals L of one meter each. This density is schematically illustrated in the upper part of Figure 2.

[0069] The density p is determined by subjecting the longitudinal distances (L, L1, L2) assigned to the event descriptions by projection onto the nearest section chain (10) to a kernel density estimation using a kernel density estimator with a predetermined bandwidth. A sequence of Dirac pulses arranged at the assigned longitudinal distances (L, L1, L2) is convolved with a non-negative, preferably symmetric kernel weighting function.

[0070] A density p can be determined based on one or more classes of event descriptions. For example, one class of event descriptions (also referred to as an event class) can be defined as braking interventions by a driver assistance system (e.g., an anti-lock braking system), and another event class can be defined as an abruptly reduced vehicle speed.

[0071] The sum of the density values ​​p thus determined across the section chain 10 is thus equal to the number of data points used (i.e., the total number of event descriptions of the respective event class(es) recorded across the section chain 10). This allows for a comparison of different road sections, a comparison of density values ​​p determined based on different data sources, and a weighting of such density values ​​p from different data sources.

[0072] Typically, the density p of event descriptions at some road network locations O1, O2 is considerably higher than the statistical average. Such particularly conspicuous locations are referred to below as peak regions R1, R2, R3. A peak region R1, R2, R3 can be caused, for example, by confusing traffic control, a particular weather exposure (e.g., a tendency toward black ice or fog formation, or ruts that increase the risk of aquaplaning in rainy weather), or by structural features of the traffic layout (lane narrowing, surface change, road damage, or similar).

[0073] To identify peak regions R1, R2, and R3, a classification of road network locations O1, O2 or of areas comprising several such road network locations O1, O2 is performed. Such a classification is also referred to as scoring.

[0074] The invention is based on the finding that scoring based on a constant threshold value e, that is to say: the classification of a road network location 01, 02 as a peak region R1, R2, R3 when the density p of event descriptions recorded there (for example: number of braking operations recorded within a radius of one meter around this road network location 01, 02) exceeds the predetermined constant threshold value e (for example: 100 braking operations), does not function satisfactorily.

[0075] In particular, such scoring requires that the number of vehicle crossings along the respective road section is known. However, this requirement is hardly feasible in practice for the technical and / or organizational-legal reasons already explained. With state-of-the-art methods, this leads to heavily trafficked road network locations 01, 02 being marked as suspicious, even though there is no increased risk of intervention by a driver assistance system, for example.

[0076] To circumvent this problem, the invention is based on the assumption that the vehicle density along a section chain 10, particularly between road network locations 01, 02 in spatial proximity, fluctuates only slightly. This assumption is particularly plausible for section chains 10 designed as highways or expressways, since for such traffic routes, a change in the number of passing vehicles is only possible at entrances and exits. In particular, this allows for good comparison of scoring values ​​between neighboring entrances or exits.

[0077] It is proposed to determine a scoring value, referred to as a Z-value or Z-score Z, for each data point. For this purpose, the deviation from a mean density p is determined in a symmetrical influence area, referred to as the bandwidth, around a road network location 01, 02 and related to a standard deviation a of the density p, also determined in this symmetric influence area.

[0078] To illustrate this procedure, it is assumed that along a road section with an approximately constant number of passing vehicles at a first to N-th road network location 01 , 02 there is a density value p1,p2, ■■■Pt - p Nof event descriptions, each reported within a predetermined bandwidth around the road network location 01, 02. From this, an empirical mean (sample mean) p^ and an empirical standard deviation <7® are calculated. Only event descriptions from those road network locations 01, 02 that lie within the predetermined bandwidth around the road network location 01, 02 for which scoring is to be performed are included in the calculation of p^ and a®. In other words: p^ and a® are calculated as a moving mean and a moving standard deviation a, respectively, over a range within the bandwidth around the respective density value p t certainly.

[0079] It is possible that the moving average p^ and / or the moving standard deviation o-® are determined by weighted averaging, in which densities p measured further away from the road network location 01, 02 are weighted less heavily than densities p measured closer. For example, a weighting of the neighboring densities p can be applied that decreases linearly with the distance from the respective road network location 01, 02.

[0080] The course of the mean value p and the standard deviation a along the section chain 10 is shown in the upper part of Figure 2 in addition to the course of the density p.

[0081] Using the mean p^ and the standard deviation is assigned to each of the density values ​​p t a standardized density value or Z-value Z is assigned: In other words, the respective density value pt of event descriptions is subjected to a statistical Z-standardization that is limited to its area of ​​influence (its bandwidth). Within this bandwidth, the determined Z-value Z is mean-free and can have a standard deviation a of 1. This allows Z-values ​​Z to be easily comparable even if they were determined for road sections with different conditions (especially with different traffic densities).

[0082] In Figure 2, the course of the Z value Z along a section chain 10 as a function of the route length L is shown in the middle area.

[0083] Using the Z-value Z, peak regions R1, R2, R3 can be particularly easily identified as statistically significant areas of a section chain 10. For example, a peak region R1, R2, R3 can be identified as a contiguous area along the section chain 10 in which the Z-value Z exceeds a predetermined threshold value 6.

[0084] As an example, a threshold value 0 = 1 is shown in the middle section of Figure 2. In other words, exceeding this threshold value 0 = 1 indicates that the density value p has deviated from the (moving) mean value p by more than the (moving) standard deviation a. However, other threshold values ​​0 can also be used.

[0085] Based on a threshold value exceedance determined in this way, the course of the Z-value Z can be transformed into a binary scoring value S:

[0086] 5> rl for Z t > 0 1 to otherwise

[0087] In the lower part of Figure 2, the course of the scoring value S along the section chain 10 is shown as an example (for 0 = 1). Based on the scoring value S, a first to third peak region R1, R2, R3 are determined.

[0088] According to the invention, a peak region R1, R2, R3 is assigned: the identifier of the respective section chain 10 (in the present case, for example, the character string “#2”), the peak start Lmin as the point of the peak region R1, R2, R3 closest to the starting point 11 and the peak end Lmax as the point of the peak region R1, R2, R3 furthest away from the starting point 11.

[0089] For improved clarity, the peak start Lmin and the peak end Lmax in Figure 2 are marked only for the second peak region R2.

[0090] Furthermore, a peak region R1, R2, R3 can be assigned an explicit length and / or a maximum value and / or a geometric description. The length specifies the distance between the peak start Lmin and the peak end Lmax. The maximum value specifies the highest density value p and / or the highest Z value Z in the peak region R1, R2, R3. The geometric description specifies the arrangement and position of the peak region R1, R2, R3 in the geocoordinate system, for example, according to the longitude and latitude of the center of the peak region R1, R2, R3.

[0091] Figure 3 schematically shows, in a simplified geocoordinate system, the course of a section chain 10 with several road sections connected by intersections, junctions, exits, ramps, and similar nodes. On both sides of the section chain 10, on the right-hand side in a direction of travel 11, 12, a Z-course Z1, Z2 of Z-values ​​Z is indicated by markings M perpendicular to the section chain 10.

[0092] The lengths of the vertical markings M assigned to the first Z-path Z1 indicate location-dependently varying Z-values ​​Z that were recorded corresponding to a first direction of travel 11. Accordingly, the lengths of the vertical markings M assigned to the second Z-path Z2 indicate Z-values ​​Z that were recorded by vehicles moving in the opposite second direction of travel 12.

[0093] In addition to the determination of peak regions R1, R2, R3 based on the exceedance of a threshold value 6, as explained in Figure 2, a representation according to Figure 3 enables a particularly clear representation of route sections with an increased number of event descriptions.

Claims

Patent claims 1. Procedure for determining an indicator value (Z) for statistical Characterization of location-related data recorded by at least one vehicle Event descriptions, characterized in that, based on a road or navigation map, those road sections free of inflow and outflow are identified whose section length is equal to or greater than a predetermined minimum section length, adjacent identified road sections are connected to form a section chain (10) with a starting point (11) and an end point (12), Event descriptions comprising a geoposition (P1, P2) and at least one parameter of a vehicle environment model of vehicles are recorded and transmitted to an analysis system, the analysis system assigns a geoposition (P1, P2) of an event description to a nearest road network location (01, 02) of a section chain (10) and by a longitudinal distance (L, L1, L2) relative to the Starting point (11) of the section chain (10) is described, for at least one section chain (10) a course of a density value (p) of event descriptions related to the longitudinal distance (L, L1, L2) to the starting point (11) is determined by means of a kernel density estimator with a predetermined bandwidth, the course of the density value (p) is transformed with a local Z standardization related to the bandwidth into a Z course (Z1, Z2) of a Z value (Z) which is mean-free related to the bandwidth.

2. Method according to claim 1, characterized in that the predetermined minimum section length is selected so that the at least one The section chain (10) is formed exclusively or predominantly from road sections which are assigned to a motorway or an expressway.

3. Method according to one of the preceding claims, characterized in that the vehicle environment model fuses a static vehicle environment and a dynamic vehicle environment.

4. Method according to one of the preceding claims, characterized in that in each case a region of road network locations (01, 02) connected along a section chain (10) between a peak start (Lmin) and a peak end (Lmax), at which the Z value (Z) exceeds a predetermined threshold value (0), preferably the value 1, is identified as a peak region (R1, R2, R3).

5. Method according to claim 4, characterized in that a peak region (R1, R2, R3) is assigned an identifier of the respective section chain (10), a longitudinal distance (L, L1, L2) of the peak start (Lmin) to the starting point (11) of the respective section chain (10) and a Longitudinal distance (L, L1, L2) of the peak end (Lmax) to the end point (12) of the respective section chain (10), a maximum value of all Z values (Z) within the peak region (R1, R2, R3) and a description of the geometry of the peak region (R1, R2, R3).

6. Method according to one of the preceding claims, characterized in that the vehicle environment model comprises at least one object detected by vehicle sensors and / or a parameter and an event description is created based on a trigger and transmitted to the analysis system, where the trigger marks an event identified as worthy of consideration based on the vehicle environment model.

7. The method according to claim 6, characterized in that the trigger marks a warning and / or an intervention of a driver assistance system.

8. Method according to one of the preceding claims, characterized in that at least one event description is assigned an event class selected from a list of event classes and only event descriptions of one or a plurality of event classes are taken into account for determining the density (p).

9. Method according to one of the preceding claims, characterized in that an event description comprises a direction of travel (11, I2) related to the direction of a section chain (10) pointing from its starting point (11) to its end point (12), and only event descriptions of the same direction of travel (11, I2) are taken into account for determining the density (p).

10. Method according to one of the preceding claims, characterized in that Z values (Z) determined at road network locations 01, 02) along at least one section chain (10) are visualized by markings (M) running perpendicular to the section chain (10), the length of which is selected proportional to the respective Z value (Z).

Images