System and method for predicting absolute and relative risks of car accidents.

专利摘要:
A system and a method are proposed for measuring and predicting absolute and relative risks of car accidents exclusively on the basis of non-insurance-related measurement data and on the basis of automated traffic pattern recognition, whereby records of accident events are generated and location-dependent probability values for those associated with the risk of a car accident specific accident conditions are determined. The proposed system (1) thus provides a raster-based (2121, 2122, 2123, 2124), technically new way of automating the risk prediction in relation to motor vehicle accidents using suitable measuring devices and systems (41, ..., 45) received environmental-based factors ( Altitude, road network, traffic data, weather conditions), including socio-economic factors that influence motor vehicle traffic and are location-dependent. In this way, predictions of the accident risk can be provided for any area.
公开号:CH714036B1
申请号:CH01489/18
申请日:2016-11-07
公开日:2020-03-31
发明作者:Chatziprodromou Iordanis；Nagel Uwe；Ventkateswaran Ramya；Larkin Peter；Elsasser Christian
申请人:Swiss reinsurance co ltd；
IPC主号:

专利说明:

Field of the Invention
The present invention relates to a system for predicting and determining absolute and relative risks of car accidents, in particular for the automated location-dependent prediction of car accidents exclusively on the basis of non-insurance-related data. Furthermore, the present invention also relates to automated systems and methods for evaluating and / or carrying out risk transfers for a vehicle between risk-exposed units and an automated risk transfer unit such as automated insurance systems. The systems and methods may use automated data collection, collection, and processing means to determine scoring factors for grid locations associated with a geographic area in which the vehicle is used. Furthermore, the system can also include suitable means for signal generation and signal transmission to electronically operated and linked systems.
State of the art
Devices and methods for automated traffic and driving pattern recognition for determining risks associated with driving or traffic patterns are complex and technically difficult or technically impossible (to the point of chaotic behavior of the traffic or driving pattern) to execute with a sufficient degree of accuracy . In addition to determining linked risks, traffic or driving pattern recognition and traffic pattern recognition is also an important part of most modern intelligent traffic systems. The detection of urban traffic conditions forms the basis of intelligent tax, control, synergy and risk assessment systems.
There are various approaches to implementing such systems in the relevant technological field. All approaches must somehow determine the at least three-dimensional space with traffic volume, average speed and load ratio. Furthermore, a classification of traffic situation patterns must take place, for example in the form of a blocked river, a tough river, a steady river and an unobstructed river. The classification can be based on historical traffic patterns, for example. To process the data, the systems can compare the classification result of different traffic model functions and thus carry out the traffic situation pattern recognition, for example via a support vector machine. The risk and traffic factors determined must therefore reflect the characteristics of the traffic conditions. The traffic model functions must be able to distinguish at least different patterns of traffic flows with high classification accuracy, the data normalization having a significant influence on the result of the classification.
In all of these control, control, synergy and risk assessment systems, the traffic ratio detection and prediction is typically based on highly dynamic factors in temporal and topographical resolution. Although traffic risk factors that are recognized or predicted in terms of temporal dynamics can be linked to a broader time frame, for example a day, a month or a year, the dynamics of traffic risk factors correlate directly with the dynamics of the traffic ratio pattern, the risk factors being greater than the assessed time frame be averaged.
Risk transfer in insurance technology, particularly for underwriting processing during risk transfer, involves the process of evaluating the value of a particular risk and, in turn, identifying monetary resources that are typically to be transferred in a periodic time frame for protection the possibility of occurrence of a risk event associated with the transferred risk. To ensure the operational stability of the risk transfer system, i.e. the insurance system, the transferred resources are defined to reflect the amount of a payout when a covered event occurs, given the likelihood of that event occurring. The process of determining the amount of resources to be transferred is called evaluation. The assessment process may include a number of variables, including experience data for a specific insured entity, experience data for a class of insured entity, investment forecasts, margin targets, and a wide variety of other data used to predict the occurrence of certain real events and the amount of damage, that are likely to result from such events are useful.
The process of historical assessment or empirical assessment includes analyzing the previous claims experience in order to determine an expected amount of resources (for example premiums collected) and / or a retroactive amount of resource adjustment. For example, a risk-exposed entity may operate a large fleet of vehicles, which risk-exposed entity may attempt to transfer vehicle risk to cover property damage and personal injury claims if a fleet vehicle is involved in an accident with another vehicle. If the fleet is large enough or the risk-exposed unit has operated the fleet long enough, sufficient historical data may be available and the expected claims for the next few years can be accurately estimated. This estimate (possibly combined with an allocation of expenditure or valuation of an administration fee) would represent the transferred monetary resources, for example the insurance premium, in an ideal scenario. At the end of the policy contract year, a surcharge or reimbursement may also be appropriate if the actual claims for the year were higher or lower than the estimated claim amount.
Automated systems for determining the amount of resources to be pooled to offset a suitable risk transfer for a vehicle, such as an insurance premium, may include, for example, the steps of (i) receiving acceleration sensors collected by using one or more vehicles Driving information, (ii) recognizing a driving and / or traffic pattern for an accident risk from driving habits of the vehicle driver based on the entered driving and / or traffic information and finally (iii) determining the insurance premium in accordance with the accident risk rate of the driver on the basis of the recognized traffic and / or driving pattern.
However, the risk transfer for a typical automobile cannot by far generate the amount of data required to produce a reliable and accurate estimate of expected claims for the vehicle or vehicles. Thus, automated insurance systems typically have to compare specific risk transfers, i.e. specific policies, in a risk pool of comparable policies in order to generate sufficient data to produce such an estimate. A mechanism for this includes an assessment of what data is available for a specific vehicle (e.g. demographic information, vehicle types and what limited risk transfer information, i.e. exposure information is available), and using this data to assign an appropriate pool of resources to the specific vehicle . The various types of data available for an automated risk transfer system to perform the assessment process are often linked to geographic locations or regions. However, this link is not consistent or even. Some property offense data is linked to a "block" of addresses in a street in a city, for example block 300-400 in the main street. Road zone data and ground level data can be saved as complex topographic maps. Loss experience data can be linked to a pair of coordinates to represent the longitude and latitude of the location of the loss event.
In systems according to the prior art, the risk transfer evaluation requires a complex search process for compiling relevant data for input in an evaluation function. For example, a risk transfer for a specific vehicle to be assessed can be linked to a specific location, for example a street address of an apartment or an office or the location where a vehicle is parked at night. To evaluate a policy for this location, a subset of the relevant data must be collected and made available for an evaluation algorithm. The collection process is often technically difficult due to the inconsistent and non-uniform association of data with geography as described above. In some cases, data is processed and aggregated by country, city and / or zip code. This aggregation is made difficult by the almost unlimited boundaries defined by districts, locations and postcodes. Furthermore, boundaries of districts, locations, and postcodes can change over time. In other cases, data is processed through an aggregated sales area.
[0010] Furthermore, according to the prior art, evaluations assigned to each area are determined essentially on the basis of the linked historical experience of claims. Existing area assessment procedures were used in state-of-the-art risk transfer systems. However, these approaches can be problematic for several reasons: (i) Geographic boundaries can change as previously described. (ii) Geographic areas can be larger than desired. (iii) The population may not be evenly distributed across these geographic areas. (iv) The historical exposure experience in these geographic areas may be limited. (v) The location of a vehicle parked in a garage does not accurately measure the geographic risk of the vehicle's location.
The document US 2014/0 365 246 A1, which corresponds to the prior art, discloses a system which provides an automated underwriting and evaluation process for risk transfer products. The system receives information about the use of a vehicle in a geographic area or at a geographic location and determines target grid cells, defined by latitude and longitude, which comprise the geographic area in which the vehicle is used. The system also determines a set of predefined data associated with the target grid cells and generates a location rating factor based on the usage information, the target cells concerned, and the set of predefined data.
Summary of the invention
An object of the present invention is to provide a system and a method for sharing the risk of risk events of a variable number of risk exposure components by ensuring dynamic, independent risk protection for the risk exposure components to cover various risks and risk categories which do not have the aforementioned disadvantages exhibit. In particular, a device and a method for automated traffic and / or driving pattern recognition and for determining risks associated with driving or traffic patterns with a high degree of temporal and spatial resolution, in particular for regions in which the amount of accessible historical data is small or even does not exist.
[0013] The present invention fulfills this object in particular with the elements of the independent claims. Further advantageous embodiments also emerge from the dependent claims and descriptions.
In particular, these objects are achieved by the invention in that with the system for the automated location-dependent prediction of absolute and relative risks of car accidents, data records of accident events can be generated exclusively on the basis of non-insurance-related data and location-dependent probability values for those with the risk of a car accident linked specific accident conditions can be determined, that a spatial grid with grid cells can be generated over a geographic area of interest with a detection unit, the area comprising at least some of the units exposed to accident risks, the grid cells of the grid being selectable and data with the system for each cell of the Grid can be assigned, and wherein data sets representative of a grid cell are assigned to a year of occurrence or measurement and are stored in a memory module of a computing unit that for each A cell population density parameter can be recorded with a settlement pattern trigger and a data record assigned to the corresponding grid cells can be assigned, population density parameters for the geographic area of interest being recorded and suitable weighting factors can be assigned in the spatial grid, taking into account the various settlement parameters, that the first high-resolution ones recorded by the first air-based measuring stations Air data can be transferred to the system and ground cover parameters can be generated and stored with the generated data set assigned to the corresponding grid cells on the basis of the first high-resolution air data, the ground cover parameters being a measure of the observable biophysical cover on the earth's surface that second high-resolution measured by second air-based measuring stations Data on light density are transferred to the system and night light parameters with the corresponding The generated data set assigned to the raster cells can be generated and stored on the basis of the second high-resolution air data on the light density, the night light parameters being generated on the basis of their weighted indication of local activity and correlation with other masses for signs of well-being that were recorded by systematically operated geodetic measuring stations third high-resolution data can be transmitted to the system and road map parameters can be generated and stored with the generated data set assigned to the corresponding raster cells on the basis of the third high-resolution data of the geodetic measuring stations, the road map parameters comprising at least one classification parameter for specifying a type of the assigned road, that of air-based Measuring stations recorded fourth measured high-resolution air data can be transferred to the system and precipitation parameters are assigned to the corresponding grid cells This generated data record can be generated and stored on the basis of the fourth high-resolution air data, the precipitation parameters generated comprising a measure of the hydrological cycle which indicates at least the distribution, amounts and strength of the local precipitation at a specific point or in a specific area of the corresponding grid cell. that fifth measured high-resolution air data recorded by fourth air-based and / or space-based measuring stations are transmitted to the system and digital height parameters can be generated and stored with the generated data set assigned to the grid cells on the basis of the fifth high-resolution air data, the digital height parameters generated being a measure of the terrain height at a specific point or in a specific area of the corresponding grid cell to provide a representation of the terrain surface include that the system has a trigger module with a hash table with a Contains a plurality of selectable morphological traffic model functions, with the data sets generated for each raster cell being filterable by predefined trigger parameters in order to trigger threshold values of the generated population density parameters, ground cover parameters, night light parameters, road map parameters, precipitation parameters and digital height parameters, the morphological traffic model functions being recorded using a scaling table based on actual accident data can be compared, and a specific morphological traffic model function can be triggered and selected by the best comparison with the accident data, and that a risk value field for each of the grid cells can be generated with an interpolation module on the basis of the data records linked to the specific grid cell, and a probability with the interpolation module at every point in the grid due to the likelihood of an accident occurring is assignable to a specific geographical location and at a specific time.
The present system provides, among other things, a technically new way of automating the risk prediction in relation to motor vehicle accidents using environmental factors (altitude, road network, traffic data, weather conditions), including socio-economic factors that influence motor vehicle traffic and are location-dependent. In this way, predictions of the accident risk can be generated for any area. The corresponding data is extracted from satellite-based raster data or road network data, which are collected by remote sensing methods and geodesy. These are then aggregated according to administrative areas or areas that are defined by any grid, or are aggregated at a street section level by capturing the environmental situation (population density, number of streets nearby, etc.) around a specific street section. In this aggregation, characteristics are created from the underlying data, which represent aspects of the environment that are relevant to the risk of motor vehicle accidents (e.g. number of intersections). The system is calibrated by comparing the characteristics of the areas or road sections with the number and type of accidents that have occurred there and linking the characteristics and accident data, for example using the machine learning methods described below. The type of accident represents the severity of the accident, described as the number of insured and fatalities involved in the accident, the geometry of the accident (rear-end collision, side collision, frontal accident, transverse collision), the number of cars involved or the damage classification. The output signaling of the system can be done, for example, as a function with values for the various characteristics as input and the accident risk parameters (for example, accident frequency and severity) as output. By comparing the risk transfer data (customer portfolio data and / or market portfolio data), different gravity metrics can be translated or transferred into monetary values, which represent the average loss per accident for different business areas.
[0016] In an alternative embodiment, the high-resolution air data include aerial images and / or satellite images and / or aerial photos. The high-resolution air data can, for example, also include aerial images and / or satellite images and / or aerial photos measured by satellites and / or aircraft and / or aircraft more easily than air or other measuring stations equipped with a balloon.
In a further alternative embodiment, the sign of other mass for signs of well-being includes highly local mass for human well-being and / or national or sub-national gross domestic product (GDP) mass, the mark being used as the weighting parameter of the system.
In a further alternative embodiment, the third high-resolution data are selected with a data extraction from an accessible high-resolution road map database.
In an alternative embodiment, the geodetic measuring stations comprise a GPS (Global Positioning System) unit or can be located by satellite imaging. Each parameter can, for example, be grouped, the combination of which is translated into a meaningful class, for example highway, sidewalk, etc. The classification of the road map parameters can, for example, values for classifying bike paths, sidewalks, highways, paths, pedestrians, main streets, residential streets, side streets , Steps, supply lines, tertiary lanes and unclassifiable road objects.
[0020] In a further alternative embodiment, the classification parameters include tag elements that enable attributes of the classification. The classification parameters can also include, for example, a measure of an average speed of a road user at the specific point of the grid cell.
[0021] In an alternative embodiment, the precipitation parameters comprise at least parameters for measuring the precipitation of rain and / or snow and / or hail.
[0022] Finally, in a further alternative embodiment, the digital height parameters also include morphological elements.
The present invention relates not only to the method according to the invention, but also to a system for executing this method and a corresponding computer program product.
[0024] Alternative embodiments of the present invention are described below with reference to examples. The examples of the embodiments are illustrated by the following attached figures:<tb> Fig. 1 <SEP> shows a block diagram which schematically shows an exemplary embodiment of a system (1) for the automated location-dependent prediction of absolute and relative risks of car accidents exclusively on the basis of non-insurance-related data. Data records of accident events are generated and location-dependent probability values for specific accident conditions associated with the risk of a car accident are determined. A spatially high-resolution grid 212 with grid cells 2121, 2122, 2123, 2124 is generated over a geographic region 21 of interest with a detection unit 2. The geographical area 21 comprises at least one section of units 70-74 that is exposed to accident risks, wherein the grid cells 2121, 2122, 2123, 2124 of the grid 212 can be selected and data about the system can be assigned to each cell 2121, 2122, 2123, 2124 of the grid 212 ,<tb> Fig. 2 and 3 <SEP> show diagrams that schematically illustrate the arrangement and resolution of data records. As an example, the figures show data extraction and grid generation for the exemplary countries China, Germany, India and Turkey. The resolution can be adjusted on the basis of dynamically triggered levels, such as community and district grids: 2 * 2, 4 * 4, 10 * 10, 15 * 15 km grids. The grids can also be selected differently, for example according to the available data quality or resolution. A suitable four-tree data structure can be calculated from the population density parameters and linked to the processing steps by system 1. In the case of grid cells of equal spacing and size, four-tree data structures are not used or are not required.<tb> Fig. 4 <SEP> shows a diagram schematically representing a grid of population data 401. System 1 extracts population density parameters 4001, 4002, 4003, 4004, a standard deviation and / or other parameters, such as biomass parameters and / or land use parameters (e.g. commercial, residential area etc.) and / or soil cover parameters (agricultural, urban, lake, etc.). The weighting factors 4011, 4012, 4013, 4014 are generated on the basis of the extracted population density parameters 4001, 4002, 4003, 4004.<tb> Fig. 5 <SEP> shows a diagram which schematically represents the image processing of first high-resolution air data 411, that is to say satellite images of ground cover data, measured by first air-based measuring stations 41. The first high-resolution air data 411 are transmitted to the system 1 and soil covering parameters 4101, 4102, 4103, 4104 are generated with the system 1, that is to say the system 1 generates defined areas of forest, rural area, urban area, cultivated land, etc.<tb> Fig. 6 <SEP> shows a diagram that schematically illustrates the image processing of satellite images of night light data, where night light values provide an indication of human wellbeing, such as gross domestic product (GDP).<tb> Fig. 7 <SEP> shows a diagram which schematically shows the data extraction from map data, road type, length and features, for example motorways, main roads, trunk roads, sidewalks etc., road crossings, one-way streets, tunnels, bridges and speed limits. The dynamically adapted map data can be based on measuring devices linked to the system 1. However, they can also be based on access to special high-resolution map data, such as OpenStreetMap (OSM) data resources. The aforementioned OSM data sample performs data collection dynamically with a large number of independent users (currently over 2 million) who collect data through manual surveying, GPS devices, aerial photos and other freely available sources. These are structured as crowdsourcing data that can be accessed under Open Database Access.<tb> Fig. 8 <SEP> shows a diagram that schematically shows the generation and extraction of the precipitation parameters 4401, 4402, 4403, 4404 and image processing of the fourth high-resolution air data 441, for example satellite images of precipitation data, the precipitation parameters 4401, 4402, 4403, 4404 for at least the average annual rainfall and total rainfall. The precipitation parameters 4401, 4402, 4403, 4404 are stored by the generated data records 2221, 2222, 2223, 2224, which are assigned to the corresponding raster cells 2121, 2122, 2123, 2124.<tb> Fig. 9 <SEP> shows a diagram that schematically shows the generation and extraction of the digital altitude parameters 4501, 4502, 4503, 4504 and the image processing of fifth high-resolution air data 451, for example satellite images, with parameters for maximum altitude, standard deviation for altitude and mean Height are generated.<tb> Fig. 10 <SEP> shows a diagram schematically illustrating the improvement in the country modeling according to the present invention. The top two images of Fig. 10 are community based, with 26% predicted with a 20% fault tolerance. The two images below are grid-based, with 87% predicted with an error tolerance of 20%.<tb> Fig. 11 <SEP> shows a diagram that schematically shows the calibration and transfer mechanisms of system 1, illustrated using the example of the prediction of accidents in Turkey with a model trained in Germany. The prediction or forecast, i.e. data processing, is based on accident data for Germany. This prediction uses characteristics for Turkey that are also available for China. For the example with a preliminary transfer effort, an error of 3 to 20% and a correlation of 96% between the prediction and the actual values are achieved. Overall trends are well captured and a good distinction is made between high and low risk regions.<tb> Fig. 12 <SEP> shows a diagram that schematically shows a risk analysis by the present system for the provinces of Jiangsu and Shandong in China. The model applied to China provides an estimate of the number of accidents for different regions. Fig. 12 shows the predicted accident distributions for the two provinces Jiangsu and Shandong.<tb> Fig. 13 <SEP> shows a diagram that schematically shows a comparison of the full model with the traffic risk factors. The present invention groups the risk factors according to those which provide a measure of the traffic risk, that is to say motorways, residential streets, bridges and tunnels. 13 shows the accident prediction and the percentage contribution by traffic risk factors. Areas with more traffic flow, as measured by these traffic risk factors, tend to have a higher accident prediction level.<tb> Fig. 14 <SEP> shows a diagram that schematically shows a comparison of risk factors with traffic flow data. Figure 14 shows in traffic flow data that neighboring Shanghai has a major impact on the south of Jiangsu Province. The traffic risk factors show a good measure of this traffic flow data.<tb> Fig. 15 <SEP> shows a diagram that schematically shows liability insurance claims in comparison to traffic risk factors. The liability insurance claims are represented by a number per period (left figure), which enables a direct comparison with the traffic risks.<tb> Fig. 16 <SEP> shows a diagram which schematically shows the processing of the second high-resolution data 421 for light density, that is to say the light data, according to the invention. The left figure shows the second high-resolution data 421 on light density, while the right figure shows the predicted or predicted accident risks. The middle figure shows a possible portfolio of transferred risks. As can be seen from Figure 16, the present invention also enables the optimization of risk transfer portfolios linked to a geographical area 21.<tb> Fig. 17 <SEP> finally shows a block diagram that schematically shows the automated data processing by generating and determining the corresponding traffic risks and traffic risk maps, includes the steps for (i) data selection and conversion, (iii) modeling of the individual countries and (iii) generating the Transfer modeling.
1 schematically shows an architecture for a possible implementation of an embodiment of the system 1, an automated location-dependent prediction of absolute and relative risks of car accidents being carried out exclusively on the basis of non-insurance-related data. The system generates data records of accident events and determines location-dependent probability values for specific accident conditions associated with the risk of a car accident.
A spatially high-resolution grid 212 with grid cells 2121, 2122, 2123, 2124 is generated over a geographic area 21 of interest with a detection unit 2, as shown by FIGS. 2 and 3. The geographical area 21 comprises at least one section of units 70-74 that is exposed to accident risks. The grid cells 2121, 2122, 2123, 2124 of the grid 212 are selectable and data can be assigned to each cell 2121, 2122, 2123, 2124 of the grid 212 via the system, and data records representative of a grid cell are assigned to a year of occurrence or measurement and stored in a memory module of a computing unit. The generation of location and resolution of data records is shown by FIGS. 2 and 3. These figures show as an example the data extraction and raster generation for the exemplary countries China, Germany, India and Turkey. The resolution can be adapted to dynamically triggered levels, such as community and district grids: 2 * 2, 4 * 4, 10 * 10, 15 * 15 km grids. The grids can also be selected differently, for example according to the available data. A suitable four-tree data structure can be generated from the population density parameters using system 1 and linked to the processing steps by system 1.
For each grid cell 2121, 2122, 2123, 2124, an ambient population density parameter 4001, 4002, 4003, 4004 is recorded with a settlement pattern trigger 40 and a data record 2221, 2222, 2223 assigned to the corresponding grid cells 2121, 2122, 2123, 2124, 2224 assigned. Population density parameters 4001, 4002, 4003, 4004 are recorded for the geographical area 21 of interest and suitable weighting factors 4011, 4012, 4013, 4014 are assigned in spatial grid 212 taking into account the various settlement parameters. With regard to the population grid used, the population grid can, for example, be essentially model-based on UN statistics and local authority data on administrative units, wherein a suitable algorithm can be used to estimate the corresponding grid densities. The raster format can be used for simple data integration and there is no need to use “imaging” in the strict sense. In a further embodiment variant, the population density parameter 4001, 4002, 4003, 4004 can be extracted, for example, by system 1 from high-resolution air data 401, as shown in FIG. 4, which include, for example, aerial images and / or satellite images and / or aerial photos. Essentially for using high resolution air data 401, 411, 412, 414, 415 of the present invention, the high resolution air data 401, 411, 412, 414, 415 from satellites and / or aircraft and / or aircraft can be lighter than air or others with a balloon equipped measuring stations include measured aerial images and / or satellite images and / or aerial photos.
The extraction of the population density parameters 4001, 4002, 4003, 4004 can be based on measured interaction between population density parameters and / or land use parameters and driving or traffic patterns. To carry out the extraction with the system 1, the system 1 can comprise variables which measure the interaction of land use and traffic behavior, that is to say traffic patterns. For extraction, population density is the primary quantifiable land use descriptor variable. System 1 can also use population density parameters to distinguish area types (city, town, suburb, city, and country). System 1 may also include other variables that may relate to the quantification of land use, including parameters for residential density and occupancy density. Other parameters and characteristics of the population or built environment, such as race, age, income and commercial employment, can also be used to weight the land use impact across different population groups. For the extraction, for example, a higher population density can be linked to a decreasing annual mileage, a greater bus availability, a decreasing dependency on vehicles occupied by one person and an increasing use of public transport. The private car remains the predominant mode of transportation for most geographic areas 21, although African Americans, Asians, and Hispanics generally use other modes of transportation. An increasing population density is typically associated with fewer passenger journeys, fewer passenger kilometers traveled and fewer passenger kilometers per journey. The least number of vehicle trips, vehicle kilometers traveled and vehicle kilometers per trip are recorded for residents of densely populated areas. Less densely populated areas tend to show more drivers among adults and more vehicles among adults.
To determine the adjusted weighting factors 4011, 4012, 4013, 4014, the small towns mentioned tend to follow national averages in relation to several traffic parameters, for example drivers among adults, vehicles distributed among adults, proportion of people working at home, and car dependency . Around 20% of small town residents travel to work differently than by private car. For residents of smaller cities, the largest number of passenger journeys from each area type can be recorded. In the case of people in suburbs, the next largest number of passenger trips can be determined. Typically, a large number of low-income residents live in small towns that have limited availability of public transportation. For the extraction, the system 1 can also identify location-specific preferences of specific segments of the population, for example. High-income households tend to live in suburbs, while middle-income households mostly live in rural areas. Low-income households are generally found in urban or rural areas. The distance and travel time to work increase with the increase in the share of trade in one area. Urban areas have the smallest proportion of working residents in census sections with over 25% in retail. Small towns have the largest share with 28.8% working residents, with over 25% of jobs in retail. Employment in trade and employment density in the working census section show some measurable correlations with driving behavior. In the home block group, increasing housing density is linked to greater availability of public transport and greater proximity to public transport. Cycling and walking increase with increasing residential density. An increasing residential density is also linked to an increasing employment density. With residential densities between 100 and 1499 housing units per square mile, people are less likely to work in jobs without a permanent job. Areas with low residential density have the largest proportion of people who work at home.
Thus, in summary, residential density parameters, commercial employment, income, area type and population density parameters provide important descriptors for traffic behavior and policy implementation and are related to the association of land use with traffic choice and behavior, whereby the data extraction by system 1 for the surrounding population density parameter 4001, 4002, 4003, 4004 and the adjusted weighting factors 4011, 4012, 4013, 4014 based on the measured variables.
As FIG. 5 shows, the system 1 provides image processing of first high-resolution air data 411, that is to say satellite images of ground cover data, measured by first air-based measuring stations 41. The first high-resolution air data 411 are transmitted to the system 1 and soil covering parameters 4101, 4102, 4103, 4104 are generated with the system 1, that is to say the system 1 generates defined areas of forest, rural area, urban area, cultivated land, etc. Soil covering parameter 4101, 4102, 4103, 4104 are stored with the generated data record 2221, 2222, 2223, 2224 assigned to the grid cells 2121, 2122, 2123, 2124 on the basis of the first high-resolution air data 411. The ground cover parameters 4101, 4102, 4103, 4104 are a measure of the observable biophysical cover on the earth's surface. Transport and mobility are related and have increased as a result of demographic, economic, land use and international developments. It can be said that transport is not only related to economic growth and social development, but is also a necessary condition for it. There are various measurement parameters relating to the land cover parameters 4101, 4102, 4103, 4104, e.g. construction of relevant infrastructures, traffic management measures (e.g. inflow control, route guidance), land use policy (e.g. city of short distances) and measures that try to influence traffic behavior ( for example tolls). The suitable measures can be selected by weighting their influence on their dependence on the function of the traffic system, that is to say a traffic analysis, for example on the basis of historical traffic data. The traffic analysis process can provide the necessary understanding of the relationship between the function of the traffic system and the underlying phenomena. As described, the process can consist, for example, of examining historical traffic data. For the analysis, the traffic analysis process can be based on the traffic situation on an average day (Annual Average Daily Traffic / AADT) or an average working day (Annual Average Weekday Traffic, Annual Average Weekday Traffic, AAWT) and with the design hour volume (Design Hour Volume, DHV) can be applied. In a second step, the time grid size can be refined. Typically, the traffic pattern is related to driving needs and traffic features. Driving demand is defined as the number of vehicles or people who want to drive past a point during a specified period. The essential traffic supply feature that influences the traffic pattern dynamics is the capacity. Capacity is defined as the maximum number of vehicles or people that can reasonably be expected to be serviced in the given period. Traffic management also measures the influence dynamics of traffic patterns. In some cases, traffic management enables more efficient use of available capacity (direct influence). Furthermore, in some cases the capacity is reduced or increased or certain journeys are stimulated or counteracted, for example through toll charges (indirect influence). Both the driving needs and the capacity of a street vary in time and space and are influenced by other external factors. Traffic is a derived need that is caused by the need or desire to perform activities in specific locations (e.g. living, working, shopping, recreation). Most variations in driving needs are due to the distribution of activities in time and space. In addition, travel needs may vary due to changes in the distribution of transportation, route selection or departure time due to external factors, previous experience or available information. The capacity of a street obviously depends on the street layout and regulations (e.g. maximum speed). With regard to temporal variations in the urban network, the current capacity is strongly influenced by traffic light cycles, with the capacity being zero at a red traffic light. In addition, weather, road construction, accidents and incidents can cause capacity to vary over time. The factors as described and the interaction between them cause the traffic pattern to vary in time and space. With regard to temporal variations, different time scales can be distinguished, which vary from minute to minute to year to year. The driving forces behind the variations differ according to the time scale. Short-term variations in city traffic are mainly due to traffic light cycles. Hour-to-hour and day-to-day variations are essentially caused by variations in travel needs, although variations in capacity (such as weather or road construction) may also play a role. Long-term variations in traffic are essentially due to long-term demographic, economic and infrastructural developments.
Travel needs and supply characteristics of urban areas differ significantly from those of highways. Therefore, an analysis of motorway traffic patterns and the associated dynamics cannot be translated directly into the urban situation. One difference between urban and freeway traffic is that several road users coexist and interact in the urban road network, e.g. pedestrians, bicycles, cars, buses, trucks, while freeways are mainly used by cars and trucks. This mix of road users also leads to relatively large differences in speed between urban road users. Another feature of the urban network is that it has many intersections. Thus, the traffic pattern in urban areas is characterized by many small disturbances, while on the other hand, motorway traffic patterns have essentially fewer disturbances, but with a greater impact. In terms of traffic demand characteristics, traffic in the urban network is essentially more diversified than traffic on highways. First of all, depending on the type of motorway, a motorway is mainly used for medium or long-distance traffic. The urban network also serves medium and long-distance traffic to and from the motorways, but also for a significant amount of local and short-distance traffic. The distribution of travel purposes is also more diversified in urban traffic. Most motorways are used for a main purpose of driving. In busy periods on working days, the main purposes of the trip are mainly work and business. In addition, rush hour traffic on weekends and during vacation periods caused by leisure traffic, for example from and to the beach, can be observed on some motorways. Most urban streets also have a significant amount of work and business traffic on weekdays. In addition to commuter traffic, shopping and leisure traffic also use the urban network to a large extent on working days.
As described above, dynamics of the traffic pattern and characteristics of urban areas differ significantly from those of highways. The difference between the characteristics of the different areas can be measured by parameters, for example for the urban traffic pattern with indicators such as traffic volume, speed, traffic jam length, deceleration and travel time. In this sense, the transmission of the first high-resolution air data 411 to the system 1 and the generation of the land cover parameters 4101, 4102, 4103, 4104 by the system 1, that is to say the detection of the defined areas of forest, country, city, cultural land, etc., is for the system 1 essential.
[0034] As an alternative embodiment, the system 1 can access high-resolution air data 411 from the European Space Agency (ESA), for example. The ESA satellites provide global ground cover maps with a spatial resolution of 300 m, which may be sufficient for the present use. The high-resolution ESA map data is created using a multi-year and multi-sensor strategy to take advantage of all appropriate data and maximize product consistency.
As FIG. 6 shows, second air-based measuring stations acquire second high-resolution data 421 for measuring light densities. The second high-resolution data 421 are transmitted to the system 1. Night light parameters 4201, 4202, 4203, 4204 are generated by the system 1 and stored with the generated data record 2221, 2222, 2223, 2224 assigned to the grid cells 2121, 2122, 2123, 2124 on the basis of the second high-resolution air data 421 on the light density. The night light parameters are generated based on their weighted indicia 4211, 4212, 4213, 4214 for local activity and correlation with other masses for indications of wellbeing. The weighted indicia 4211, 4212, 4213, 4214 may include examples of other indicative measures of welfare, such as highly local measures of human welfare and / or the national or sub-national gross domestic product (GDP). GDP or gross domestic product per capita is typically such a measure of national and human wellbeing worldwide, with night light values providing a measure of human wellbeing, such as GDP. The correlation of the other mass of signs of well-being, such as GDP, and the traffic pattern of the region can only be considered in its typically limited correlation. For example, GDP provides a good measure of all economic activity, but not of the local distribution of this activity.
[0036] As an alternative embodiment, the second high resolution data 421 may be based on the US Defense Meteorological Satellite Program and / or other sources. In general, more light and higher light intensity on a satellite image, for example measured in pixels per square kilometer, correlate with higher levels of development. This correlation can be illustrated, for example, with reference to FIG. 6, which shows North Korea and South Korea, North Korea being an almost black area while South Korea shining in light. Other comparisons can also illustrate the correlation, such as the comparison between a brightly lit Tokyo and the center of Africa. One of the advantages of using satellite measurement data is that satellites provide welfare data more directly than geodetic ground stations, which can take years for the latter. Therefore, satellite measurement data are more suitable for dealing with the dynamics of traffic patterns. Although a square kilometer is typically about the highest resolution available for high-resolution data 421 for measuring light output, measurement data from newer satellites, such as data from the National Polar Orbiting Partnership between NASA and NOAA, provide higher resolution images, which is associated with the execution of the present invention and the system 1 may be more advantageous.
Third high-resolution data 431 are recorded by systematically operated geodetic measuring stations 43 and transmitted to the system 1. On the basis of the third high-resolution data 431 from geodetic measuring stations 43, road map parameters 4301, 4302, 4303, 4304 are generated and stored using the generated data record 2221, 2222, 2223, 2224 assigned to the corresponding raster cells 2121, 2122, 2123, 2124. The road map parameters 4301, 4302, 4303, 4304 include at least one classification parameter 4311, 4312, 4313, 4314 for specifying a type of the assigned road. The third high-resolution data 431 can be selected, for example, with a data extraction from an available high-resolution road map database. The geodetic measuring stations 43 can, for example, comprise a global positioning system (GPS) unit or can be located by satellite imaging. The classification parameters 4311, 4312, 4313, 4314 of the road map parameters 4301, 4302, 4303, 4304 can, for example, values for classifying bike paths, sidewalks, highways, paths, pedestrians, main streets, residential streets, side streets, steps, supply lines, tertiary lanes and unclassifiable street objects include. The classification parameters 4311, 4312, 4313, 4314 also include tag elements that enable attributes of the classification. The classification parameters 4311, 4312, 4313, 4314 can also include a measure for an average speed of a road user at the specific point of the grid cell 2121, 2122, 2123, 2124.
Fig. 7 shows schematically the data extraction from map data, type of road, length and features, such as highways, main roads, highways, sidewalks, etc., intersections, one-way streets, tunnels, bridges and speed limits. The dynamically adjusted map data can on the System 1 linked measuring devices are based. However, they can also be based on access to special high-resolution map data, such as OpenStreetMap (OSM) data resources. The aforementioned OSM data sample performs data collection dynamically with a large number of independent users (currently over 2 million) who collect data through manual surveying, GPS devices, aerial photos and other freely available sources. Thus, the geodetic measuring stations 43 can comprise systematic geodesy using devices and measuring stations such as a GPS handheld device, a notebook, a digital camera or a voice recording device. However, the third high-resolution data 431 can also be generated by aerial photography, satellite imaging and data from other sources that have additional important data sources and are automatically imported into the third high-resolution data 431. Special processes can be provided, for example, to handle automated imports and to avoid technical problems. In the example of OSM, the data is structured that can be accessed under Open Database Access.
Map data of the third high-resolution data 431 is usually collected using a GPS unit, although this is not absolutely necessary if an area has already been acquired by satellite imaging. Once the third high resolution data 431 has been collected, it is entered into a data store. Initially, no information is available on the type of route being transmitted, which means that it can be, for example, a motorway, a sidewalk or a river. In a second step, the routes and objects are identified automatically and semi-automatically. In particular, the identification process includes the placement and processing of objects such as schools, hospitals, taxi ranks, bus stops, restaurants, etc., which can be done, for example, by an expert recognition system. As an alternative embodiment, the third high-resolution data 431 entered in the data memory can use, for example, a topological data structure with some core elements. Such core elements can include nodes, for example. Nodes are points with a geographic position, stored as coordinates (pairs of latitude and longitude). In addition to their use in this way, they can be used to display map features of no size, such as points of interest or mountain peaks. Paths are another possible core element. Paths can be defined as ordered lists of nodes that represent a polyline or possibly a polygon if they form a closed loop. They can be used to represent linear features such as roads and rivers and areas such as forests, parks, parking lots and lakes. Furthermore, the core elements can include relations. Relations are ordered lists of nodes, paths and relations, whereby each element (relations and paths) can optionally have a «role» (a sequence). Relations are used to represent the relationship of existing nodes and paths. Examples include road turning restrictions, routes that span multiple existing paths (such as a trunk road), and areas with potholes. The core elements can furthermore comprise tags. Tags can be key-value pairs (any two sequences). They can be used to store metadata about map objects (such as type, name and physical properties). Typically, tags are not stand-alone, but are always linked to an object, i.e. a node, a path or a relation.
Fourth high-resolution air data 441 are measured by space and / or air-based measuring stations 44 and are queried to system 1. In addition, measurement data from soil-based measuring stations can be used. On the basis of the fourth high-resolution air data 441 and / or ground-based data, precipitation parameters 4401, 4402, 4403, 4404 with the generated data record 2221, 2222, 2223, 2224, which is assigned to the corresponding raster cells 2121, 2122, 2123, 2124, on the Generated and stored on the basis of the fourth high-resolution air data 441. The precipitation parameters 4401, 4402, 4403, 4404 generated include a measure of the hydrological cycle 4411, 4412, 4413, 4414, the at least distribution 4421, 4422, 4423, 4424, quantities 4431, 4432, 4433, 4434 and strength 4441, 4442, 4443 , 4444 of the local precipitation at a specific point or in a specific region of the corresponding grid cell 2121, 2122, 2123, 2124. The precipitation parameters 4401, 4402, 4403, 4404 can, for example, include at least parameters for measuring the precipitation of rain and / or snow and / or hail. FIG. 8 illustrates such a generation and extraction of the precipitation parameters 4401, 4402, 4403, 4404 and image processing of the fourth high-resolution air data 441, for example satellite images of precipitation data, the precipitation parameters 4401, 4402, 4403, 4404 for at least the average annual rainfall and the total Rainfall provides. Correlations between weather-related data and traffic patterns are known. However, they are typically known in connection with weather-related influences and delays in air traffic patterns, while the background of the correlation between weather-related influences and delays in air traffic patterns is completely different and essentially relates to the airports concerned. However, weather influences, for example represented by the hydrological cycle 4411, 4412, 4413, 4414 described above, the precipitation distribution 4421, 4422, 4423, 4424, quantities 4431, 4432, 4433, 4434 and strength 4441, 4442, 4443, 4444 influence the performance of the freeway systems every day and every hour. Rain, snow, ice and Ä. are responsible, in part or in full, for over 1.5 million freeway accidents and over 600,000 injuries and 7,000 deaths on U.S. roads each year.
Furthermore, in the United States alone, automobile drivers lose approximately 1 billion hours a year in traffic jams due to adverse weather. Weather is effectively the second most important cause of one-time freeway congestion, which is responsible for around 25 percent of delays. Studies have shown that adverse weather significantly increases average travel times depending on the area selected, for example, 14 percent in the Washington DC area and 21 percent in Seattle, WA. Rush hour traffic in Washington DC can increase by 24 percent in rain or snow. Despite the effects of adverse weather on traffic patterns and traffic, systems for traffic pattern recognition and prediction typically do not consider the links between weather and traffic flow. Accurate and timely road and weather data is important because it enables you to manage infrastructure in real time in response to existing and upcoming weather conditions and to warn motorists of changes in weather and road conditions. Advances in intelligent traffic systems, road weather information systems, weather and traffic data collection, and forecasting technologies must be based on a better understanding of how drivers behave in adverse weather and how their decisions affect traffic flow. By extracting and generating the precipitation parameters 4401, 4402, 4403, 4404 on the basis of the fourth high-resolution air data 441, the present invention fully takes into account the weather-related correlation and does not have the disadvantages of the systems according to the prior art. As an alternative embodiment, the present invention may further include means for real time changing of traffic light and inflow control timing, operating automated defrosting systems, and setting various speed limits that enable the signaling of system 1 to be widely used.
The fourth high resolution data 441 may further include weather and traffic data from static and fixed devices such as video cameras, traffic counters, loop detectors, airport weather stations, and environmental sensor stations. The fourth high-resolution data 441 can likewise be acquired at least in part by traffic and weather information supplied by moving vehicles. Thus, the present invention can take into account the effects of adverse weather on the macroscopic (overall) traffic flow and quantified changes in traffic speed, capacity and density in correlation with the generated precipitation parameters 4401, 4402, 4403, 4404. It takes into account that the correlation of the precipitation parameters 4401, 4402, 4403, 4404 (rain or snow) with the traffic pattern does not necessarily influence the density of the traffic flow, but does affect the speed with free traffic flow, the speed with the capacity and the capacity. Most of these parameters vary with the amount of precipitation. Although a 12-20 percent decrease in capacity occurs in snow conditions, the reduction in capacity does not normally depend on the amount of snow (or speed of snow). A local weighting of the precipitation parameters 4401, 4402, 4403, 4404 can be advantageous, since it was established that precipitation parameters 4401, 4402, 4403, 4404 can include strong geographical correlations. Observations show that in a first area (colder region) there are greater decreases (for example of about 20 percent) in the speed with free flow of traffic and the speed in capacity in snow than in another comparable area (for example about 5 percent in the warmer area) . One possible explanation is that drivers who are more used to snow are more aware of the dangers and therefore drive more slowly. In any case, it can be advantageous, particularly for traffic pattern estimation and prediction, that local dependencies of precipitation parameters 4401, 4402, 4403, 4404 are taken into account by system 1. The data processing used must therefore be calibrated for a variety of local conditions and traffic patterns for implementation and evaluation, in particular if the system 1 is used in the context of regional planning and regional operation. As an alternative embodiment, the fourth high resolution data 441 can be accessed using the appropriate data from the European Center for Medium-Range Weather Forecasts (ECMWF) and can be transferred to system 1.
Fifth high-resolution air data 451 are measured by fourth air-based measuring stations 45 and transmitted to system 1. Digital height parameters 4501, 4502, 4503, 4504 are generated and stored with the generated data record 2221, 2222, 2223, 2224 assigned to the grid cells 2121, 2122, 2123, 2124 on the basis of the fifth high-resolution air data 451. The digital height parameters 4501, 4502, 4503, 4504 can further comprise, for example, morphological elements.
9 shows the generation and extraction of the digital height parameters 4501, 4502, 4503, 4504 and the image processing of fifth high-resolution air data 451, for example satellite images, parameters for maximum altitude, standard deviation for altitude and mean altitude being generated. The generated digital height parameters 4501, 4502, 4503, 4504 include a measure 4511, 4512, 4513, 4514 for the terrain height at a specific point or in a specific area of the corresponding grid cell 2121, 2122, 2123, 2124 to provide a representation of the terrain surface. The digital height parameters 4501, 4502, 4503, 4504 provide a digital representation of a terrain surface that is generated from terrain elevation data. In the present case, the digital height parameters 4501, 4502, 4503, 4504 represent the surface of the earth including all objects on it. This contrasts, for example, with digital terrain parameters, which represent the pure surface of the ground without objects such as systems and / or buildings. The digital height parameters 4501, 4502, 4503, 4504 can represent the surface as a grid (a network of squares, also called a height map when displaying the height) or as a vector-based triangular irregular network (TIN). System 1 can generate digital elevation parameters 4501, 4502, 4503, 4504 in numerous ways, but using remote sensing or direct geodetic data. A method for generating the digital elevation parameters 4501, 4502, 4503, 4504 is, for example, using an artificial aperture radar in which two orbits of a radar satellite, or a single orbit if the satellite is equipped with two antennas, collects sufficient data for the digital map or digital altitude parameters 4501, 4502, 4503, 4504 in the ten-fold kilometer range on one side with a resolution of about ten meters. Other types of stereoscopic pairs can be used using the digital image correlation method, whereby two optical images with different angles, captured in the same pass of an airplane or an Earth observation satellite, are acquired. Other methods of generating digital elevation parameters 4501, 4502, 4503, 4504 may include, for example, interpolating digital contour maps that can be created by direct geodesy of the ground surface. This method is particularly useful in mountainous regions where interferometry is not always satisfactory. The digital height parameters 4501, 4502, 4503, 4504 include that the height is continuously available at any location in the examination area. The quality of the digital height parameters 4501, 4502, 4503, 4504 correlates with how accurate the height at each pixel is (absolute accuracy) and how exactly the morphology is displayed (relative accuracy). Several factors play an important role in the quality of the digital elevation parameters 4501, 4502, 4503, 4504, such as terrain roughness, acquisition density (elevation data collection method), grid resolution or pixel size, interpolation algorithm, vertical resolution, terrain analysis algorithm, quality masks related to coastline, lakes, snow , Clouds, correlation, etc. As an alternative embodiment, the fifth high resolution air data 451 and / or the applied digital elevation model can be accessed by the United States National Aeronautics and Space Administration (NASA).
The system 1 comprises a trigger module 3 with a hash table 31 with a plurality of selectable morphological traffic model functions 311, 312, 313 etc. For each raster cell 2121, 2122, 2123, 2124, the data records 2221, 2222, 2223, 2224 generated are carried out Predefined triggering parameters 321, 322, 323 etc. are filtered in order to generate threshold values for the generated population density parameters 4001, 4002, 4003, 4004, ground covering parameters 4101, 4102, 4103, 4104, night light parameters 4201, 4202, 4203, 4204, street map parameters 4301, 4302, 4303, 4304 To trigger precipitation parameters 4401, 4402, 4403, 4404 and digital height parameters 4501, 4502, 4503, 4504, the morphological traffic model functions 311, 312, 313 etc. are compared with a scaling table 33 on the basis of recorded actual accident data 331. A specific morphological traffic model function 311, 312, 313 etc. is triggered and selected by the best comparison with the accident data 331.
A risk value field 50 for each of the grid cells 2121, 2122, 2123, 2124 is generated with an interpolation module 5 on the basis of the data records 2221, 2222, 2223, 2224 linked to the specific grid cell 2121, 2122, 2123, 2124, and a probability 521 is assigned 52 to the interpolation module 5 each point in the grid 212 based on the probability of an accident occurring at a specific geographical location and at a specific time.
Reference list
[0047]<tb> 1 <SEP> System for determining absolute and relative risks of car accidents<tb> 2 <SEP> registration unit<tb> <SEP> 21 <SEP> Geographical area<tb><SEP> <SEP> 212 <SEP> Spatially high resolution grid<tb><SEP><SEP> <SEP> 2121, 2122, 2123, 2124 <SEP> grid cells<tb><SEP><SEP> <SEP> 2221, 2222, 2223, 2224 <SEP> Generated data record<tb> 3 <SEP> release module<tb> <SEP> 30 <SEP> Morphological function memory<tb> <SEP> 31 <SEP> hash sections<tb><SEP> <SEP> 311, 312, 313 etc. <SEP> Selectable morphological traffic model function<tb> <SEP> 32 <SEP> trip parameter table<tb><SEP> <SEP> 321, 322, 323 etc. <SEP> tripping parameters with threshold values<tb> <SEP> 33 <SEP> scaling table<tb><SEP> <SEP> 331 <SEP> Recorded actual accident data<tb> 4 <SEP> measuring and tripping units<tb> <SEP> 40 <SEP> pattern triggers<tb><SEP> <SEP> 401 <SEP> High-resolution density data<tb><SEP><SEP> <SEP> 4001, 4002, 4003, 4004 <SEP> population density parameters<tb><SEP><SEP> <SEP> 4011, 4012, 4013, 4014 <SEP> weighting factors<tb> <SEP> 41 <SEP> First air-based measuring stations<tb><SEP> <SEP> 411 <SEP> First high-resolution data<tb><SEP><SEP> <SEP> 4101, 4102, 4103, 4104 <SEP> Soil coverage parameters<tb> <SEP> 42 <SEP> Second air-based measuring stations<tb><SEP> <SEP> 421 <SEP> Second high-resolution data<tb><SEP><SEP> <SEP> 4201, 4202, 4203, 4204 <SEP> night light parameters<tb><SEP><SEP> <SEP> 4211, 4212, 4213, 4214 <SEP> Weighted indicator parameters<tb> <SEP> 43 <SEP> Geodetic measuring stations<tb><SEP> <SEP> 431 <SEP> Third high-resolution data<tb><SEP><SEP> <SEP> 4301, 4302, 4303, 4304 <SEP> road map parameters<tb><SEP><SEP> <SEP> 4311, 4312, 4313, 4314 <SEP> Classification parameters of road map parameters<tb> <SEP> 44 <SEP> Third air-based measuring stations<tb><SEP> <SEP> 441 <SEP> Fourth high-resolution data<tb><SEP><SEP> <SEP> 4401, 4402, 4403, 4404 <SEP> precipitation parameters<tb><SEP><SEP> <SEP> 4411, 4412, 4413, 4414 <SEP> Hydrological cycle parameters<tb><SEP><SEP> <SEP> 4421, 4422, 4423, 4424 <SEP> precipitation distribution factors<tb><SEP><SEP> <SEP> 4431, 4432, 4433, 4434 <SEP> quantity factors<tb><SEP><SEP> <SEP> 4441, 4442, 4443, 4444 <SEP> strength factors<tb> <SEP> 45 <SEP> Fourth air-based measuring stations<tb><SEP> <SEP> 451 <SEP> Fifth high-resolution data<tb><SEP><SEP> <SEP> 4501, 4502, 4503, 4504 <SEP> Digital height parameters<tb><SEP><SEP> <SEP> 4511, 4512, 4513, 4514 <SEP> measuring factors for the terrain height<tb> 5 <SEP> interpolation module<tb> 6 <SEP> data transmission network<tb> <SEP> 61 <SEP> activation device<tb> <SEP> 62 <SEP> alarm device<tb> <SEP> 63 <SEP> Mobile access device<tb> <SEP> 64 <SEP> input / output device

权利要求:
Claims (12)
[1]
1. System (1) for the automated, location-dependent prediction of absolute and relative risks of car accidents exclusively on the basis of non-insurance-related data, records of accident events being able to be generated and location-dependent probability values for specific accident conditions associated with the risk of a car accident being ascertainable, characterized in thatthat a spatially resolving grid (212) with grid cells (2121, 2122, 2123, 2124) can be generated over a geographic area (21) of interest with a detection unit (2), the area (21) comprising at least a part of units (70- 74) which are exposed to an accident risk, the grid cells (2121, 2122, 2123, 2124) of the grid (212) being selectable and data about the system (1) for each cell (2121, 2122, 2123, 2124) of the grid (212) can be assigned, and wherein data records representative of a raster cell are assigned to a year of occurrence or measurement and are stored in a memory module of a computing unit,that for each raster cell (2121, 2122, 2123, 2124) an ambient population density parameter (4001, 4002, 4003, 4004) can be recorded with a settlement pattern trigger (40) and a data set assigned to corresponding raster cells (2121, 2122, 2123, 2124) (2221, 2222, 2223, 2224) can be assigned, population density parameters (4001, 4002, 4003, 4004) being recorded for the geographic area (21) of interest and suitable weighting factors (4011, 4012, 4013, 4014) in the spatial grid (212) taking into account the various settlement parameters, and the surrounding population density parameters (4001, 4002, 4003, 4004) using the system (1) of spatially resolving air monitoring data (401, 411, 412, 414, 415) measured with satellite or measuring stations are extractable, based on measured interaction between the surrounding population density parameters (4001, 4002, 4003, 4004) and / or land use parameters and measured Fah r or traffic patterns,that first detected, spatially resolving air monitoring data (411) can be transmitted to the system (1) by first air-based measuring stations (41) and ground covering parameters (4101, 4102, 4103, 4104) are assigned to the corresponding grid cells (2121, 2122, 2123, 2124) Data record (2221, 2222, 2223, 2224) can be generated and stored on the basis of the first spatially resolving air monitoring data (411), the ground cover parameters (4101, 4102, 4103, 4104) being a measure of the observable biophysical cover on the earth's surface,that second spatially resolving data (421), which measures the luminance and is recorded by second air-based measuring stations (42), can be transmitted to the system (1) and night light parameters (4201, 4202, 4203, 4204) with the corresponding grid cells (2121, 2122, 2123, 2124) assigned data record (2221, 2222, 2223, 2224) can be generated and stored on the basis of the second spatially resolving air monitoring data (421) on the luminance, the night light parameters (4201, 4202, 4203, 4204) based on their weighted proxy ( 4211, 4212, 4213, 4214) for local activity and correlation with other proxy measurements for local common good,that third spatially resolving data (431) recorded by systematically operated geodetic measuring stations (43) can be transmitted to the system (1) and that road map parameters (4301, 4302, 4303, 4304) are assigned to the corresponding grid cells (2121, 2122, 2123, 2124) Data record (2221, 2222, 2223, 2224) can be generated and stored on the basis of the third spatially resolving data (431) from the geodetic measuring stations (43), the road map parameters (4301, 4302, 4303, 4304) having at least one classification parameter (4311, 4312, 4313, 4314) to indicate a type of the assigned street,that fourth measured spatially resolving air monitoring data (441) recorded by air-based measuring stations (44) can be transmitted to the system (1) and precipitation parameters (4401, 4402, 4403, 4404) with the data record assigned to the corresponding grid cells (2121, 2122, 2123, 2124) (2221, 2222, 2223, 2224) can be generated and stored on the basis of the fourth spatially resolving air monitoring data (441), the precipitation parameters (4401, 4402, 4403, 4404) being a measure of the hydrological cycle (4411, 4412, 4413, 4414) include that at least distribution (4421, 4422, 4423, 4424), amounts (4431, 4432, 4433, 4434) and strength (4441, 4442, 4443, 4444) of local precipitation at a specific point or area the corresponding grid cell (2121, 2122, 2123, 2124) indicates,that fifth spatially resolving air monitoring data (451) measured by fourth air-based measuring stations (45) can be transmitted to the system (1) and digital height parameters (4501, 4502, 4503, 4504) with the data set assigned to the grid cells (2121, 2122, 2123, 2124) (2221, 2222, 2223, 2224) can be generated and stored on the basis of the fifth spatially resolving air monitoring data (451), the digital height parameters (4501, 4502, 4503, 4504) generated being a measure (4511, 4512, 4513, 4514) for the terrain height at a specific point or in a specific area of the corresponding grid cell 2121, 2122, 2123, 2124) for providing a representation of the terrain surface,that the system (1) comprises a trigger module (3) with a hash table (31) with a plurality of selectable morphological traffic model functions (311, 312, 313), whereby for each grid cell (2121, 2122, 2123, 2124) the grid cells ( Data records (2221, 2222, 2223, 2224) assigned to 2121, 2122, 2123, 2124) can be filtered by predefined triggering parameters (321, 322, 323), around threshold values of the generated population density parameters (4001, 4002, 4003, 4004), land cover parameters (4101 , 4102, 4103, 4104), night light parameters (4201, 4202, 4203, 4204), street map parameters (4301, 4302, 4303, 4304), precipitation parameters (4401, 4402, 4403, 4404) and digital altitude parameters (4501, 4502, 4503, 4504), whereby the morphological traffic model functions (311, 312, 313) can be compared with a scaling table (33) on the basis of recorded actual accident data (331), and wherein a specific morphological traffic model function (311, 312, 313 ) can be triggered and selected by the best comparison with the accident data (331), andthat a risk value field (50) for each of the raster cells (2121, 2122, 2123, 2124) with an interpolation module (5) on the basis of the data records (2221, 2222, 2223, 2221, 2122, 2123, 2124) assigned to the specific 2224) can be generated (51), and a probability (521) can be assigned (52) to the interpolation module (5) each point in the grid (212) on the basis of the probability of an accident occurring at a specific geographical location and at a specific time.
[2]
2. System (1) according to claim 1, characterized in that the high-resolution air monitoring data (401, 411, 412, 414, 415) comprise aerial images and / or satellite images and / or aerial photos.
[3]
3. System (1) according to claim 1 or 2, characterized in that the high-resolution air monitoring data (401, 411, 412, 414, 415) of satellites and / or aircraft and / or aircraft lighter than air or other equipped with a balloon Measuring stations include measured aerial images and / or satellite images and / or aerial photos.
[4]
4. System (1) according to claim 1 to 3, characterized in that the weighted proxy (4211, 4212, 4213, 4214) for other proxy measurements for local common good comprises highly local mass for human well-being and / or the national or sub-national gross domestic product GDP .
[5]
5. System (1) according to claim 1 to 4, characterized in that the third high-resolution data (431) are selected with a data extraction from an accessible high-resolution road map database.
[6]
6. System (1) according to claim 1 to 5, characterized in that geodetic measuring stations (43) comprise a global positioning system unit, GPS unit, or can be located by satellite imaging.
[7]
7. System (1) according to claim 1 to 6, characterized in that the classification parameters (4311, 4312, 4313, 4314) of the road map parameters (4301, 4302, 4303, 4304) values for classifying bike paths, sidewalks, highways, paths, Pedestrians, main streets, residential streets, side streets, steps, supply lines, tertiary lanes and unclassifiable street objects.
[8]
8. System (1) according to claim 7, characterized in that the classification parameters (4311, 4312, 4313, 4314) comprise tag elements that enable attributes of the classification.
[9]
9. System (1) according to claim 1 to 8, characterized in that the classification parameters (4311, 4312, 4313, 4314) comprise a measure of an average speed of a road user at the specific point of the grid cell (2121, 2122, 2123, 2124).
[10]
10. System (1) according to claim 1 to 9, characterized in that precipitation parameters (4401, 4402, 4403, 4404) comprise at least parameters for measuring the precipitation of rain and / or snow and / or hail.
[11]
11. System (1) according to claim 1 to 10, characterized in that the digital height parameters (4501, 4502, 4503, 4504) further comprise morphological elements.
[12]
12. A method for the automated, location-dependent prediction of absolute and relative risks of car accidents exclusively on the basis of non-insurance-related data, data records of accident events being generated and location-dependent probability values being determined for specific accident conditions associated with the risk of a car accident, characterized in thatthat a spatially resolving grid (212) with grid cells (2121, 2122, 2123, 2124) is generated over a geographic area (21) of interest with a detection unit (2), the area comprising at least a part of units (70-74) , which are exposed to an accident risk, the grid cells (2121, 2122, 2123, 2124) of the grid (212) being selectable and data being assignable with the system to each cell (2121, 2122, 2123, 2124) of the grid (212), and wherein data records representative of a raster cell are assigned to a year of occurrence or measurement and are stored in a memory module of a computing unit,that for each raster cell (2121, 2122, 2123, 2124) an ambient population density parameter (4001, 4002, 4003, 4004) is recorded with a settlement pattern trigger (40) and a data record assigned to the corresponding raster cells (2121, 2122, 2123, 2124) 2221, 2222, 2223, 2224) is assigned, whereby population density parameters (4001, 4002, 4003, 4004) are recorded for the geographic area (21) of interest and suitable weighting factors (4011, 4012, 4013, 4014) in the spatial grid (212) below Taking into account the various settlement parameters, and the surrounding population density parameters (4001, 4002, 4003, 4004) can be extracted using the system (1) of spatially resolving air monitoring data (401, 411, 412, 414, 415) measured with satellites or measuring stations are based on the measured interaction between the surrounding population density parameters (4001, 4002, 4003, 4004) and / or land use parameters and the measured driving mode r traffic patterns,that first recorded spatially resolving air monitoring data (411) are transmitted to the system by first air-based measuring stations (41), and ground covering parameters (4101, 4102, 4103, 4104) with the data record (2221 assigned to the corresponding grid cells (2121, 2122, 2123, 2124) , 2222, 2223, 2224) are generated and stored on the basis of the first spatially resolving air monitoring data (411), the ground cover parameters (4101, 4102, 4103, 4104) being a measure of the observable biophysical cover on the earth's surface,that second spatially resolving data (421) on luminance recorded by second air-based measuring stations (42) are transmitted to the system (1) and night light parameters (4201, 4202, 4203, 4204) with the corresponding raster cells (2121, 2122, 2123, 2124) assigned data record (2221, 2222, 2223, 2224) are generated and stored on the basis of the second spatially resolving air monitoring data (421) on the luminance, the night light parameters (4201, 4202, 4203, 4204) on the basis of their weighted proxy (4211, 4212, 4213, 4214) for local activity and correlation with other proxy measurement throws for local common good,that third spatially resolving data (431) recorded by systematically operated geodetic measuring stations (43) are transmitted to the system (1) and road map parameters (4301, 4302, 4303, 4304) are assigned to the corresponding grid cells (2121, 2122, 2123, 2124) Data record (2221, 2222, 2223, 2224) is generated and stored on the basis of the third spatially resolving data (431) from the geodetic measuring stations (43), the road map parameters (4301, 4302, 4303, 4304) at least one classification parameter (4311, 4312, 4313, 4314) to indicate a type of the assigned street,that fourth measured, spatially resolving air monitoring data (441) recorded by air-based measuring stations (44) are transmitted to the system (1) and precipitation parameters (4401, 4402, 4403, 4404) are assigned to the corresponding grid cells (2121, 2122, 2123, 2124) Data set (2221, 2222, 2223, 2224) is generated and stored on the basis of the fourth spatially resolving air monitoring data (441), the precipitation parameters (4401, 4402, 4403, 4404) being a measure of the hydrological cycle (4411, 4412, 4413 , 4414) include that at least distribution (4421, 4422, 4423, 4424), amounts (4431, 4432, 4433, 4434) and strength (4441, 4442, 4443, 4444) of local precipitation at a specific point or in a specific one Area of the corresponding grid cell (2121, 2122, 2123, 2124) indicatesthat fifth measured spatially resolving air monitoring data (451) recorded by fourth air-based measuring stations (45) are transmitted to the system (1) and digital height parameters (4501, 4502, 4503, 4504) are assigned to the grid cells (2121, 2122, 2123, 2124) Data record (2221, 2222, 2223, 2224) is generated and stored on the basis of the fifth spatially resolving air monitoring data (451), the digital altitude parameters (4501, 4502, 4503, 4504) being generated a measure (4511, 4512, 4513, 4514 ) for the terrain height at a specific point or in a specific area of the corresponding grid cell 2121, 2122, 2123, 2124) for providing a representation of the terrain surface,that the system (1) comprises a trigger module (3) with a hash table (31) with a plurality of selectable morphological traffic model functions (311, 312, 313), the assigned data records (2121, 2122, 2123, 2124) for each raster cell (2121, 2122, 2123, 2124) 2221, 2222, 2223, 2224) are filtered by predefined trigger parameters (321, 322, 323), around threshold values of the generated population density parameters (4001, 4002, 4003, 4004), ground cover parameters (4101, 4102, 4103, 4104), night light parameters (4201 , 4202, 4203, 4204), road map parameters (4301, 4302, 4303, 4304), precipitation parameters (4401, 4402, 4403, 4404) and digital altitude parameters (4501, 4502, 4503, 4504), whereby the morphological traffic model functions (311, 312, 313) with a scaling table (33) on the basis of recorded actual accident data (331), and wherein a specific morphological traffic model function (311, 312, 313) by the best comparison with the Un case data (331) is triggered and selected, andthat a risk value field (50) for each of the grid cells (2121, 2122, 2123, 2124) with an interpolation module (5) on the basis of the data records (2221, 2222, 2223) assigned with the specific grid cell (2121, 2122, 2123, 2124) , 2224) is generated (51), and a probability (521) with (52) is assigned to the interpolation module (5) each point in the grid (212) on the basis of the probability of an accident occurring at a specific geographical location and at a specific time .

类似技术:

---

公开号 | 公开日 | 专利标题

---

CH714036B1|2020-03-31|System and method for predicting absolute and relative risks of car accidents.

---

DE60319993T2|2009-04-09|DEVICE AND METHOD FOR TRANSPORT INFORMATION PROCESSING

---

DE102007003147B4|2013-06-06|Driving Assistance System

---

DE10261826B4|2007-05-31|Navigation device and method for searching for a route

---

Tam et al.2008|Using automatic vehicle idenification data for travel time estimation in Hong Kong

---

Saenz-de-Miera et al.2012|The responsibility of tourism in traffic congestion and hyper-congestion: A case study from Mallorca, Spain

---

EP1489552B1|2007-05-02|Method and apparatus for improving the identification and/or re-identification of objects in image processing

---

EP1099203B1|2003-09-17|Method for detecting a vehicle traffic status and system for detecting said traffic status

---

Makarova et al.2019|Improving Road Safety by Affecting Negative Factors.

---

de Vos et al.2018|Parking occupancy and external walking costs in residential parking areas

---

Salisu et al.2020|Traffic survey analysis: implications for road transport planning in Nigeria

---

D'Ambrosio2011|Methodology for collecting vehicle occupancy data on multi-lane interstate highways: a ga 400 case study

---

Moudon et al.2005|A transportation-efficient land use mapping index

---

Dheenadayalu et al.2004|Analysis of link capacity estimation methods for urban planning models

---

Basuki2017|The Willingness to Walk of Urban Transportation Passengers |

---

Nissen et al.2020|How does weather affect the use of public transport in Berlin?

---

Laoye et al.2016|Indices of Traffic Congestion on Major Roads in Akure, a Developing City in Nigeria

---

Bashbash2017|Study to Establish Traffic Statistical Records in Gaza City, Palestine

---

Smith2018|Development and Evaluation of a Prototype Travel Time Reliability Monitoring System for Freeway Facilities in North Carolina

---

Rondon2014|The impact of weather conditions on urban travel speed using ANPR observations.

---

Zhang et al.2014|Develop a Plan to Collect Pedestrian Infrastructure and Volume Data for Future Incorporation into Caltrans Accident Surveillance and Analysis System Database

---

Bashit et al.2018|Geographic information system for mapping potential of traffic congestion |

---

Saw et al.2019|Travel time variability analysis based on level of service: a case study of mixed traffic urban arterial corridors

---

Das et al.2013|Defining Level of Service Criteria of Urban Streets Using Neural Gas Clustering

---

Thill et al.2011|Comparing approaches to winter highway maintenance operations through user mobility performance

同族专利:

---

公开号 | 公开日

---

US20190283745A1|2019-09-19|

---

CN109952592A|2019-06-28|

---

WO2018083352A1|2018-05-11|

---

EP3535728A1|2019-09-11|

---

WO2018082784A1|2018-05-11|

---

EP3535727A1|2019-09-11|

---

US10850731B2|2020-12-01|

---

CN109923575A|2019-06-21|

---

US20190325740A1|2019-10-24|

---

CH714036A1|2018-05-11|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

---

US6186793B1|1995-11-07|2001-02-13|Randall E. Brubaker|Process to convert cost and location of a number of actual contingent events within a region into a three dimensional surface over a map that provides for every location within the region its own estimate of expected cost for future contingent events|

---

US7809586B2|2002-11-27|2010-10-05|Computer Sciences Corporation|Computerized method and system for estimating an effect on liability using a comparison of the actual speed of a vehicle in an accident and time and distance traveled by the vehicles in a merging vehicle accident|

---

US10157422B2|2007-05-10|2018-12-18|Allstate Insurance Company|Road segment safety rating|

---

US20140095214A1|2012-10-03|2014-04-03|Robert E. Mathe|Systems and methods for providing a driving performance platform|

---

CA2835574A1|2012-11-28|2014-05-28|Craig Arnatt|Method and system for the creation and delivery to a user of an automated, real time, online insurance policy and certificate of insurance for rental of a venue and/or equipment|

---

US10262373B2|2013-06-07|2019-04-16|State Farm Mutual Automobile Insurance Company|Systems and methods for grid-based insurance rating|

---

US20140372150A1|2013-06-14|2014-12-18|Hartford Fire Insurance Company|System and method for administering business insurance transactions using crowd sourced purchasing and risk data|

---

US10169821B2|2013-09-20|2019-01-01|Elwha Llc|Systems and methods for insurance based upon status of vehicle software|

---

US9305317B2|2013-10-24|2016-04-05|Tourmaline Labs, Inc.|Systems and methods for collecting and transmitting telematics data from a mobile device|

---

US20160189303A1|2014-03-21|2016-06-30|Gil Emanuel Fuchs|Risk Based Automotive Insurance Rating System|

---

US10170005B1|2017-09-11|2019-01-01|Honeywell International Inc.|Vehicle conflict detection|US10339544B2|2014-07-02|2019-07-02|WaitTime, LLC|Techniques for automatic real-time calculation of user wait times|

---

US11087403B2|2015-10-28|2021-08-10|Qomplx, Inc.|Risk quantification for insurance process management employing an advanced decision platform|

---

US10489976B2|2017-08-11|2019-11-26|Jing Jin|Incident site investigation and management support system based on unmanned aerial vehicles|

---

US10948305B1|2017-11-30|2021-03-16|United States Automobile Association |Driving assistance systems and methods|

---

US11248922B2|2018-11-01|2022-02-15|International Business Machines Corporation|Personalized social navigation coach|

---

CN111724596A|2020-06-23|2020-09-29|上海电科智能系统股份有限公司|Method for intelligently, accurately and automatically identifying and prejudging bottleneck region of expressway|

---

CN112242059B|2020-09-30|2021-10-01|南京航空航天大学|Intelligent decision-making method for unmanned vehicle based on motivation and risk assessment|

---

CN112382091A|2020-11-11|2021-02-19|北京世纪高通科技有限公司|Road water accumulation early warning method and device|

---

CN112541408B|2020-11-30|2022-02-25|北京深睿博联科技有限责任公司|Feasible region identification method, device, equipment and computer readable storage medium|

法律状态:

优先权:

---

申请号 | 申请日 | 专利标题

---

PCT/EP2016/076783|WO2018082784A1|2016-11-07|2016-11-07|System and method for predicting of absolute and relative risks for car accidents|

---