Apparatus and method for automated traffic and driving pattern recognition and location-dependent measurement of absolute and/or relative risk probabilities for car accidents

A method for automated location-dependent forecasting of absolute and relative risks for car accidents based on exclusively non-insurance related data by means of a measuring apparatus, wherein data records of accident events are generated and location-dependent probability values for specific accident conditions associated with the risk of car accident are determined, wherein

• a spatial high-resolution grid with grid cells is generated over a geographical area of interest by means of a capturing unit, said area including at least a portion of units exposed to accident risks, wherein the grid cells of the grid are selectable and data are assignable via the apparatus to each cell of the grid, and wherein data records representative of a grid cell are assigned to a year of occurrence or measurement and are saved in a memory module of a processing unit,
• for each grid cell static condition parameters are measured, at least comprising static road condition parameters, and dynamic condition parameters are measured, at least comprising temporal fluctuation parameters of weather conditions and temporal fluctuation parameters of temperature and temporal fluctuation parameters of daylight intensity and temporal fluctuation parameters of traffic condition, by means of measuring stations for a geographical area of interest, and transferred to the measuring apparatus, and assigned to generated data records assigned to the corresponding grid cells,
• the measured condition parameters at least comprise 6 types of condition parameters, wherein a first type of condition parameter provides a measure for a point in time for an accident, a second type of condition parameter provides a measure for a structural environmental parameter, a third type of condition parameter provides a measure for a temporal traffic condition parameter, a forth type of condition parameter provides a measure for a temporal road condition parameter, a fifth type of condition parameter provides a measure for a temporal weather condition parameter, and a sixth type of condition parameter provides a measure for a vehicle type parameter,
• for each grid cell and for each type of condition parameter of a specific data record customized weighting factors are assigned in said spatial high-resolution grid accounting for diverse condition patterns,
• the measuring apparatus comprises a machine-learning or machine-driven trigger module, wherein, by means of pattern recognition of the weighted, diverse condition patterns, a probability value is triggered and/or assigned by means of an interpolation module and the machine-learning or machine-driven trigger module to each point in said grid, giving a measure for the probability of the occurrence of an accident at a given geographical location and time.

A measuring apparatus for automated traffic and driving pattern recognition and location-dependent measurement of absolute and/or relative risk probabilities for car accidents based on geospatial contexting and based on exclusively non-insurance related data, wherein data records of accident events are generated and location dependent probability values for specific accident conditions associated with the risk of car accident are determined, wherein

• a spatial high-resolution grid with grid cells is generated over a geographical area of interest by means of a capturing unit, said area including at least a portion of units exposed to accident risks, wherein the grid cells of the grid are selectable and data are assignable via the measuring apparatus to each cell of the grid, and wherein data records representative of a grid cell are assigned to a year of occurrence or measurement and are saved in a memory module of a calculating unit,
• for each grid cell, an ambient population density parameter is captured by means of a settlement pattern trigger and assigned to a generated data record assigned to the corresponding grid cells, wherein population density parameters are captured for the geographical area of interest and customized weighting factors are assigned in said spatial high-resolution grid accounting for diverse settlement patterns,
• first measured aerial high-resolution data captured by first air-based measuring stations are transferred to the measuring apparatus and land cover parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said first aerial high-resolution data, wherein the land cover parameters are a measure for the observable bio-physical cover on the earth's surface,
• second high-resolution data on light density captured by second air-based measuring stations are transferred to the measuring apparatus and nighttime light parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said second aerial high-resolution data on light density, wherein said nighttime light parameters are generated based on their weighted proxy for local activity and correlation to other welfare proxy measures,
• third high-resolution data captured by systematically performing ground survey measuring stations are transferred to the measuring apparatus and road map parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said third high-resolution data of the ground survey measuring stations, wherein the road map parameters comprise at least one classification parameter indicating a type of the assigned road,
• fourth measured aerial high-resolution data captured by air-based measuring stations are transferred to the measuring apparatus and precipitation parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said fourth aerial high-resolution data, wherein the generated precipitation parameters comprise a measure of the hydrological cycle giving at least the local precipitation's distribution, amounts and intensity at a specific point or area of the corresponding grid cell,
• fifth aerial high-resolution data measured by fourth air-based measuring stations are transferred to the measuring apparatus and digital elevation parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said fifth aerial high-resolution data, wherein the generated digital elevation parameters comprise a measure for the terrain elevation at a specific point or area of the corresponding grid cell providing a representation of the terrain's surface,
• the measuring apparatus comprises a trigger module with a hash table with a plurality of selectable morphological traffic model-function, wherein for each grid cell, the generated data records are filtered by predefined trigger parameters triggering threshold values of the generated population density parameters, the land cover parameters, the nighttime light parameters, the road map parameters, precipitation parameters, and the digital elevation parameters, wherein the morphological traffic model-functions are matched by means of a scaling table based on captured actual accident data, and wherein a specific morphological traffic model-function is triggered and selected by best matching to the accident data, and
• a risk-value field for each of the grid cells is generated by means of an interpolation module based on the data records associated with the specific grid cell, and a probability is assigned by means of the interpolation module to each point in said grid, giving the probability of the occurrence of an accident at a given geographical location and time.

A method for automated traffic and driving pattern recognition and location-dependent measurement of absolute and/or relative risk probabilities for car accidents based on geospatial contexting and based on exclusively non-insurance related data, wherein data records of accident events are generated and location-dependent probability values for specific accident conditions associated with the risk of car accident are determined, wherein

• a spatial high-resolution grid with grid cells is generated over a geographical area of interest by means of a capturing unit, said area including at least a portion of units exposed to accident risks, wherein the grid cells of the grid are selectable and data are assignable by means of the measuring apparatus to each cell of the grid, and wherein data records representative of a grid cell are assigned to a year of occurrence or measurement and are saved in a memory module of a calculating unit,
• for each grid cell, an ambient population density parameter is captured by means of a settlement pattern trigger and assigned to a generated data record assigned to the corresponding grid cells, wherein population density parameters are captured for the geographical area of interest and customized weighting factors are assigned in said spatial high-resolution grid accounting for diverse settlement patterns,
• first measured aerial high-resolution data captured by first air-based measuring stations are transferred to the measuring apparatus and land cover parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said first aerial high-resolution data, wherein the land cover parameters are a measure for the observable bio-physical cover on the earth's surface,
• second high-resolution data on light density captured by second air-based measuring stations are transferred to the measuring apparatus and nighttime light parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said second aerial high-resolution data on light density, wherein said nighttime light parameters are generated based on their weighted proxy for local activity and correlation to other welfare proxy measures,
• third high-resolution data captured by systematically performing ground survey measuring stations are transferred to the measuring apparatus and road map parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said third high-resolution data of ground survey measuring stations, wherein the road map parameters comprise at least one classification parameter indicating a type of the assigned road,
• fourth measured aerial high-resolution data captured by air-based measuring stations are transferred to the measuring apparatus and precipitation parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said fourth aerial high-resolution data, wherein the generated precipitation parameters comprise a measure of the hydrological cycle giving at least the local precipitation's distribution, amounts and intensity at a specific point or area of the corresponding grid cell,
• fifth measured aerial high-resolution data captured by fourth air-based measuring stations are transferred to the measuring apparatus and digital elevation parameters are generated and stored by means of the generated data record assigned to the corresponding grid cells based on said fifth aerial high-resolution data, wherein the generated digital elevation parameters comprise a measure for the terrain elevation at a specific point or area of the corresponding grid cell providing a representation of the terrain's surface,
• the measuring apparatus comprises a trigger module with a hash table with a plurality of selectable morphological traffic model-functions, wherein for each grid cell, the generated data records are filtered by predefined trigger parameters triggering threshold values of the generated population density parameters, the land cover parameters, the nighttime light parameters, the road map parameters, the precipitation parameters, and the digital elevation parameters, wherein the morphological traffic model-functions are matched by means of a scaling table based on captured actual accident data, and wherein a specific morphological traffic model-function is triggered and selected by best matching to the accident data, and
• a risk-value field for each of the grid cells is generated by means of an interpolation module based on the data records associated with the specific grid cell, and a probability is assigned by means of the interpolation module to each point in said grid, giving the probability of the occurrence of a accident at a given geographical location and time.