Methods, devices, systems and computer program products for estimating road surface temperature

By using vehicle sensors to establish a relationship between ambient conditions and road surface temperature, the method addresses the reliance on road weather stations, providing accurate temperature estimation without direct road sensor data.

DE102019135751B4Active Publication Date: 2024-11-14NIRA DYNAMICS AB
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
DE102019135751
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Filing Date
2019-12-23
Publication Date
2024-11-14
Estimated Expiration
2039-12-23

AI Technical Summary

Technical Problem

Existing methods for estimating road surface temperature rely heavily on the availability of road weather stations, which can be unreliable due to geographical limitations, temporary unavailability, or data restrictions, leading to inaccurate or unavailable temperature estimates.

Method used

A method utilizing vehicle-mounted environmental sensors to collect data on ambient conditions and establish a relationship with road surface temperature through neural networks or filtering methods, enabling estimation without direct road sensor data.

Benefits of technology

Enables accurate estimation of road surface temperature even in areas lacking road weather stations by leveraging comprehensive environmental data, improving reliability and reducing processing requirements in vehicles.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

Method for estimating a road surface temperature of a geographical location (10,11), comprising: - Retrieving (31) reference environmental sensor data providing information about environmental conditions near mobile vehicles (13,14) from a data warehouse, - Retrieving (32) reference road sensor data, which provide at least information about a road surface temperature in the vicinity of distributed road sensors (12), from the data warehouse, - Establishing (33) a relationship between the reference environmental sensor data and the road surface temperature based on the reference environmental sensor data and the reference road sensor data, - Generating (34) a data structure (22) encoding the established relationship, - Collecting (35) current vehicle environment sensor data that provide information about current environmental conditions in the vicinity of one or more vehicles, in which one or more vehicles, - Determine (36) a road surface temperature estimate for the geographic location using the data structure and the collected current vehicle environment sensor data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical area

[0001] The present invention generally relates to estimating a road surface temperature of a geographic location. Background of the invention

[0002] Modern vehicles are equipped with a multitude of sensors that provide direct or indirect information about the current conditions of the vehicle and its surroundings, for example, to adapt vehicle behavior accordingly. One parameter of interest is the current temperature of the road surface, as this can influence tire-road interaction (e.g., friction) and thus affect vehicle behavior, for example, during ABS braking.

[0003] However, ordinary vehicles typically do not have sensors that allow direct measurement of road surface temperature. To estimate the surface temperature of a road or roadway section at a given geographical location, conventional approaches rely on road weather stations, such as the Road Weather Information System (RWIS), which uses a plurality of environmental sensor stations distributed along roads to take measurements and reports these measurements to a central server for analysis (e.g., using a model) and distribution (e.g., to vehicles).

[0004] Consequently, these conventional approaches rely on the presence and availability of data from road weather stations, with the accuracy of estimation being related to the geographical density of road weather stations. Achieving high accuracy requires a high density of road weather stations. However, these data may be unavailable for numerous reasons, such as remote roads not being equipped with RWIS stations; RWIS stations being temporarily unavailable for technical, economic, or environmental reasons; an RWIS data provider can only make the data available to a limited group of participants; the data are only representative of a limited geographical area; or a variety of other reasons.

[0005] For global coverage, a user of RWIS data may need to obtain such data from multiple local providers, and the user may need several different models describing the dependencies at each geographic location.

[0006] As can be seen from these examples, a common problem with road surface temperature estimation is its reliance on the availability of road weather stations. Therefore, there is great interest in providing alternative ways to estimate road surface temperature.

[0007] Therefore, an object of the present invention is to provide improved solutions for estimating a road surface temperature of a geographical location.

[0008] DE 10 2017 214 053 A1 relates to methods for determining a coefficient of friction for a contact between a tire of a vehicle and a road surface and methods for controlling a vehicle function of a vehicle.

[0009] DE 10 2018 203 807 A1 concerns methods and devices for detecting and evaluating road conditions and weather-related environmental influences.

[0010] DE 10 2018 008 788 A1 concerns methods for determining road surface roughness.

[0011] JP 2002- 48 878 A relates to methods and apparatus for predicting road surface temperature or road surface condition. Summary of the invention

[0012] Methods, apparatus, systems and computer program products are disclosed to at least partially overcome the deficiencies of the above and others.

[0013] According to a first aspect, a method for estimating a road surface temperature of a geographical location is disclosed.

[0014] The method comprises retrieving reference environmental sensor data and retrieving reference road sensor data. The environmental sensor data provides information about environmental conditions in the vicinity of mobile vehicles, and the reference road sensor data provides information about at least a road surface temperature in the vicinity of distributed road sensors. The retrieval can be performed, in particular, in a data warehouse (central database).

[0015] Data indicative of environmental conditions is such that it can be collected using at least vehicle sensors. Non-limiting examples of environmental conditions may include ambient air temperature, solar radiation, GPS location, time of day, date, humidity, barometric pressure, wind speed, rain intensity, vehicle speed, fog, surface water amount, snow amount, presence of ice, traffic intensity (e.g., vehicles per unit time), wind direction, salt amount, surface type (such as asphalt, concrete, gravel, etc.), and / or an estimated coefficient of friction. Preferably, several of the above-mentioned environmental conditions are used. The use of multiple environmental conditions increases the data base upon which a relationship can be built, thereby increasing the reliability and accuracy of the road surface temperature estimation using such comprehensive information.

[0016] Examples of distributed road sensors may include the aforementioned Road Weather Stations (RWIS). Other examples include sophisticated vehicle-mounted sensors that, due to their vehicle-mounted nature, are dynamically distributed across the road network and can be configured to provide data indicating road surface temperature based on a number of distributed reference vehicles.

[0017] The method further includes establishing a relationship between the reference environmental sensor data and the road surface temperature based on the reference environmental sensor data and the reference road sensor data, and generating a data structure encoding the established relationship. Examples of establishing a relationship may include, as described below, training a neural network or applying a filtering method. Such relationships and corresponding data structures may thus make it possible to specify the dependence of the road surface temperature on one or more measurable input variables.

[0018] For example, a neural network can be trained on reference environmental sensor data and reference road sensor data (including road surface temperature), so that the trained neural network can use (current) vehicle environmental sensor data as input to provide a road surface temperature as output. Such a neural network thus encodes the dependence of the road surface temperature on environmental sensor data.

[0019] The method further includes collecting current vehicle environmental sensor data indicative of current environmental conditions in the vicinity of one or more vehicles, in the one or more vehicles. Similar to the environmental conditions mentioned above with reference to the reference environmental sensor data, non-limiting examples of current environmental conditions may include ambient air temperature, solar radiation, GPS location, time of day, date, humidity, barometric pressure, wind speed, rain intensity, vehicle speed, fog, surface water amount, snow amount, presence of ice, traffic intensity (e.g., vehicles per unit time), wind direction, salt amount, surface type (such as asphalt, concrete, gravel, etc.), and / or an estimated coefficient of friction.

[0020] Preferably, the current vehicle environmental sensor data can be of the same type as the reference environmental sensor data. For example, both data sets can provide information about the ambient air temperature, regardless of the specific sensor used to collect this type of data. This makes it possible to feed the same type of data, namely (current vehicle) environmental sensor data, into the relationship established using the (reference) environmental sensor data.

[0021] The method further includes determining a road surface temperature estimate for the geographic location using the data structure and the collected current vehicle environmental sensor data. In particular, the collected current vehicle environmental sensor data may be used as the input of the data structure. The relationship encoded by the data structure thus allows a road surface temperature estimate for the geographic location to be obtained as the output.

[0022] Consequently, determining the road surface temperature estimate may not involve the use of actual road sensor data (once the relationship is established), making it possible to estimate the surface temperature even if no distributed road stations are present or available at the geographical location where the surface temperature is to be estimated.

[0023] In some embodiments, the method further comprises sending the collected current vehicle environment sensor data from the one or more vehicles to the data warehouse. This allows for centralizing the determination of the road surface temperature estimate, thereby reducing the processing requirements in the vehicles and increasing the amount of current vehicle environment sensor data by utilizing such data from multiple vehicles.

[0024] This may further enable environmental data of a geographical location to be collected from multiple vehicles, particularly over an extended period of time, to build a time series of current vehicle environmental sensor data of a geographical location. For the purposes of the present disclosure, the term "time series" refers to a set of data points that provide information about the measurable values ​​(e.g., of environmental conditions, such as ambient air temperature) at a plurality of points in time, regardless of whether all measurements are performed by the same sensor or by multiple sensors, and regardless of whether the plurality of points in time are continuous or discrete, or a combination thereof.

[0025] In particular, a time series can be measured in its entirety or partially measured and then completed by interpolation or extrapolation. An example duration of a time series can range from one day to several months, preferably between 3 days and 30 days, i.e., a duration of (approximately) 10 days. Such durations can be suitable, for example, for capturing freezing and melting behavior in the soil.

[0026] In some embodiments, the method further comprises generating road condition data based on the determined road surface temperature estimate and sending the generated road condition data to at least one of the vehicles and / or other vehicles. This enables the estimated surface temperature to be applied in a connected driving environment, for example, by performing the road surface temperature estimation remotely from the vehicles and sending information about road conditions to the vehicle(s). The vehicle(s) may use this road condition data to control vehicle behavior, for example, to consider how the road conditions would affect potential ABS braking.

[0027] In some embodiments, the retrieved reference environmental sensor data may comprise a (first) time series that provides information about environmental conditions at a (first) plurality of points in time. This makes it possible to establish the relationship based on a more comprehensive database, which in particular shows the temporal development of environmental conditions. The inventors have recognized that such a temporal development encodes valuable information about the microclimate typical of geographical locations that exhibit such a temporal development. For example, the presence of a nearby lake not only influences environmental conditions (such as the humidity in the air, which is represented by humidity measurements), but can also influence the rate of heating / cooling of the road surface temperature. Such a dependency can be encoded in the relationship, e.g.by training a neural network with time series of reference environmental sensor data. The use of time series therefore makes it possible to capture physical dynamics (such as ground freezing), which implicitly contains information about meteorological features.

[0028] Additionally or alternatively, the collected current vehicle environmental sensor data may include a (second) time series that provides information about current environmental conditions at a (second) plurality of points in time. This allows the road surface temperature estimate to be determined based on a more comprehensive database, particularly showing the temporal evolution of environmental conditions. Similar to the establishment of the relationship, such a temporal evolution encodes valuable information about the microclimate of the specific geographical location in question.

[0029] Preferably, both establishing the relationship and determining the road surface temperature estimate may be based on such time series.

[0030] In some embodiments, the relationship can be established by training a neural network, where the data structure is representative of the neural network. Alternatively, it can be established by applying a filtering method, in particular a Kalman filter, a Bayesian filter, or a non-statistical filter, where the data structure is representative of the output of the filter. This makes it possible to specify the dependence of the road surface temperature with a minimum of hypothesis or knowledge about the dependence (in contrast, for example, to a regression analysis, which requires an assumption regarding a parametric function describing such a dependence).

[0031] In some embodiments, the reference environmental sensor data, the reference road sensor data, and / or the current vehicle environmental sensor data are location-specific. For example, knowledge of the location with respect to the reference environmental sensor data and the reference road sensor data makes it easier to establish a connection between these two data sets. Likewise, knowledge of the location with respect to the current vehicle environmental sensor data makes it possible to use this additional information as another input to the relationship when determining a road surface temperature estimate.

[0032] In some embodiments, the reference road sensor data, in addition to road surface temperature, may further provide information about one or more of the following environmental conditions: ambient air temperature, ground temperature at one or more depths, solar radiation, humidity, wind speed, GPS location, time, date, barometric pressure, rain intensity, fog, surface water volume, snow volume, ice presence, traffic intensity (e.g., vehicles per unit time), wind direction, salt volume, surface type (such as asphalt, concrete, gravel, etc.), and an estimated coefficient of friction. It may be collected by one or more of a variety of means, such as dedicated sensors, cameras, weather stations, visual inspection, lidar, radar, or a fusion of data from multiple sources.This makes it possible to take into account any discrepancy between the values ​​of these environmental conditions indicated by the reference road sensor data and the values ​​of these environmental conditions indicated by the reference environmental sensor data.

[0033] In some embodiments, the method may further comprise retrieving reference weather data indicative of general weather conditions from the data warehouse and receiving current weather data indicative of current general weather conditions at the geographical location. In such cases, establishing the relationship may further be based on the reference weather data. This allows a more comprehensive relationship to be established between the road surface temperature and the input data. For example, weather data (which, unlike road data, is typically readily and globally available without significant limitations) may be used as a further input parameter to be used both during the establishment of the relationship (e.g., when training a neural network) and during the determination of a road surface temperature estimate.In such cases, in particular, determining the road surface temperature estimate may further be based on the current weather data.

[0034] Non-limiting examples of reference weather data or current weather data include ambient air temperature, sunshine, cloud cover, geographic area, humidity, wind speed, time of day, date, atmospheric pressure, rainfall intensity, fog, surface water level, snow level, presence of ice, temperature profile at different altitudes, wind direction and / or surface temperature.

[0035] According to a second aspect, a computer program product is disclosed. The computer program product comprises program code that, when executed in a computing device, is configured to perform the steps of a method according to the first aspect.

[0036] According to a third aspect, a device for estimating a road surface temperature of a geographical location is disclosed. The device comprises a data storage unit, a communication interface, and a processing unit. The data storage unit is configured to store a data structure encoding a relationship between reference environmental sensor data measurable by vehicles and a road surface temperature. The communication interface is configured to receive current vehicle environmental sensor data that provides information about current environmental conditions in the vicinity of the vehicle.

[0037] The processing unit is communicatively coupled to the vehicle sensor and the data storage unit and is configured to determine a road surface temperature estimate for the geographic location using the data structure and based on the collected current vehicle environmental sensor data.

[0038] Such a device thus makes it possible to fully exploit the benefits of the relationship encoded in the data structure, regardless of whether the relationship is established (and received by the device) by the device or by another device.

[0039] According to a fourth aspect, a data warehouse unit is disclosed for generating a data structure configured to (i) obtain reference environmental sensor data providing information about environmental conditions in the vicinity of vehicles, (ii) obtain reference road sensor data providing at least information about a road surface temperature prevailing in the vicinity of and measured by road sensors, (iii) establish a relationship between the reference environmental sensor data and the road surface temperature based on the reference environmental sensor data and the reference road sensor data, (iv) generate a data structure encoding the established relationship, and (v) send the data structure to the device. Such a data warehouse unit makes it possible to generate a data structure encoding a relationship for a plurality of users or user devices.The data structure may, for example, be sent to a device according to the third aspect or any other device configured to use the relationship encoded in the data structure.

[0040] According to a fifth aspect, a system for estimating a road surface temperature of a geographical location is disclosed. The system comprises a device according to the third aspect and a data warehouse unit according to the fourth aspect. The data warehouse unit is communicatively coupled to the device to send the data structure of the data warehouse unit to the device. Such a system enables the full benefits of all of the above aspects to be utilized, as it comprises both the data warehouse unit and the device.

[0041] In some embodiments, the system may further include a vehicle sensor configured to collect current vehicle environmental sensor data indicative of current environmental conditions near the vehicle and communicatively coupled to the device's communication interface.

[0042] According to a sixth aspect, the use of a system according to the fifth aspect for estimating a road surface temperature is disclosed. Brief description of the drawings

[0043] The following detailed description refers to the enclosed drawings in which: Fig. 1 schematically shows a situation with roads and vehicles in which methods, data warehouse units or systems according to the embodiments can be used; Fig. 2 shows a box diagram of a system including a device and a data warehouse unit according to embodiments; Fig. 3 shows a general flow diagram of a method according to the embodiments; Fig. 4 shows a general flow diagram of a method according to a particular embodiment; and Fig. 5 schematically shows a situation with roads and vehicles in which methods, devices or systems according to the embodiments can be used. Detailed description of the embodiments

[0044] Fig. 1 schematically shows a situation with roads and vehicles in which methods, data warehouse units or systems according to the present disclosure can be used.

[0045] In particular, Fig. 1, a first geographical location 10. In the illustrated example, the first geographical location 10 is a road segment. The present disclosure relates to a geographical location with a substantially arbitrary degree of granularity. A geographical location can be, for example, a federal state, a region, a municipality, a street, a road segment, an area unit of a grid square, an intersection, etc. Preferably, the geographical location is a road segment.

[0046] The road section 10 is provided with a stationary road sensor 12 which is configured to measure reference road sensor data which provide at least information about a road surface temperature at the road section 10.

[0047] Although the Fig. 1 represents a stationary measuring station, the function of the stationary road sensors can additionally or alternatively be performed by other distributed road sensors, such as sensors of a mobile reference station, e.g. a specially equipped reference vehicle.

[0048] In any case, the reference road sensor data is sent from the stationary road sensor 12 to a data warehouse unit 16 for further processing.

[0049] Fig. 1 further shows another road section 11 which is not provided with a sensor capable of measuring the road surface temperature.

[0050] Two vehicles 13, 14 approach the road section 10. Both vehicles 13, 14 are equipped with respective reference environmental sensors configured to collect reference environmental sensor data, such as an ambient air temperature sensor (not shown).

[0051] Vehicle 13 may reach road section 10 earlier than vehicle 14. Once vehicle 13 is in the vicinity of road section 10, it collects reference environmental sensor data, such as a first ambient air temperature, at a first time. The reference environmental sensor data collected by the first vehicle 13 is sent by vehicle 13 to data warehouse unit 16, for example, together with additional information such as location and / or time / date.

[0052] Once the vehicle 14 is in the vicinity of the road section 10, it also collects reference environmental sensor data, such as a second ambient air temperature, at a second time. The reference environmental sensor data collected by the second vehicle 14 is sent by the vehicle 14 to the data warehouse unit 16, for example, along with additional information such as location and / or time / date.

[0053] In other embodiments not shown, the collection of reference environmental sensor data may additionally or alternatively be performed by the road sensor 12. This allows the two data sets for establishing a relationship (i.e., reference environmental sensor data and reference road sensor data) to be collected at the same location (namely, where the sensor 12 is located), thereby increasing the accuracy of the relationship to be established.

[0054] Referring again to the description of Fig. 1, the transmission to the data warehouse unit 16 can be done wirelessly, e.g. via the Internet 15.

[0055] In the data warehouse unit 16, the data collected by the road sensor 12 and the vehicles 13, 14 are further processed, as described in detail below, to establish a relationship between them and to generate a data structure encoding this relationship.

[0056] As will be apparent from the following description, the teachings of the present disclosure thus enable such a relationship and data structure to be used to reliably predict or estimate the road surface temperature at other geographical locations nearby (such as road segment 11) or distant (such as in another country).

[0057] For example, by establishing a relationship between the road surface temperature and the vehicle sensor data of the geographic location 10, the present disclosure enables vehicle ambient sensor data to be collected near another geographic location 11 (which does not have a road sensor for measuring the road surface temperature) and to determine a road surface temperature estimate using the established relationship without requiring road sensor data.

[0058] In the example of Fig. 1, the first vehicle 13 reaches the second geographical location 11 before the second vehicle 14. This allows the first vehicle 13 to collect current vehicle environment sensor data near the geographical location 11. This data can be processed by a (not shown) Fig. 1) apparatus (or method) according to the present disclosure may be used to generate a road surface temperature estimate.

[0059] Fig. Figure 2 shows a box diagram of a system 20 for estimating a road surface temperature of a geographic location. The system 20 includes a data warehouse unit 21, a device 24, and a vehicle sensor 28. In other examples not shown, the system may not include the vehicle sensor 28.

[0060] The data warehouse unit 21 is used to create a data structure. The data warehouse unit 21 comprises a data storage unit 22 and a processing unit 23.

[0061] The data memory 22 is configured to store reference environmental sensor data (in Fig. 2 designated as "Ref data a", "Ref data b", etc.) that provide information about environmental conditions in the vicinity of vehicles, which are measured by vehicle sensors. The data memory 22 is further configured to receive reference road sensor data that provide at least information about a road surface temperature (in Fig. 2 with T ofl,a , T ofl,b etc.) that prevails near road sensors and is measured by them.

[0062] The processing unit 23 is configured to establish a relationship between the reference environmental sensor data and the road surface temperature based on the information in the data memory 22. It is further configured to generate a data structure encoding the established relationship.

[0063] The data warehouse unit 21 is communicatively coupled to the device 24 in order to send the data structure of the data warehouse unit to the device 24, in particular to a data storage unit 25 of the device.

[0064] The data warehouse unit 21 thus generates a data structure encoding a relationship between environmental data and a road surface temperature and sends it to the device 24, which is configured to use the relationship encoded in the data structure.

[0065] The device 24 is used to estimate a road surface temperature of a geographical location. The device 24 comprises a data storage unit 25, a communication interface 26, and a processing unit 27.

[0066] The data storage unit 25 is configured to store the data structure sent by the data warehouse unit 21. As stated above, the data structure encodes the relationship between reference environmental sensor data measurable by vehicles and a road surface temperature.

[0067] The communication interface 26 is used to receive current vehicle environmental sensor data, which provides information about the current environmental conditions in the vicinity of the vehicle. This current vehicle environmental sensor data is measured by one or more vehicle sensors 28, which transmit this data to the communication interface 26.

[0068] The processing unit 27 of the device is communicatively coupled to the vehicle sensor 28 and to the data storage unit 25 via the interface 26. The processing unit 27 is configured to determine a road surface temperature estimate for the geographic location using the data structure and based on the collected current vehicle environmental sensor data.

[0069] The device 24 thus serves to use the relationship encoded in the data structure to estimate the road surface temperature.

[0070] Fig. 3 shows a general flow diagram of a method 30 for estimating a road surface temperature of a geographic location.

[0071] The method 30 includes a step 31 for retrieving reference environmental sensor data and a step 32 for retrieving reference road sensor data. The reference environmental sensor data provides information about environmental conditions in the vicinity of mobile vehicles, and the reference road sensor data provides at least information about a road surface temperature in the vicinity of distributed road sensors, such as road weather stations (e.g., RWIS).

[0072] In the example shown, steps 31 and 32 are performed substantially in parallel. In other examples, these steps may be performed sequentially (first step 31 and then step 32, or first step 32 and then step 31) or only partially in parallel. In further examples, step 32 of retrieving reference road sensor data (including at least a road surface temperature in the vicinity of distributed road sensors) may further include a step (similar to step 31) of retrieving reference environmental sensor data indicative of environmental conditions (such as a time series of ambient air temperature in the vicinity of vehicles) measured by distributed road sensors.

[0073] Referring again to Fig. 3, the method 30 further includes a step 33 for establishing a relationship between the reference environmental sensor data and the road surface temperature based on the reference environmental sensor data and the reference road sensor data. It further includes a step 34 for generating a data structure encoding the established relationship.

[0074] The set of steps 31 to 34 can, for example, be performed in a data warehouse and viewed as a method for creating a relationship data structure.

[0075] The method 30 further includes a step 35 for collecting current vehicle environmental sensor data indicative of current environmental conditions in the vicinity of one or more vehicles in the one or more vehicles.

[0076] The method 30 further includes a step 36 for determining a road surface temperature estimate for the geographic location using the data structure and the collected current vehicle environmental sensor data. In particular, the collected current vehicle environmental sensor data can be used as the input of the data structure. The relationship encoded by the data structure thus allows a road surface temperature estimate of the geographic location to be obtained as the output.

[0077] As from Fig. As can be seen from Figure 3, determining 36 the road surface temperature estimate does not involve the use of actual road sensor data. Data from distributed road stations is used only in steps 32 and 33 to establish the relationship.

[0078] The set of steps 35, 36 may, for example, be performed in a network backend and viewed as using a relationship data structure to estimate a road surface temperature.

[0079] Fig. 4 shows a flowchart of a specific method 40 for estimating a road surface temperature of a geographical location. The method 40 may be used as a preferred embodiment of the method described in Fig. 3 can be considered in the more general procedure 30 shown.

[0080] The method 40 includes a step 41 for retrieving reference environmental sensor data in the form of a time series of the reference ambient air temperature. The reference environmental sensor data thus provides information about the ambient temperature condition in the vicinity of mobile vehicles at multiple points in time. In preferred examples, the reference environmental sensor data may include other variables, such as time series of humidity, air pressure, wind speed, rain intensity, fog, surface water quantity, snow quantity, ice presence, wind direction, and / or salt quantity. For illustrative purposes, however, the present example focuses on reference environmental sensor data in the form of a time series of the reference ambient air temperature.

[0081] The method 40 further includes a step 42 for retrieving reference road sensor data in the form of a time series of road surface temperature. The reference road sensor data thus provides information about the road surface temperature in the vicinity of distributed road sensors, such as road weather stations (e.g., RWIS).

[0082] In the example shown, ambient air temperature is used as an example environmental condition. Additionally or alternatively, other examples of environmental conditions may be used, including solar radiation, GPS location, time, date, humidity, barometric pressure, wind speed, rain intensity, vehicle speed, fog, surface water volume, snow volume, ice presence, traffic intensity (e.g., vehicles per unit time), wind direction, salt volume, surface type (such as asphalt, concrete, gravel, etc.), and an estimated friction coefficient.

[0083] The method 40 further includes a step 43 for training a neural network. By training the neural network, a relationship is established between the ambient air temperature measured by the vehicles (reference ambient sensor data) on the one hand and the road surface temperature measured by the distributed road sensors on the other hand.

[0084] By training the neural network using the ambient air temperature (measured by vehicles) and the reference road surface temperature, it is possible to specify the dependence of the road surface temperature on the measurable input variable, ambient air temperature. In other examples not shown, additional input variables can be used. In any case, the trained neural network can be used to output a road surface temperature estimate using a time series of current vehicle ambient sensor data as input.

[0085] In the example shown, the neural network can be trained on the reference environmental sensor data and the reference road sensor data (including the road surface temperature), so that the trained neural network can use (current) vehicle environmental sensor data as input to output a road surface temperature. Such a neural network thus encodes the dependence of the road surface temperature on environmental sensor data.

[0086] Particularly with reference data in the form of a time series, the neural network to be trained can be a recurrent neural network (i.e., one that has not only forward connections to downstream layers but also feedback connections to the same layer or upstream layers). This makes it possible to incorporate temporal dynamic behavior.

[0087] Thus, in the illustrated case of training a neural network with a time series of reference environmental sensor data, it can be used to output a road surface temperature estimate using a time series of current vehicle environmental sensor data as input.

[0088] As an alternative to training a neural network, other embodiments (not shown) may include applying a filtering method.

[0089] The method 40 further comprises a step 44 for generating a data structure encoding the trained neural network and a step 45 for collecting current vehicle environmental sensor data in the form of a time series of the current ambient air temperature.

[0090] The time series of the current ambient air temperature provides information about the current ambient conditions in the vicinity of one or more vehicles at multiple points in time. Since each vehicle is moving and only stays in a given geographical location for a limited period of time, it is preferable to measure the ambient air temperature in multiple vehicles at multiple points in time, so that the time series is created (or estimated) based on data from multiple vehicles at multiple points in time.

[0091] Such a time series of environmental conditions may also be measured for other examples of current environmental conditions, which further include solar radiation, GPS location, time, date, humidity, air pressure, wind speed, rain intensity, vehicle speed, fog, surface water amount, snow amount, presence of ice, traffic intensity (e.g., vehicles per unit time), wind direction, salt amount, surface type (such as asphalt, concrete, gravel, etc.), and / or an estimated friction coefficient.

[0092] Furthermore, the method 40 includes a step 46 for determining a road surface temperature estimate for the geographic location using the trained neural network and the collected current ambient air temperature data. In particular, the collected current ambient air temperature data can be used as input to the trained neural network. The relationship encoded by the neural network thus makes it possible to obtain a road surface temperature estimate for the geographic location as output.

[0093] Although this is not explicitly stated in Fig. As illustrated in Figure 4, the method 40 may further comprise retrieving reference weather data providing information about general weather conditions (e.g., humidity), which is similar to the retrieval according to steps 41, 42. Consequently, establishing the relationship is further based on the reference weather data (in addition to the reference environmental sensor data). Thus, the relationship is established using further weather data as (another) input.

[0094] In such cases, method 40 may further include receiving current weather data indicative of current general weather conditions at the geographic location, similar to the collection of step 45. Thus, step 46 of determining the road surface temperature estimate would be further based on the current weather data (in addition to the current vehicle environmental sensor data).

[0095] Fig. 5 schematically shows a situation with roads and vehicles in which methods, devices or systems according to the present disclosure may be used.

[0096] Fig. In particular, Figure 5 shows an apparatus 50 for estimating a road surface temperature of a geographic location. The apparatus 50 stores a data structure (not shown) that encodes the relationship between environmental sensor data measurable by vehicles and a road surface temperature.

[0097] A first vehicle 52 has passed a geographical location 51. The first vehicle 52 has collected current vehicle environment sensor data near the geographical location 51 and sent this data to the device 50.

[0098] Although the geographical location 51 is not provided with conventional means for determining a road surface temperature, the device 50 is configured to estimate the road surface temperature using the data structure and based on the transmitted vehicle environment sensor data of the geographical location 51.

[0099] In other examples (not shown), the device may base the estimation on a plurality of measurements from a plurality of vehicles, e.g., in the form of a time series of vehicle environmental sensor data.

[0100] Once the device 50 has estimated the road surface temperature, it may generate road condition data (e.g., indicating that the road is icy if the estimated road surface temperature indicates the presence of ice) and send this road condition data (and / or the estimated road surface temperature) to a second vehicle 53 that is currently approaching the geographic location 51.

Claims

[1] A method for estimating a road surface temperature of a geographical location (10,11), comprising: - retrieving (31) reference environmental sensor data providing information on environmental conditions in the vicinity of mobile vehicles (13, 14) in a data warehouse, - retrieving (32) reference road sensor data, which provide at least information about a road surface temperature in the vicinity of distributed road sensors (12), in the data warehouse, - establishing (33) a relationship between the reference environmental sensor data and the road surface temperature based on the reference environmental sensor data and the reference road sensor data, - generating (34) a data structure (22) encoding the established relationship, - collecting (35) current vehicle environment sensor data that provide information about current environmental conditions in the vicinity of one or more vehicles, in the one or more vehicles, - determining (36) a road surface temperature estimate for the geographic location using the data structure and the collected current vehicle environmental sensor data. [2] The method of claim 1, wherein the method further comprises: - Sending the collected current vehicle environment sensor data from the one or more vehicles to the data warehouse. [3] The method of claim 1 or 2, wherein the method further comprises: - Generating road condition data based on the determined road surface temperature estimate and - Sending the generated road condition data to at least one of the vehicles and / or other vehicles. [4] Method according to one of the preceding claims, wherein - the retrieved reference environmental sensor data comprise a first time series that provides information about environmental conditions at a first plurality of points in time, - the collected current vehicle environmental sensor data comprise a second time series that provides information about current environmental conditions at a second plurality of points in time. [5] Method according to one of the preceding claims, wherein the relationship is established (33) by: (a) training (43) a neural network, wherein the data structure is representative of the neural network, or (b) applying a filtering method, in particular a Kalman filter or a Bayesian filter or a non-statistical filter, wherein the data structure is representative of the output of the filter. [6] Method according to one of the preceding claims, wherein the reference environmental sensor data, the reference road sensor data and / or the current vehicle environmental sensor data are location-specific. [7] A method according to any one of the preceding claims, wherein the determining (36) does not involve the use of current road sensor data. [8] Method according to one of the preceding claims, wherein the reference environmental sensor data and / or the vehicle environmental sensor data and / or the reference road sensor data provide information about: - Ambient air temperature, solar radiation, GPS location, time, date, humidity, air pressure, wind speed, rain intensity, vehicle speed, fog, surface water quantity, snow quantity, presence of ice, traffic intensity, wind direction, salt quantity, surface type and / or an estimated coefficient of friction. [9] A method according to any one of the preceding claims, wherein the method further comprises: - Retrieving reference weather data that provides information about general weather conditions in the data warehouse and - Receiving current weather data that provides information about current general weather conditions at the geographical location, and - the establishment of the relationship is further based on the reference weather data and - determining the road surface temperature estimate is further based on current weather data. [10] A computer program product comprising program code adapted, when executed in a computing device, to perform the steps of a method according to any one of the preceding claims. [11] Apparatus (24, 50) for estimating a road surface temperature of a geographical location (10, 11, 51), comprising: - a data storage unit (25) configured to store a data structure encoding a relationship between reference environmental sensor data measurable by vehicles and a road surface temperature, - a communication interface (26) for receiving current vehicle environment sensor data providing information about current environmental conditions in the vicinity of the vehicle, - a processing unit (27), wherein the processing unit -- capable of communication with the vehicle sensor and the data storage unit and -- is configured to determine a road surface temperature estimate for the geographic location using the data structure and based on the collected current vehicle environmental sensor data. [12] Data warehouse unit (21) for generating a data structure (22) which is arranged to - to obtain reference environmental sensor data that provide information about environmental conditions near vehicles, - to obtain reference road sensor data which provide at least information on a road surface temperature prevailing in the vicinity of road sensors and measured by them, - to establish a relationship between the reference environmental sensor data and the road surface temperature based on the reference environmental sensor data and the reference road sensor data in the data warehouse, - to generate a data structure (22) encoding the relationship established and - send the data structure to the device. [13] A system for estimating a road surface temperature of a geographical location, comprising: - a device (24) according to claim 11 and - a data warehouse unit (21) according to claim 12, - wherein the data warehouse unit is communicatively coupled to the device to send the data structure of the data warehouse unit to the device. [14] System according to the preceding claim, further comprising: - a vehicle sensor (28) configured to collect current vehicle environmental sensor data providing information about current environmental conditions in the vicinity of the vehicle and communicatively coupled to the communication interface of the device. [15] Use of a system according to claim 13 or 14 for estimating a road surface temperature.

Citation Information

Patent Citations

  • Method for determining a coefficient of friction for contact between a tire of a vehicle and a road surface and method for controlling a vehicle function of a vehicle

    DE102017214053A1

  • Methods for determining road surface roughness

    DE102018008788A1

  • Method and device for detecting and evaluating road conditions and weather-related environmental influences

    DE102018203807A1

  • Method and device for predicting road-surface temperature or road-surface state

    JP2002048878A

  • JP002002048878A