Patent Application
20250109962
The disclosure relates to a method for operating a computer-based system in a vehicle, for determining an updated and optimized forecast relating to at least a weather or a road condition of a road section in a network of roads, said method comprising the following steps: providing a first set of data from a remote provider of information related to at least a weather or a road condition in at least said road section, said first set of data corresponding to said forecast; transmitting said first set of data to said vehicle; and providing a second set of data based on at least the operation of the vehicle or present conditions in the surroundings of the vehicle.
The disclosure also relates to a computer-based system for carrying out a method as mentioned above.
It is of great importance that today's infrastructure with networks of roads for different types of vehicles can be used with a high level of safety and efficiency. For this reason, it is important to provide updated and detailed knowledge regarding the status of such a network of roads, both for vehicle drivers and for road transport companies and authorities which are dependent on road transports.
For example, it is of importance for a vehicle driver to have knowledge of weather-related conditions such as the presence of snow or ice on a particular road section. Such information is crucial both as regards the safety when travelling on the road in question and also as regard the expected time it may take to travel a certain distance to reach a certain goal.
In particular, there is a need for drivers and owners of trucks, buses and other private companies and public authorities in the transport business to be quickly informed in the event that weather-related problems arise on certain roads and also how and when the roads can be expected to be available for travel. Furthermore, it is financially beneficial for trucking companies if accurate and updated information related to the status of road networks is available. Such companies can increase their efficiency by redirecting trucks to roads being in good driving conditions if, for example, the weather suddenly changes. Also, maintenance companies can increase their efficiency by prioritizing certain roads which may be in need of quick and efficient maintenance efforts.
For drivers of for example police cars, ambulances and fire trucks, information regarding the condition of certain roads or networks of roads is of high importance. With information describing for example that certain roads are very icy or snowy, the driver can decide whether an alternative road might be faster or safer, or both.
According to known technology, the status and condition of the road in a defined road network at a given point in time can be determined with the aid of knowledge of the current and expected weather, and also by means of other criteria which determines the condition and status of a particular road. Such criteria may for example include the occurrence of ice or water on the surface of the road, the expected friction between the tyres of a vehicle and the road surface, the air temperature and also the road temperature. Many other criteria may also be taken into account.
The condition of the road surface also depends on e.g. whether there has been any road service recently involving, for example, snow ploughs and similar equipment.
In order to obtain accurate information regarding the status and condition of a road network in a certain geographic location, information related to forecasts regarding weather conditions can generally be provided by means of institutions such as national weather agencies. Such institutions rely on weather stations which provide information about the weather in a certain region. However, such information can be unreliable. One source of uncertainty is the fact that such forecast information is often locally based. For roads located just a short distance from the weather station, the weather situation may be completely different. Also, changes in weather conditions or road conditions may occur very quickly. Normally, weather information from the above-mentioned institutions is updated every hour. During such a time period, the weather may change considerably in different road sections. This also contributes to the uncertainty of the forecasts from the institutions.
It should be noted that it is also previously known to provide enriched forecasts, which take into account additional parameters and also aggregated data from for instance fixed weather stations.
Also, information regarding, and detection of, other events such as activities related to road maintenance can be helpful to obtain an accurate forecast.
For the reasons stated above, there is a general requirement to obtain improved and optimized information related to weather and road status which is available from fixed weather stations and institutions offering such information.
The patent document U.S. Pat. No. 9,799,219 teaches an arrangement in which weather data from a weather communication centre is transmitted to a remote server and further to navigation units which are arranged in vehicles. Each vehicle which uses the arrangement also has a number of control units which are used to provide data which is related to local weather conditions. Said document also teaches that weather data can be modified in the server based on data from the vehicle.
Even though the system according to U.S. Pat. No. 9,799,219 constitutes a useful arrangement for providing weather-related data, there is a need for improved arrangements and methods within the technical field in question. Generally, there is a need for quick and correct adjustments of preliminary weather forecasts, in order to decrease the degree of uncertainty of such forecasts and to improve the weather and road information for different types of users of such information.
In accordance with the disclosure, there is provided a method and a system having a purpose of solving one or more of the drawbacks of known devices within this field.
For this reason, and in accordance with the disclosure, there is provided a method for operating a computer-based system in a vehicle, for determining an updated and optimized forecast relating to at least a weather or a road condition of a road section in a network of roads. The method comprises the following steps: providing a first set of data from a remote provider of information related to at least a weather or a road condition in at least said road section, said first set of data corresponding to said forecast; transmitting said first set of data to said vehicle; and providing a second set of data based on at least the operation of the vehicle or present conditions in the surroundings of the vehicle. In particular, the method furthermore comprises the following steps: combining, in said computer-based system, said first set of data with said second set of data for obtaining said updated and optimized forecast; and providing the updated and optimized forecast to a user.
Certain advantages are achieved by means of the dispenser system according to the disclosure. In particular, due to the fact that a computer-based system in each of the actual vehicles which are travelling on the network of roads can be arranged for providing said updated and optimized forecast, a very quick and highly accurate correction and update of a preliminary forecast in these vehicles can be made. Also, the fact that the update is carried out within the hardware of the vehicle itself allows the method to be carried out without the need for a continuous online connection with any external server system.
According to an embodiment, the first set of data may be provided in the form of a set of parameters (P) indicating road condition information being associated with said road section and being valid during a pre-determined period of time.
According to an embodiment, said first set of data may be provided in the form of at least one of the following parameters (P1-P7), which are provided as either values or probability distributions:
According to an embodiment, at least one of said parameters (P1-P7) may be associated with a predicted uncertainty related to a deviation of a parameter value or a probability related to a parameter, at a given point in time.
According to an embodiment, access to said parameters may be allowed via an online application programming interface (API).
According to an embodiment, said second set of data (10) may be provided by onboard system data obtained in said vehicle; said onboard system data corresponding to a road condition presently existing in the vicinity of said vehicle, an environmental condition in the surroundings of the vehicle, or a current condition of operation of the vehicle, or any combination of said conditions.
According to an embodiment, said onboard system data may be provided in the form of measurement data from at least one sensor system arranged in the vehicle.
According to an embodiment, said measurement data may be provided from at least one of the following sensors or signals being associated with the sensor system in the vehicle:
According to an embodiment, said first set of data may be combined with said second set of data in a fusion process which may involve a pre-processing stage or a fusion algorithm stage or a post-processing stage, or a combination of said stages.
According to an embodiment, an output set of data may be fed from said fusion process to at least one of the following destinations:
According to an embodiment, there may be provided a pre-processing stage which compensates for detected faults or ignorances in said computer-based system or related to said second set of data.
According to an embodiment, there may be provided a fusion algorithm stage comprising a prediction and updating process implemented by means of Kalman filtering and smoothing, or any other suitable similar algorithm.
According to an embodiment, there may be provided a post-processing stage utilizing a fused posterior distribution of output data.
According to an embodiment, there is further provided the steps of: providing a set of one or multiple probable forecasts, generally corresponding to the first set of data, as a first input in the form of a prior distribution to a fusion algorithm stage; wherein, in the case of the set of probable forecasts being constituted by one single probable forecast, that forecast is instantly set as the prior distribution; providing, as a second input to the fusion algorithm stage, observations from measurements by means of said sensor system in said vehicle; combining said first input and said second input into a posterior distribution; and wherein the combined forecast is gradually, for each update or combination, improved to better resemble the reality as based on said measurements, and compensate for any incorrections in the original forecast, in the fusion algorithm stage.
According to an embodiment, there is further provided the steps of: providing a post-processing step in which the posterior distribution is further processed; wherein, in the case of the set of probable forecasts being constituted by one single probable forecast, that forecast is instantly set as the prior distribution; wherein, in the case of the set of probable forecasts being constituted by multiple probable forecasts, the posterior distribution is matched to the most similar forecast, or a mix of the probable forecasts, based on said observations and inherit the rest of its properties, which have not been measured and thus updated.
According to an embodiment, there is further provided the steps of: providing said first set of data from said remote provider of information related to at least a weather or a road condition in upcoming road sections, said first set of data corresponding to a forecast; updating, in said computer-based system, said first set of data for upcoming road sections with information about at least said second set of data or the difference between said second set of data and the first set of data for the current road section where the second set of data is perceived, wherein the distance to the upcoming road section, and the difference in local road climatological factors as compared with the current road section, is taken into consideration for obtaining said updated and optimized forecast for the upcoming road sections.
The disclosure also relates to a computer-based system in a vehicle, for determining an updated and optimized forecast relating to at least a weather or a road condition of a road section in a network of roads. The system comprises: a vehicle communication unit which is configured for receiving a first set of data of information related to at least a weather or a road condition in at least said road section from a remove provider, said first set of data corresponding to said forecast; and a set of sensor devices providing a second set of data based on at least the operation of the vehicle or present conditions in the surroundings of the vehicle. Furthermore, said system is furthermore configured to combine said first set of data with said second set of data for obtaining said updated and optimized forecast, and for providing the updated and optimized forecast to a user.
Further advantages and advantageous features of the embodiments contemplated herein are disclosed in the following description and in the dependent claims.
The disclosure will be described in greater detail below with reference to the figures shown in the appended drawings.
Different aspects of the present disclosure will be described more fully hereinafter with reference to the enclosed drawings. The disclosure can be realized in many different forms and should not be construed as being limited to the embodiments below.
With initial reference to
In the context of this disclosure, the term “road section” is used to describe a selected section, distance or length of a road in which it is possible to prepare a forecast of for example a weather condition or a road surface condition. More precisely, a road section 3 can be located where a vehicle 2 is currently travelling, or it can alternatively be where a vehicle 2 is intended to be travelling at a future point in time, as shown in
A road section 3 may vary in length, suitably from a few meters up to 750 meters based for example on the property variation of the surface of the road section 3 in question.
Structures such as for example a bridge has unique properties from its connecting roads and will be isolated and may thus be regarded as a unique road section, whereas a strip such as a flat uniform highway most likely results in longer road sections.
According to a further example, the term “road section” can also be used to specify sections on virtually all the roads which are located within a given radius from the current position of a certain vehicle in the network 1 of roads.
As mentioned initially, vehicles 2 which are travelling on the network 1 of roads may obtain access to data related to weather and road status which is supplied via a remote provider of such information, as shown in a simplified form by means of reference numeral 4 in
As shown in
The remote provider 4 is arranged for transmitting data related to information corresponding to a forecast of weather conditions or road surface conditions, or both, via the communication unit 4c and the communication channel 5. Such information is here referred to as a “first set of data”, as shown with reference numeral 6 in
In addition to data related to a forecast of weather conditions and road surface conditions, the remote provider 4 may also transmit other related information, such as information on the formation of traffic congestion and queues.
Furthermore, each one of the vehicles 2 which are within the network 1 or roads and within the transmission range of the communication unit 4c is provided with its own vehicle communication unit 7, which is adapted for wireless communication. The vehicle communication unit 7, which is shown for only one of the vehicles 2 in
The forecast information from the remote provider 4 is suitably transmitted to one or more vehicles 2 as a result of the vehicle 2 sending a request for such information via the vehicle communication unit 7 to the remote provider 4.
Furthermore, each vehicle 2 as shown in
Furthermore, each vehicle 2 which is configured for operating in accordance with the disclosure, is provided with a computer-based control system 10 which is arranged in the vehicle 2 and which is configured for receiving and processing the first set of data 6 and the second said of data 9 in order to provide an updated and optimized forecast, i.e. as compared with the original forecast provided by the remote provider 4. The control system can furthermore be configured for transmitting a request directed to the remote provider 4 in order to obtain forecast data in that or those road sections 3 which may be relevant for the corresponding vehicle 2 to travel on.
In summary, the present disclosure relates to a method for operating a computer-based system 10 in a vehicle 2 for determining an updated and optimized forecast relating to at least a weather or a road condition of a road section 3 in a network 1 of roads. Generally, the method comprising the following steps: providing a first set of data 6 from a remote provider 4 of information related to at least a weather or a road condition in at least said road section 3, wherein the first set of data 6 corresponds to said forecast; transmitting the first set of data 6 to the vehicle 2; and providing a second set of data 9 based on at least the operation of the vehicle 2 or conditions in the surroundings of the vehicle 2. Furthermore, as generally described above, the method also comprises the following steps: combining, in said computer-based system 10, said first set of data 6 with said second set of data 9 for obtaining the updated and optimized forecast; and providing the updated and optimized forecast to a user.
The fact that the process of combining the first set of data 6 and the second set of data 9 is carried out in each one of the vehicles 2, which are configured for implementing the above-mentioned method, leads to a process of optimizing a forecast from the remote provider 4 which is implemented in the vehicle's 2 own hardware and software. This means that this optimizing process is independent from continuous communication with, for example, a centralized computer server having forecast information or other information regarding the road condition in the road sections which are of interest. More precisely, the process according to this disclosure is implemented partly by means of the centralized computer server 4a in order to obtain general forecast information. However, this is not needed during the process of updating and optimizing the forecast. This means that a further advantage is provided during the updating and optimizing process (which is implemented in the vehicle 2) since it is much faster than a corresponding process implemented by means of communication with a cloud server. This means that an estimated and optimized condition for a road section 3 can be determined before the vehicle 2 reaches said road section 3.
According to the embodiment shown in
As a non-exhaustive list of examples, the following parameters (P1-Pn) can be provided and transmitted by the remote provider 4:
The parameters (P1-Pn) can be provided either in the form of values or in the form of probability distributions. As an example, a first parameter (P1) can be provided in the form of a temperature value being expressed as an expected temperature T with a range within the temperature is expected to be, i.e. T1±x degrees centigrade. As a further example, a parameter such as “road condition class” mentioned above may be provided in the form of a set of probabilities, such as the following information: “probability for dry road=0%”;
“probability for wet road=10%”; “probability for snow on the road=70%”; “probability for ice on the road=20%”.
Consequently, each parameter is generally associated with a predicted uncertainty which is related to a deviation of the parameter value or a probability related to a parameter, and which is valid at a given point in time.
According to an embodiment, the first set of data 6 is in the form of at least one of said parameters (P1-P7) and is provided as a probability distribution for each one of those parameters (P1-P7) which are applied.
Furthermore, the parameters from the remote provider 4 are available over an online application programming interface (API). The coordinates in latitudes and longitudes for the position of interest can be sent by a vehicle 2 to the API which finds the nearest road section and returns all parameters associated with that road section. Consequently, a prediction of the current state of the road section as regards weather and road condition is provided by the remote provider 4.
Furthermore, and according to the embodiment shown in
For example, the sensor system 8 may comprise a laser sensor which can be used together with a camera and a computing unit in order to predict the road condition up to approximately 20 meters in front of the vehicle 2. Generally, the sensor system 8 could be configured to provide measurement data from a number of sensors or signals in the vehicle 2 so as to provide, for example, one or more of the following signals:
The measurement data which is provided by means of the sensor system 8 will be used to provide information of at least one of the following parameters which are related to the road section 3:
Preferably, the parameters originating from the sensor system 8 are accompanied with a predicted uncertainty describing what the deviation is for each parameter at a given time instance. The same principles regarding a predicted uncertainty which is applied to different parameters are used for the first set of data 6 and the second set of data 9.
According to an embodiment, the first set of data 6 are combined, or fused, with said second set of data 9 in a fusion process, which now will be described in greater detail with reference to
Suitably, the fusion system 12 forms part of the above-mentioned computer-based system in each vehicle 2 (see also
Consequently, the input section 13 will receive a preliminary forecast which will constitute one or multiple forecasts given different possible scenarios, each with a given probability.
The forecast/-s will constitute a prediction of the road condition prior to making any observations which are related to the vehicle 2 and the sensor system 8.
Also, data from the sensor system 8 is also fed to the input section 13, i.e. in the form of the above-mentioned second set of data 9. This means that the input section 13 is fed with observations made from a sensor system 8 providing information on the road weather conditions, or alternatively an environmental condition in the surroundings of the vehicle 2, or alternatively a current condition of operation of the vehicle 2, installed on the vehicle 2. The sensor observations will be used to update the prediction given by the forecast.
The input information to the fusion system 12 constitutes all valuable and fusible signals provided by forecast and the onboard sensor together with its corresponding uncertainties. Examples of signals are listed below:
Some information delivered by the forecast is constant, however still desirable to provide to a user. Such signals will not be processed by the fusion system 12, but simply passed through.
With reference to
As mentioned above, the first set of data 6 is combined with the second set of data 9 in a fusion process for obtaining an updated and optimized forecast. According to an embodiment, this process involves a pre-processing stage, a fusion algorithm stage and a post-processing stage. Generally, however, it should be noted that any combination of these stages can in principle be used. According to an embodiment which is shown in
The input to the pre-processing step 18 is received within the host vehicle domain through a receiver 21 adapted for the forecast API 22.
The sensor system 8 input is received through a connection within the software of the computer system 10 of the vehicle 2 (see
In the pre-processing step 18 we draw some conclusions on the sensor input 8 and may potentially apply prevention of observed misbehavior.
A part of the post-processing step 20 is to counter-act and prevent observed scenarios of misbehavior. In other words, this step compensates for any faults or ignorance in the fusion system 12 or in any of the sensor 8 inputs.
The main input to the fusion system 12 is the forecast from the remote provider 4. The forecast will constitute of one or multiple forecasts given different scenarios and provided with a corresponding probability, as described above.
There are multiple possible procedures for updating and optimizing one or multiple forecasts from the remote provider 4. An example procedure is that the set of probable forecasts are combined and provided as the prior distribution to the fusion step. In the case of the set of probable forecasts constituting of simply one forecast, that forecast would instantly be set as the prior distribution. The combined forecast is altered to more resemble the measurement in the fusion step. In the post-processing step, the posterior distribution would be matched to the most similar forecast (or potentially a mix of them) and inherit the rest of its properties, which has not been measured and thus updated.
The fusion algorithm process 19 may constitute a customized version of a classic sensor fusion algorithm, e.g., Kalman filtering, made suitable for the present disclosure. Classic filtering may be divided in two parts, being the prediction step, where we predict what road condition we will face resulting in a predicted density, followed by the update step, where we update our prediction with measurements which result in a posterior density being our output for each time step as well as the input to the prediction step for the next time step.
It should be noted that
To apply this general theory in the context of edge computed road condition great insights in climatology and empirical studies of comprehensive data are required. An initial approach is that a vehicle will include road condition information in the fusion calculations for a predefined range, e.g. 1 km, and predictions will be made separately for the road sections contained within that range. As soon as a new road section appears within the predefined 1 km range, that section will be initiated with the predicted forecast. When updating the road section, the distance to it as well as the climatological difference to the current section will be considered. Consequently, it can be learned, from the present road section, how the upcoming sections might differ from the forecast prior to facing them.
The prediction step (see
It should be noted that the term “nowcast” refers to the prediction of all road sections in the present point of time. This is as opposed to a “forecast”, which describes certain conditions before they actually occur.
The prediction step is shown in the equations below according to principles in which the prediction step equations are associated with a standard Kalman filter.
Furthermore, in an update step we update our prior distribution with observations made in real-time from the output of information from the sensor system 8 in the vehicle 2, such as from a laser, for instance. The observations made are associated with a measurement model matrix and an estimated measurement noise determining to what extent the measurement should be trusted, and how to interpret it.
Combining of the measurement model matrix, measurement noise and measurement in traditional Kalman filtering are done by calculating the product of the Kalman Gain, Kk, and the Innovation, Vk. Doing so, a correction term is obtained in order to update the prior mean. Merging the prior mean and correction term result in a posterior mean {circumflex over (x)}kjk, which may be seen in the below equations.
In order to achieve the posterior covariance, Pkjk, we subtract the prior covariance with the Kalman Gain, Kk, and the Innovation covariance, Sk, who reflect to what extent the measurement is informative.
The combination of posterior mean and covariance constitutes both the output from the fusion algorithm at each time step as well as the input to the prediction in the upcoming time step.
The components in the Kalman filter represent different information about the system. The Kalman Gain, Kk, determines how much we should trust the new information. The Innovation covariance, Sk, is the predicted covariance of the measurement. The Innovation, Vk, captures the new information in our measurement.
The update step equations are suitably associated with a standard Kalman filter.
Furthermore, in the post-processing step 20 (see
This is where we apply the posterior distribution from the fusion algorithm and the insights learnt from it to a well-balanced set of upcoming road segments. The difference in road segment specification will be taken into consideration while applying deep knowledge in road climatology when doing so. For small deviations in risk-free situations between the prediction and the update this will have little impact, however, to give insights into when the update is of greatest utility a few examples have been listed below.
The host vehicle travels along a highway in minus degrees air temperature, but where the road surface temperature is well above zero degrees Celsius. We have predicted the road to most likely be completely dry but start to make continuous detections of wet road. A scenario with non-dry road is therefore favored. This makes little difference for safety on the current section, but a nearby bridge with less heat capacity will likely be frozen under these conditions. Normally this would be no issue due to the road surface temperature being above zero degrees. However, we know that an upcoming road section contains a bridge which are known for being immensely affected by air temperature and wind likely reducing the road surface temperature in comparison to adjacent road sections. We would in this case be able to predict a substantial risk of slipperiness on the bridge, where there previously was none, and take precautions.
The host vehicle travels along a highway being predicted as filled with snow. We make detections of no snow as well as wet surface on a road of temperature below zero degrees Celsius. We are then able to draw the conclusion that a winter maintenance measure has been carried out and furthermore determine what measure it was. The same measure may then be applied for each road section up until the next highway exit or even further if we also know the standard winter maintenance route for that road.
Regarding signal update, the posterior distribution of our signals might be vastly different from one of the fusion inputs. We might find scenarios where, for instance, numerical signals are conflicting with the road condition of highest probability. If we know the numerical signals to be more certain than the road condition classification, then the road condition probabilities may be altered to the corresponding signals of high certainty.
Furthermore, a part of the post-processing step will be to counter-act and prevent observed scenarios of misbehavior. In other words, this step compensates for any faults or ignorance in the fusion system or in any of the sensor inputs.
The designated consumer or user of the process implemented by the fusion system 12 is an ADAS or AD application within the automotive industry, as indicated by means of reference numeral 23 in
Furthermore,
The output from the fusion system 12 will result in all the input parameters and its corresponding uncertainties if any, however refined and fused from the two or multiple sources. According to the disclosure, the fused result will be much more accurate, less uncertain and adapt to unexpected scenarios than previously known systems and methods. Some of the key deliverables are documented below.
The output will be consumed by both the host application 23 (see
Certain key deliverables will now be discussed.
Regarding the output of road friction, the system may face three different setups and will deliver the parameter accordingly. The three setups available are the following:
For the different setups our system may either apply different confidence to the estimates of the upcoming conditions or deliver three different signals corresponding to the current setup. Regarding validation, it is difficult to validate performance of the entire system due to a lack of ground truth reference, however the performance of fusion is on the contrary able to validate with the later observed data.
A signal describing the risk of reduced friction coefficient, potentially to multiple extents. It could either be a flag describing whether the vehicle may face reduced friction coefficient or not. It could also constitute of multiple flags describing whether the vehicle may face reduced friction coefficient below [1, 0.8, 0.6, 0.4, 0.2] for instance. It could furthermore constitute of a probability between 0 and 1 to whether the vehicle faces slippery conditions.
The road condition class parameter contains information regarding what road condition that the vehicle most likely faces. The different classes contained are the following:
The principles of a method according to the disclosure have now been described. It should be noted that the disclosure also relates to a computer-based system 10 in a vehicle 2, for determining an updated and optimized forecast relating to at least a weather or a road condition of a road section 3 in a network 1 of roads. The system 10 comprises: a vehicle communication unit 7 which is configured for receiving a first set of data 6 of information related to at least a weather or a road condition in at least said road section 3 from a remove provider 4, said first set of data 6 corresponding to said forecast; and a set 8 of sensor devices providing a second set of data 9 based on at least the operation of the vehicle 2 or present conditions in the surroundings of the vehicle 2. Furthermore, said system 10 is configured to combine said first set of data 6 with said second set of data 9 for obtaining said updated and optimized forecast, and for providing the updated and optimized forecast to a user.
According to an embodiment, the method according to the disclosure comprises steps as follows. Initially, it should be mentioned that a set of one or multiple probable forecasts (f1, f2, f3) is provided. These probable forecasts generally correspond to the first set of data 6 as mentioned above, and are used as a first input, in the form of a prior distribution, to a fusion algorithm stage 19. Furthermore, observations (m1) from measurements by means of the sensor system 8 in a vehicle 2 in question are provided as a second input to the fusion algorithm stage 19. The first input and the second input are then combined into a posterior distribution. Furthermore, the combined forecast is then gradually improved, for each update or combination (i.e each step of combining). This is carried out so that the combined forecast better resembles the reality, i.e., the current situation, as based on the measurements from the sensor system 8, which is ensured to have lower error rate than the forecast. This means that the previously made forecast is improved and updated based on information from measurements in vehicles which reflect a real, presently occurring, situation. For example, information regarding road or weather status or conditions which are sensed by the vehicle's set of sensors 8—and which are not taken into account in the original forecast, or prior distribution—can be used to update, improve and optimize the original forecast so that it more accurately reflects the reality. Also, the combined forecast will be used to compensate for any occurring incorrections in the original forecast. This is suitably implemented by means of the process in the fusion algorithm stage 19.
According to a further embodiment, the method according to the disclosure comprises steps as follows. A post-processing step 20 is carried out, in which the posterior distribution is further processed. In such a step, the posterior distribution will be matched to the most similar forecast, or a mix of the probable forecasts, based on said observations (m1) and inherit the rest of its properties, which have not been measured and thus updated. Climatological know-how will, furthermore, be applied to draw conclusions and further alter the updated forecast, i.e., if a salting activity can be concluded to have occurred, then that gives the possibility to alter parameters which have not been measured, e.g., the amount of residual salt on the road.
According to a further embodiment, the method according to the disclosure comprises steps as follows. There is provided a first set of data 6 from the remote provider 4 of information related to at least a weather or a road condition in upcoming road sections 3, wherein the first set of data 6 corresponds to a forecast. Furthermore, the first set of data 6 for upcoming road sections is updated, in the computer-based system 10, with information about at least said second set of data 9 or the difference between said second set of data 9 and the first set of data 6 for the current road section where the second set of data 9 is perceived. In such a step, the distance to the upcoming road section (i.e. the distance between the current road section and the road section for which a forecast is to be updated), and the difference in local road climatological factors, i.e. by comparing the road section for which a forecast is to be updated and the current road section in terms of e.g. altitude, nearby vegetation, road conductivity, geographical position, shadowing, etc., are taken into consideration for obtaining said updated and optimized forecast for the upcoming road sections.
The invention is not limited to the embodiments described above but can be varied within the scope of the appended claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| 22152737.7 | Jan 2022 | EP | regional |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/SE2023/050051 | 1/20/2023 | WO |
