Adaptive method and device for prediction of a weather characteristic of a surface of a road segment
An adaptive method and device for predicting a weather-related characteristic of a surface of a segment of a road network. The method includes obtaining a location and of measuring a weather-related characteristic of the surface of the roadway of a road segment on which a measuring vehicle is traveling; predicting a weather-related surface characteristic of the road segment using a weather-observation history and a first prediction model associated with the road segment; associating a second prediction model with the road segment when a difference between the measured characteristic and the predicted characteristic is greater than a threshold; and transmitting to a vehicle a prediction made by applying the associated model to the weather-observation history.
1. An adaptive method for predicting a weather-related characteristic of a surface of a segment of a road network, the method comprising:
receiving, from a measuring vehicle,
a location of the measuring vehicle, and
a second weather-related characteristic of the surface of the road segment, which second characteristic is measured by a sensor of the measuring vehicle at said location;
determining a road segment on which the first vehicle is traveling on the basis of the received location;
selecting, from a plurality of prediction models, a first prediction model associated with the determined road segment;
predicting a first weather-related surface characteristic by applying the first prediction model to a weather-observation history obtained for the determined segment;
computing a first value E 1 representative of the difference between the first weather-related characteristic predicted for the road segment and the second weather-related characteristic measured by the measuring vehicle; and
when the computed first value E 1 is greater than a first threshold:
predicting a second weather-related characteristic by applying at least one second prediction model to the weather-observation history of the determined segment,
computing a second value E 2 representative of the difference between the weather-related characteristic measured by the measuring vehicle and the predicted second weather-related characteristic;
when the second value E 2 is less than the first value E 1 , associating the second prediction model with the road segment; and
transmitting, to at least one second vehicle traveling on the segment, a weather-related surface characteristic predicted by applying the prediction model associated with the segment to the weather-observation history of the segment.
2. The method as claimed in claim 1 , wherein when the first value E 1 is less than or equal to the second value E 2 , and comprised between the first threshold S 1 and a second threshold S 2 , the first prediction model is retrained using the weather-observation history of the segment and surface characteristics measured by the measuring vehicle.
3. The method as claimed in claim 2 , further comprising, when the first value E 1 is less than or equal to the second value E 2 and greater than the second threshold S 2 :
training a third prediction model using a plurality of weather-observation histories and a plurality of surface characteristics measured by at least one measuring vehicle, and
associating the third model with the road segment.
4. The method as claimed in claim 2 , wherein the first and second thresholds are defined depending on a value representative of the reliability of the measurement of the second weather-related surface characteristic by the measuring vehicle.
5. The method as claimed in claim 1 , wherein the weather-forecast history is obtained for a time window the duration of which varies as a function of a contextual datum.
6. A non-transitory processor-readable data medium on which is stored a computer program comprising instructions for executing a predicting method as claimed in claim 1 .
7. An adaptive device for predicting a weather-related characteristic of a surface of a segment of a road network, comprising a communication interface, a processor and a memory, the processor and the communication interface being configured by instructions stored in the memory and being configured to implement:
receiving, from a measuring vehicle,
a location of the measuring vehicle, and
a second weather-related characteristic of the surface of the road segment, which second characteristic is measured by a sensor of the measuring vehicle at said location;
determining a road segment on which the first vehicle is traveling on the basis of the received location;
selecting, from a plurality of prediction models, a first prediction model associated with the determined road segment;
predicting a first weather-related surface characteristic by applying the first prediction model to a weather-observation history obtained for the determined segment;
computing a first value E 1 representative of the difference between the first weather-related characteristic predicted for the road segment and the second weather-related characteristic measured by the measuring vehicle; and
when the computed first value E 1 is greater than a first threshold:
predicting a second weather-related characteristic by applying at least one second prediction model to the weather-observation history of the determined segment,
computing a second value E 2 representative of the difference between the weather-related characteristic measured by the measuring vehicle and the predicted second weather-related characteristic;
when the second value E 2 is less than the first value E 1 , associating the second prediction model with the road segment; and
transmitting, to at least one second vehicle traveling on the segment, a weather-related surface characteristic predicted by applying the prediction model associated with the segment to the weather-observation history of the segment.
8. A server comprising such a predicting device as claimed in claim 7 .