Method and device for predicting a connection quality to a cellular network

A device for predicting a value representative of a quality of connection of a vehicle to a cellular network from a target location, the device comprising a memory in which are stored instructions, a communication module and a computer suitable for implementing, when it is configured by the instructions:

• a phase of training a prediction model comprising the following steps:
• receiving, via the communication module, from at least one training vehicle connected to a particular access point, a location datum of the training vehicle, an identifier of the access point to which the training vehicle is connected, and a value representative of a quality of the connection of the training vehicle to said access point,
• determining, via a characteristic-extracting module, a characteristic of propagation of a radio signal between the received location and the identified access point, comprising at least one length of a rectilinear propagation path of the signal from the identified access point to the received location, the signal being transmitted through a first topographic element, and at least one characteristic of the first topographic element,
• training the prediction model on the basis of the determined characteristic of propagation associated with the value representative of a quality of the received connection;
• and
• a phase of predicting a second value representative of a quality of connection to a cellular network at the target location on a road network, comprising the following steps:
• identifying a target cellular access point accessible from the target location,
• determining, via the characteristic-extracting module, a second characteristic of propagation of a radio signal between the target location and the identified target access point, comprising at least one length of a rectilinear propagation path of the signal from the identified access point to the target location, the signal being transmitted through a first topographic element, and at least one characteristic of the first topographic element,
• predicting a value representative of a quality of connection to the target cellular access point from the target location by applying the prediction model to the second characteristic of propagation;
• and
• wherein a vector of characteristics that is determined for a given location and an identified access point comprises:
• a length of a reflected propagation path of a signal from the access point to the given location, the signal being reflected from a second topographic element, the length of the reflected propagation path being a sum of a distance from the access point to the second topographic element and a distance from the second topographic element to the given location, and
• an angle of reflection of the signal from the second topographic element.

A non-transitory data medium containing computer-program instructions that, when executed by a processor, are configured to implement operations for predicting a value representative of a quality of connection of a vehicle to a cellular network from a target location on a road network, the operations comprising:

• a phase of training a prediction model comprising the following steps:
• receiving, from at least one training vehicle connected to a particular access point, a location datum of the training vehicle, an identifier of the access point to which the training vehicle is connected, and a value representative of a quality of the connection of the training vehicle to said access point,
• determining a characteristic of propagation of a radio signal between the received location and the identified access point, comprising at least one length of a rectilinear propagation path of the signal from the identified access point to the received location, the signal being transmitted through a first topographic element, and at least one characteristic of the first topographic element,
• training the prediction model on the basis of the determined characteristic of propagation associated with the value representative of a quality of the received connection;
• and
• a phase of predicting a second value representative of a quality of connection to a cellular network at the target location on a road network, comprising the following steps:
• identifying a target cellular access point accessible from the target location,
• determining a second characteristic of propagation of a radio signal between the target location and the identified target access point, comprising at least one length of a rectilinear propagation path of the signal from the identified access point to the target location, the signal being transmitted through a first topographic element, and at least one characteristic of the first topographic element,
• predicting a value representative of a quality of connection to the target cellular access point from the target location by applying the model to the second characteristic of propagation;
• and
• wherein a vector of characteristics that is determined for a given location and an identified access point comprises:
• a length of a reflected propagation path of a signal from the access point to the given location, the signal being reflected from a second topographic element, the length of the reflected propagation path being a sum of a distance from the access point to the second topographic element and a distance from the second topographic element to the given location, and
• an angle of reflection of the signal from the second topographic element.

A road map comprising at least, for a location on the map, a value representative of a quality of connection to a cellular network, said value being obtained using a method comprising:

• a phase of training a prediction model comprising the following steps:
• receiving, from at least one training vehicle connected to a particular access point, a location datum of the training vehicle, an identifier of the access point to which the training vehicle is connected, and a value representative of a quality of the connection of the training vehicle to said access point,
• determining a characteristic of propagation of a radio signal between the received location and the identified access point, comprising at least one length of a rectilinear propagation path of the signal from the identified access point to the received location, the signal being transmitted through a first topographic element, and at least one characteristic of the first topographic element,
• training the prediction model on the basis of the determined characteristic of propagation associated with the value representative of a quality of the received connection;
• and
• a phase of predicting a second value representative of a quality of connection to a cellular network at the target location on a road network, comprising the following steps:
• identifying a target cellular access point accessible from the target location,
• determining a second characteristic of propagation of a radio signal between the target location and the identified target access point, comprising at least one length of a rectilinear propagation path of the signal from the identified access point to the target location, the signal being transmitted through a first topographic element, and at least one characteristic of the first topographic element,
• predicting a value representative of a quality of connection to the target cellular access point from the target location by applying the model to the second characteristic of propagation;
• and
• wherein a vector of characteristics that is determined for a given location and an identified access point comprises:
• a length of a reflected propagation path of a signal from the access point to the given location, the signal being reflected from a second topographic element, the length of the reflected propagation path being a sum of a distance from the access point to the second topographic element and a distance from the second topographic element to the given location, and
• an angle of reflection of the signal from the second topographic element.