Radio coverage map generation

1. A method of generating a radio coverage map, the method comprising:

obtaining image data of a geographical area, the image data comprising:

a representation of an environment in the geographical area;

an indication of one or more transmission point locations corresponding to locations of one or more transmission points in a wireless communications network; and

a direction of transmissions from the one or more transmission points; and

applying a generative model to the image data, to generate a radio coverage map of the geographical area, wherein the generative model comprises generative adversarial networks, and

wherein the generative model is trained on a set of training data comprising image data of geographical areas and corresponding radio coverage maps acquired through simulation based on one or more radio propagation models.

2. The method according to claim 1 , wherein the image data comprises a plurality of pixels and a plurality of layers, each layer comprising, for each of the plurality of pixels, respective values for one or more parameters.

3. The method according to claim 2 , wherein one or more first layers of the plurality of layers comprise the representation of the environment, and wherein one or more second layers of the plurality of layers comprise the indication of the one or more transmission point locations.

4. The method according to claim 3 , wherein the one or more first layers comprise respective predefined values for predefined types of object belonging to the environment in the geographical area.

5. The method according to claim 3 , wherein the one or more first layers comprise values for a height of an object belonging to the environment in the geographical area.

6. The method according to claim 3 , wherein the one or more second layers further comprise, for each of the one or more transmission points, an indication of one or more of:

a height of a transmission point;

a frequency of transmissions from the transmission point; and

a power of transmissions from the transmission point.

7. The method according to claim 1 , wherein the generative model is trained on a set of training data further comprising:

image data of the geographical areas and the corresponding radio coverage maps acquired through measurement.

8. The method according to claim 1 , wherein the generative adversarial networks comprise a generating network and a discriminating network, and wherein the generative adversarial networks are trained by:

generating, using the generating network, based on training image data of the geographical area, a fake radio coverage map;

determining, using the discriminating network, whether each of a plurality of training radio coverage maps is real or fake, the plurality of training radio coverage maps comprising the fake radio coverage map and at least one real radio coverage map; and

updating the generating network and the discriminating network based on whether the determination by the discriminating network is correct or incorrect.

9. The method according to claim 1 , wherein the radio coverage map comprises an indication of received signal strength over the geographical area.

10. The method according to claim 9 , wherein the indicated received signal strength at a particular location is the received signal strength associated with a transmission point having the highest received signal strength at the particular location.

11. The method according to claim 1 , wherein applying the generative model to the image data to generate the radio coverage map comprises:

for each of a plurality of transmission points in the wireless communications network, applying the generative model to image data comprising an indication of a single, first transmission point location of a plurality of transmission point locations corresponding to a location of a single, first transmission point of the plurality of transmission points to generate an intermediate radio coverage map in respect of the first transmission point; and

combining respective intermediate radio coverage maps to generate the radio coverage map.

12. The method according to claim 1 , wherein applying the generative model to the image data to generate the radio coverage map comprises:

applying the generative model to image data comprising an indication of a plurality of transmission point locations corresponding to the locations of a plurality of transmission points to generate the radio coverage map.

13. A radio coverage map generator, comprising:

an obtaining module configured to obtain image data of a geographical area, the image data comprising:

a representation of an environment in the geographical area;

an indication of one or more transmission point locations corresponding to locations of one or more transmission points in a wireless communications network; and

a direction of transmissions from the one or more transmission points; and

an applying module configured to apply a generative model to the image data, to generate a radio coverage map of the geographical area, wherein the generative model comprises generative adversarial networks, and

wherein the generative model is trained on a set of training data comprising image data of geographical areas and corresponding radio coverage maps acquired through simulation based on one or more radio propagation models.

14. The radio coverage map generator according to claim 13 , wherein the image data comprises a plurality of pixels and a plurality of layers, each layer comprising, for each of the plurality of pixels, respective values for one or more parameters.

15. The radio coverage map generator according to claim 14 , wherein one or more first layers of the plurality of layers comprise the representation of the environment, and wherein one or more second layers of the plurality of layers comprise the indication of the one or more transmission point locations.

16. The radio coverage map generator according to claim 13 , wherein the generative model is trained on a set of training data further comprising:

the image data of the geographical areas and the corresponding radio coverage maps acquired through measurement.

17. The radio coverage map generator according to claim 13 , wherein the generative adversarial networks comprise a generating network and a discriminating network, and wherein the generative adversarial networks are trained by:

generating, using the generating network, based on training image data of the geographical area, a fake radio coverage map;

determining, using the discriminating network, whether each of a plurality of training radio coverage maps is real or fake, the plurality of training radio coverage maps comprising the fake radio coverage map and at least one real radio coverage map; and

updating the generating network and the discriminating network based on whether the determination by the discriminating network is correct or incorrect.

18. The radio coverage map generator according to claim 13 , wherein the radio coverage map comprises an indication of received signal strength over the geographical area.

19. The radio coverage map generator according to claim 18 , wherein the indicated received signal strength at a particular location is the received signal strength associated with a transmission point having the highest received signal strength at the particular location.

20. The radio coverage map generator according to claim 13 , wherein the applying module is configured to apply the generative model to the image data to generate the radio coverage map by:

for each of a plurality of transmission points in the wireless communications network, applying the generative model to image data comprising an indication of a single, first transmission point location of a plurality of transmission point locations corresponding to a location of a single, first transmission point of the plurality of transmission points to generate an intermediate radio coverage map in respect of the first transmission point; and

combining respective intermediate radio coverage maps to generate the radio coverage map.