Figure 1: Number of NRA using given thresholds for GSM, UMTS and LTE

From a consumer perspective, the signal power received may be of less relevance that the probability of successfully connect to the service. Therefore, some NRAs are considering that a given area is covered when the probability of connecting to the service (e.g. the call success rate for voice service) is higher than a given threshold.

Further, 7 out of 33 NRAs use multi-level thresholds to define coverage. BEREC estimates that this practice gives more accurate information to the public. Considering this, BEREC recommends NRAs to choose a multi-level thresholding approach.

In order to improve the information on mobile coverage given to the public, there would seem to be merit in NRAs consider specifying at least three levels of mobile coverage (for example: basic / good / excellent). With regards to the publication of coverage maps this may enable end-users to come to well-founded decisions when choosing their MNO. It may also help identify areas where mobile coverage is available, but could still be improved, which would be of interest for MNOs as well as for the governments.

BEREC recognizes that mobile signal prediction enables the estimation of mobile coverage over the whole of the geographic surface of a given country. The NRAs should note that signal predictions are a statistical representation of the coverage achieved in practice.

NRAs should base coverage estimation (numerical data or maps) on coverage calculations/predictions, whenever it is not economically or technically possible to carry out field measurements of the whole country. Such predictions may be generated from modelling a number of relevant input parameters[6] to predict the level of available mobile signal power or service accessible in any given area.

An NRA may elect to:

1. generate coverage predictions and publish information themselves (maps and/or metrics about mobile coverage);

2. obtain the results of predictions from the operators and publish information themselves (maps and/or metrics about mobile coverage); or

3. use a third party to generate the coverage predictions and publish information (maps and/or metrics about mobile coverage).

In all cases, the NRA should be confident of the accuracy of the mobile signal predictions they use – see CP3.

The answers to the survey showed that 10 out of 33 NRAs use their own mobile signal prediction software which they combine with network data that is mainly provided by the mobile operators (e.g. base station locations, antenna parameters, frequencies). The software predicts mobile coverage, whose output can be referred to as theoretical mobile coverage, or mobile signal predictions.

These are mainly used to verify whether coverage obligations are being fulfilled by the operator. More generally, this approach allows the verification of national population or geographic area being covered by a pre-defined service parameter.

23 out of 33 NRAs surveyed do not perform calculations of mobile coverage by themselves. Here, the mobile operators provide the coverage mainly based on estimation/simulation tool. This theoretical mobile coverage which is sent to the NRA, can then be used in a similar way as above.

It is clear that in the majority of countries each operator or NRA has its own way of calculating mobile coverage. Whether the NRA performs the calculation itself or obtain the coverage information from the operators, it should ensure the accuracy of the information. This would help ensure comparability between the operators. See CP3.

It is worth noting that theoretical/predication-based mobile coverage information is the only known methodology that enables NRAs to derive an estimate of mobile coverage over 100% of their country land mass. Thus, theoretical/predication-based mobile coverage information should be used unless field measurements are feasible over 100% of the geography.

The constraints referred above concerning theoretical mobile coverage are related to the following:

- the digital terrain model which can have different resolutions and sources,

- the digital terrain clutter which can have different classes and sources,

- the software tool that can be from different providers,

- the propagation model (e.g. ITU.R P.1812, Okumura-hata, Crosswave model, ray tracing) than can be different and configured in different ways (e.g. the diffraction model + sub-path attenuation),

- the radiation pattern of the antennas,

-the sensitivity (received signal power) per service and respective link budget, and

- other planning parameters (e.g. penetration losses, body loss, height gain, coverage reliability, fading margins, etc.).

All these requirements have varying impact on theoretical mobile coverage information and may benefit from a more harmonised approach in the future.

A theoretical mobile coverage, by itself, should be understood as an estimation. To obtain the theoretical mobile coverage there is a need to specify the mobile coverage in terms of a particular target, which may be received signal power, percentage successful voice service or data service of certain downlink data rate. Variations exist amongst the NRAs whether coverage is estimated on a QoS or QoE basis.

NRAs should verify the reliability of mobile coverage information using, where appropriate, field measurements, noting that for technical and resourcing reasons it may not be possible to make widespread measurements[7]. Measurements by drive-testing offer an effective method of testing the accuracy of mobile signal predictions. NRAs should ensure statistical robustness of the measurement methodology and of the measurement processing and analysis.

The outcome of the survey[8] shows that NRAs can have similar reasons for using field measurements. Indeed, the NRAs carry out measurements[9] to:

- verify that MNOs comply with their license obligations on mobile coverage,

- safeguard the end users’ rights to have transparent information on mobile networks’ coverage and performance. In several countries, NRAs measure mobile coverage as part of resolving consumer complaints with respect to the mobile networks performance, and

- verify the reliability of the mobile coverage maps provided or/and published by the MNOs by collecting in-field measurements through drive testing across a representative sample of a given country where appropriate. This kind of measurements can act as incentive and strengthen the competition between operators because informing consumers helps them choosing the right service on their living area. Thus, publishing maps displaying mobile coverage of each of the MNOs in a market pushes them to constantly improve their networks.

The radio signal level and its quality do not guarantee that a mobile user can effectively access and use mobile services. Therefore, approximately half of the NRAs determine whether a location is covered, by measuring parameters that are more related to the QoS/QoE provided at that location. In few countries, NRAs aligned with MNOs on the exact set of parameters to be measured in the field, nonetheless these parameters differ across European states.

Drive testing is a standard method that should be used to monitor mobile coverage in the field – it may inform the level of mobile coverage (end/or service availability) along routes or at specific locations and may provide an opportunity to reflect the end users’ perception and experience of the mobile networks. Drive testing refers to the outdoor measurements, performed as static or in-move measurements. They typically include vehicle(s) equipped with at least a RF scanning equipment. Besides, this method ensures that all MNNOs are measured at the same time in a same location and, thus, gives a comparable picture of MNOs.

One NRA reported using crowdsourcing for providing some form of testing of coverage maps.

To realize the benefit of dissemination of mobile coverage information, NRAs should strive to provide easy-to-access accurate mobile coverage information to the widest possible range of consumers. The information published would be in particular useful for end-users to be able to compare the coverage provided by the operators and therefore to compare their services.

NRAs should consider a range of methods of access such as via their own and third-party website and apps, with the aim of maximization of coverage information accessibility by the widest range of consumers possible. Providing the information on mobile coverage in open data would be in particular beneficial as it can easily be widely reused by external parties. Third-party publication of NRAs’ coverage information should reference the source and whether the information has been tested for accuracy.

When NRAs publish coverage maps, they should consider the following:

1. Provide for consumer confidence in the maps by:

- stating if the data used to create the maps were provided by the MNOs or by the NRA itself.

- stating if the data used for the creation of the maps were tested for accuracy.

- describing the method used to test the data for accuracy (theoretical calculation and/or drive tests etc.), as well as the extent to which the data were tested for accuracy.

2. Increase the effectiveness of information conveyed by the maps by:

-providing consumers with the choice to select/unselect the available services and/or technology.

- displaying layers for different levels of coverage by exploiting the notion of multi-thresholding.

- providing a coverage map resolution of 100m or lower.

NRAs mobile coverage maps are also meant to complement and feed the EU Mapping Platform which collects data about mapping broadband coverage and quality of service (for fixed and mobile services). The platform is carried out in close cooperation with relevant Member State administrations (ministries) and with BEREC.