Solar radio emission as disturbance of mobile radio networks
Radio measurements of cellular networks and solar noise
All cellular networks constantly monitor the radio frequencies allocated to the communication services offered to customers and constantly receive, from each user equipment (UE), such as smartphones, the signal level of the cells around it and the indicators of associated channel quality.
UEs periodically send a large number of measurements from Layer 2—MAC (3GPP TS 36.321)—and Layer 3—Radio Resource Control (3GPP TS 36.331) and, with MDT, UEs periodically share their measurements with the network (when positioning system (GPS) receiver is activated, UE measurements are also geolocated).
This standard mechanism creates billions of daily measurements describing the radio evolution around each cell, essential information for optimizing the use of the frequency band and combating the impact of radio interference and (inevitably) environmental electromagnetic noise such as than thermal noise, originating both from the Earth and from the sky.
Big Data analysis, a fairly recent but now consolidated industrial trend, allows new directions of investigation capable of exploiting massive radio data, including less frequent events, such as (but not only) disturbances of the solar activity.
The study of solar effects at 2695 MHz on 4G LTE systems by means of MDT paves the way for the study of effects for the new 5G radio access adopting the sub-millimeter band below 6 GHz and, in the near future, the band millimeter above 25 GHz.
The sun could be more disturbing at the higher 5G and 6G frequencies and in the meantime mobile radio services are extending their reach to include critical low latency uses (eg in contexts such as: autonomous driving, remote surgery, active networks energy and water, etc.).
RAN LTE installations are typically used in a three-sector configuration14 to optimize capacity and radio coverage, combining three sectors of 120° (eg half power at +/- 60°), the whole territory (360°) around. The cells are characterized by their physical position, by the frequency band allocated to communicate with the mobile terminals, by the horizontal orientation (azimuth, direction of maximum gain of the antenna) and by the vertical orientation, technically also called the inclination. Typically, BS cellular antennas use negative tilt (i.e. pointing towards streets and buildings where people live) or zero tilt (i.e. pointing towards the horizon) . Therefore, in general, the cells do not point skyward and the antenna gain decreases more steeply in the vertical direction (eg, half power at +/− 10° degrees). In practice, solar radio emissions could directly inject noise into radiomobile antennas, even when the cell azimuth matches the azimuth of the Sun, only at sunrise or sunset, when the tilt is somehow aligned with the solar elevation (zero) in the sky.
Nevertheless, we cannot forget that solar radio waves coming from the sky and impacting the ground (or roofs, buildings, roads, etc.) are subject to reflections and more generally to scattering. Indeed, considering the UHF wavelength typical of many mobile bands around the world, solar radio waves could also be bent by edge diffraction all along the Earth’s surface or by diffraction from human infrastructure on the surface. Earth (e.g. buildings, walls, cars). Electromagnetic scattering, the phenomenon that makes it possible to serve a huge mass of mobile terminals without the need for line-of-sight of the antenna, weakens the role of BS cell antenna tilt as protection against radio disturbances from the sun, allowing injection into the antennas of the BS cell as part of the solar radio disturbance when the tilt is not aligned with the elevation of the Sun. The only practical protection against solar radio disturbances remains the horizontal orientation of the antenna of the BS cell (azimuth of the cell).
To isolate the noise disturbances of the sun on the radiomobile RAN from other interferences, two criteria are taken into account. The first criterion takes advantage of the abundance of mobile radio cells insisting on a sufficiently large geographical area (see Fig. 1). Indeed, a single cell can be affected by local noise around the cell, while thousands of cells distributed over a wide area make it possible to attenuate local interference phenomena.
The second criterion exploits the abundance of different cell orientations (azimuth), always allowing (at each moment of the day) to select a specific subgroup of cells which, at that moment, result (horizontally) oriented towards the azimuth of the Sun and another subgroup of cells oriented in the opposite direction. Using this approach, the RF power of solar noise can be tracked throughout the day (see Fig. 2), selecting, at each instant, the appropriate subgroup of cells and analyzing the differences in behavior between the two subgroups. The sun-exposed subgroup is characterized by the fact that the maximum antenna gain is in the direction of the sun. In contrast, the other subgroup of cells exhibits the minimum antenna gain in the direction of the sun. Effects depending specifically on the orientation of the cells towards the Sun can emerge from this dynamic division of the cells in the two groups, by analyzing the radio measurements produced collectively by the single group of cells and comparing the two results.
In terms of frequency, the choice fell on the analysis of the 2.6 GHz band (2500–2690 MHz), which was allocated by the World Radiocommunication Conference (WRC) in 2000 for mobile communications services terrestrial. Analysis of the 2.6 GHz band has two advantages. The first solar radio emission is generally stronger in the 2.6 GHz band, compared to other lower frequencies currently used for mobile radio communications. Second, the 2.6 GHz band is currently the closest LTE band to the frequencies (in the 100 MHz range of 2.75 GHz to 2.85 GHz) that the Penticton Radio Observatory continuously monitors. The solar radio flux at 2.8 GHz is also called the F10.7 index. The index (10.7 being the wavelength in centimeters) has been an excellent indicator of solar activity since 1947, easily and reliably measured day-to-day (see Fig. 3) in all types of weather and well correlated with the number of sunspots15.16.
The Penticton F10.7 index, expressed in ufs, is made up of three daily measurements of solar flux, each representing a one-hour average. The time difference (9h) between Canada (Penticton) and Italy only guarantees the contemporaneity of the observation of the Sun (in summer) for the first daily Canadian measurement at 10h00 in Penticton (17h00 UT), time which corresponds to 7:00 p.m. :00 in Italy.
A general characterization of solar noise at 2.6 GHz can be seen in Fig. 4, showing the great difference existing, at these frequencies, between the conditions of calm Sun and active Sun.
To study the influence of the sun on mobile radio services, a relevant radio measurement is the uplink signal interference noise ratio (UL_SINR), i.e. the ratio (usually expressed in dB) between the power that a BS cell receives when a signal is transmitted (uplink direction) by a mobile terminal, and the interference and noise that the same cell detects when communicating with the mobile terminal on the same Resource Blocks ( RB being the smallest unit of radio resources that can be allocated to a user, 180 kHz wide).
By exploiting the UL_SINR measurements on the UpLink RBs allocated to 2.6 GHz, we have defined a different index, the “UL_SINR Index”, to describe the influence of solar activity on mobile radios. The UL_SINR index is constructed by taking the difference between the median of the UL_SINR distribution measurements linked to the cells which result directed towards the Sun azimuth (with a tolerance of ± 60°) and the median of the UL_SINR distribution measurements linked to the cells which result directed in the opposite direction (always with a tolerance of ± 60°). By averaging these differences between the medians of the UL_SINR distribution over one hour and changing the group of cells according to the course of the Sun throughout the day, we obtain a value, the UL_SINR index (expressed in dB ), which can be compared to the corresponding F10 .7 Index (of the same time). The idea behind this index is to maximize the detectability of solar noise against mobile phone interference, which is usually randomly distributed around the RBs assigned by the serving cell.
Considering that mobile radio cells are distributed throughout the territory in a way which tends to maximize radio coverage and mobile telephone traffic and at the same time to minimize reciprocal cellular interference, there is no reason to have a median UL_SINR distribution that depends on a specific cell orientation direction when we consider a high number of cells and measurements to fill these distributions. Therefore, the general expectation is that the UL_SINR index is close to zero. When a difference exists, and this difference is negative, it means that it operates a cause which makes the UL_SINR lower for cells oriented Sun and higher for cells oriented in the opposite direction. In other words, the anti-correlation between the Penticton F10.7 index and the UL_SINR index shows the existence of an influence of the Sun on the quality of the mobile radio uplink channel, the SINR being one of the indicators of good or bad communication on the radio (uplink).