As an addendum regarding the ability of EM radiation to pass through objects and tendency to be line of sight at gsm frequencies. First of all , let us think about wavelength ( or its inverse frequency)All EM radiation in freespace away from any objects , by that I mean many wavelengths distance, propagatesnaturally in a straight line away from its source. This is true for all frequencies, ie highest frequencies are no more straight than lower frequencies. However, the direction of propagation can be modied by the environment (or medium) in which the EM radiation is travelling. We are now no longer talking about "free space". Here on earth we have objects and mediums which can be closer in distance to the source than one wavelength. Typically for what we call "Long wave" radiation less than say 500 Hz (600 metres) very long distances can be achieved - well beyond the horizon as the EM radiation tends to be bent by the curvature of the earth. This is to do with the conductivity of the earth ( ground) surface) as it causes the wavefront to slow slightly a the point it penetrates the ground ( induced current voltage in that medium). The proper term is diffraction - the same principle as a beam of light passing through a prism. This is called the ground wave mode of propagation . At much lower frequencies where the wavelength is say 60 miles the EM radiation can propagate easily right around the world, this is because the Earth and ionosphere create a wave guide - similar to the hollow tubes that used for very high frequencies - typically above a Ghz. This is called ionospheric propagation and is widely used to communicate with submarines for obvious reasons There is also another ionospheric mode which allows "short and medium " wavelength long distance propagation, by multiple reflections of the wavefront between earth and the ionoshere. These "non straight" line modes exist purely because of the presence and characteristics of a medium or mediums and their relationship to the wavelength of the EM radiation in question. In the case of a GSM signal - you are correct - these are pretty much straight line line of sight propagations, except where reflections occur. This is because the wavelength is short enough such that the diffraction effects from the ground and atmoshere are much less significant at these frequecies. The EM energy absorbed by objects is also related to their size / dimensions in terms of the wavelength. Metal lengths and loops of steel framed buildings absorb or reflect more energy the closer they are to the wavelength or sub wavelengths - particularly maximising at 1/2 and 1/4 wavelength. The object's conductivity is also a factor as it readily absorbs energy . You can see why gsm and microwave signals are attenuated more by trees , leaves fog mosture and cannot penetrate so well, concrete buildings with metal rebar mesh etc and how such structures tend to reflect rather than absorb , due to having sub wavelength resonances. Distances are maintained by the use of fucused arrays are explianed before.