When Terrestrial RAN Meets NTN: A New Mobility Problem for NTN Handover

On Monday I wrote about handover basics and Event A3. That logic has served us well on the ground. Put a LEO satellite in the picture and a lot of it starts to break.

In a terrestrial cell, RSRP drops clearly as the UE moves from center to edge, so comparing levels makes sense. Seen from a satellite hundreds of kilometers up, the difference between center and edge can be just a few dB. Pure RSRP triggers lose most of their meaning.

Then there is movement. With LEO the cell itself can move over the user at several kilometers per second, so a phone sitting still on a table may still need a handover every few minutes. Add long propagation delay, Doppler, and timing advance that keeps changing, and mobility becomes a very different job.

3GPP has been responding release by release:

🔹 Release 17 brought NR NTN together with conditional handover triggers based on time and on UE distance to the cell reference point, beyond signal level alone. 🔹 Release 18 went further with NTN to TN mobility improvements and RACH-less handover, which cuts signaling when many users move at once.

That last point worries me most for real deployments. When a satellite beam leaves an area, it can take a whole population of UEs with it at the same time, and the terrestrial layer has to absorb that load gracefully.

For those of us coming from SON, this is a familiar challenge in a new geometry. Neighbor relations, mobility policies and load balancing will need to understand ephemeris and orbit timing, not only radio measurements.

(For how SON itself has to adapt when the network moves, see NR Over Non-Terrestrial Networks: Can SON Work When The Network Is Moving?.)

𝗧𝗵𝗲 𝗻𝗲𝘅𝘁 𝗺𝗼𝗯𝗶𝗹𝗶𝘁𝘆 𝗲𝘅𝗽𝗲𝗿𝘁𝘀 𝘄𝗶𝗹𝗹 𝘁𝗵𝗶𝗻𝗸 𝗶𝗻 𝗼𝗿𝗯𝗶𝘁𝘀.

#NTN #5G #NonTerrestrialNetworks #LEO #Handover #3GPP #RAN #SON