Cell hand-off behavior


I started a thread last week,


about the difference between the cell location and cell tower location. The OpenCellId data includes lat/lon estimates of the center of the cell. I’m more interested in the location of the cell’s transmitter (the location of the cell tower). I tried to encourage a conversation about how OpenCellId data might be re-analyzed to estimate cell tower locations (in addition to the current cell centers).

There doesn’t appear to be any interest in that matter. @Sagar responded cordially, but no other discernable interest, so I 'll let it go.

Maybe there is more interest in another topic I’ve been trying to understand: hand-off from cell to cell.

Theoretically, a stationary cellular modem will scan for a sufficiently strong signal from a Mobile Network Operator (MNO), register, and camp there. If the modem is using a multi-SIM, capable of accessing multiple MNOs, at a given location there may be signals available from multiple MNOs, and for any one of those MNO there may be strong enough signals coming from multiple towers, so the modem may have many choices regarding cell registration, and must pick one. Once registered, one might think that stationary modem would remain registered on that cell indefinitely. but that’s not the case; more in a minute.

For a mobile cellular modem, once registered, the modem can only remain registered with a cell while it is within the cell. As the modem reaches the cell boundary and the signal grows weak, the modem must be handed-off to the equipment of another cell with a stronger signal. If the hand-off occurs before the signal from the first cell grows so week it drops then there is no loss in connectivity. If no hand-off occurs, before the tower and modem can no longer communicate then the connection drops and the modem in a cellular dead zone (from its perspective).

So I thought that if a cellular modem is stationary, it would remain registered with a single cell for an extended period. I was, however, surprised by my own observations, checking cell registration every 5 seconds, that the modem may be handed back and forth among two of more towers over and over.

You might also think (as I did) that when heading in a straight line, say on a north-south road, the modem would be handed off every kilometer or two to a new tower. But that’s not what happens either. The registration bounces around between various sets of towers which change as the modem continues its journey. What’s most surprising is how frequently (while heading north, for example) hand-off will occur to a cell it was registered on when it was surprisingly farther south. Also, I had expected to see hand-offs occurring at longer distance intervals than they actually do. Initially, I was checking for hand-offs every 30 seconds, but hand-off were so much more frequent than I expected that I thought I might be missing some (more than one hand-off in 30 seconds), so I started checking registration every 10 seconds, and then every 5.

Hand-off algorithms are a topic of research and there are many papers on the subject. The selection and implementation of hand-off algorithms is not uniform, so hand-off behavior can very from one MNO to another, and even tower by tower.

I’ll drag anyone any further into the weeds unless there is interest in this topic.

If you are interested in this topic then, what the hell, why not reply.

If there is no interest in this topic, and no further interest (save the one reply) to my previous topic, then I’ll than the moderators for creating this forum, conclude my interests do not coincide with this community and search for another.