Connected Mode Mobility
How the Network Steers an Active Connection
In LTE and 5G NR, a device in the RRC_CONNECTED state has active radio bearers and an established context in the serving base station. Because that context cannot simply follow the device autonomously, the serving cell takes responsibility for mobility. It sends the UE a measurement configuration listing the carrier frequencies and neighbor cells to monitor, the filtering coefficients to apply, and the reporting events that should generate a measurement report. The UE measures reference signal received power and quality on the serving and neighbor cells, applies Layer 3 filtering, and reports back. The serving eNB or gNB then decides whether to keep the UE, change its measurement configuration, or hand it over.
A handover decision is followed by a preparation phase in which the source cell signals the target over the X2 or Xn interface, the target admits the bearers and allocates resources, and the source sends the UE an RRC reconfiguration message carrying the target cell identity and a dedicated random access preamble. The UE detaches from the source, performs random access on the target, and resumes data. During the brief gap between leaving the source and synchronizing to the target, user data cannot flow; minimizing this interruption is a central goal of every mobility enhancement, from RACH-less handover to dual-connectivity make-before-break.
The decision itself is governed by measurement events. The most common is the A3 event, where a neighbor becomes a configured offset stronger than the serving cell. Companion events include A2 (serving cell drops below a threshold, often used to start inter-frequency measurements) and A5 (serving cell below threshold 1 while a neighbor exceeds threshold 2). Tuning the offsets, hysteresis, and time-to-trigger trades off handover responsiveness against the risk of ping-pong, where a UE oscillates between two cells of similar strength.
Measurement Event and Handover Equations
Mn + Ofn + Ocn − Hys > Mp + Ofp + Ocp + Off
A3 event leaving condition:
Mn + Ofn + Ocn + Hys < Mp + Ofp + Ocp + Off
Layer 3 filtered measurement:
Fn = (1 − a) × Fn−1 + a × Mn, a = (1/2)(k/4)
Where Mn, Mp = measured RSRP of neighbor and serving cell (dBm); Of, Oc = frequency- and cell-specific offsets (dB); Hys = hysteresis (1 to 3 dB); Off = A3 offset (2 to 4 dB); the condition must hold for the time-to-trigger; k = filterCoefficient. Example: with Of and Oc both 0, Off = 3 dB and Hys = 1 dB, the neighbor at Mn = −82 dBm and serving at Mp = −88 dBm give Mn − Mp = 6 dB > Off + Hys = 4 dB, so the report fires after the TTT expires.
Connected Mode vs. Idle Mode Mobility
| Attribute | Connected Mode (Handover) | Idle Mode (Reselection) |
|---|---|---|
| RRC state | RRC_CONNECTED | RRC_IDLE / RRC_INACTIVE |
| Decision authority | Network (source cell) | UE (autonomous) |
| Trigger | A3 / A5 measurement report | Broadcast priorities & S-criteria |
| Active session | Preserved (bearers moved) | None to preserve |
| Typical latency | 30 to 50 ms (intra-freq) | Seconds (no urgency) |
| Signaling | X2/Xn + RRC reconfiguration | None until next TAU |
| Ping-pong control | Hysteresis + time-to-trigger | Reselection timer Treselection |
Frequently Asked Questions
What is the difference between connected mode and idle mode mobility?
In connected mode the network is in control: the UE has an active RRC connection and bearers, so the serving cell configures measurement events, then the source cell decides when to hand the UE over and signals the handover command. In idle mode the UE is in control, performing autonomous cell reselection on broadcast priorities, and the network only learns of the change at the next tracking area update. Connected mode preserves active sessions with tens-of-milliseconds interruption; idle reselection has no session to protect and can take seconds.
How does the A3 event trigger a handover in connected mode?
The A3 event fires when a neighbor becomes offset better than the serving cell: Mn + Ofn + Ocn − Hys > Mp + Ofp + Ocp + Off, with Hys around 1 to 3 dB and Off around 2 to 4 dB. The condition must hold for the time-to-trigger window (40 to 640 ms) before the UE sends a measurement report, after which the eNB or gNB issues the RRC handover command. Hysteresis and time-to-trigger together suppress reports from short fades and prevent ping-pong between similar-strength cells.
What causes handover interruption time and how is it minimized?
Interruption is the gap where the UE cannot pass data while it detaches from the source, achieves uplink sync with the target via random access, and resumes on the target. Baseline LTE intra-frequency handover runs roughly 30 to 50 ms; inter-frequency adds measurement gap overhead. 5G NR cuts this with RACH-less handover, make-before-break dual connectivity, and Layer 1/2 triggered mobility, approaching zero interruption by holding the source link until the target is ready. Conditional handover hides preparation latency by pre-configuring the target.