L1/L2 Triggered Mobility – the new way of doing handover in 5G Advanced

The topic of mobility and handover has always been a primary focus of the standardization body 3GPP, through different generations of telecommunication standards. When a mobile terminal moves from the coverage area of one cell to another, handover needs to be performed. The main purpose of handover is to ensure that each terminal is always connected to the cell with the best signal quality. To keep services working, the handover needs to be done as quickly as possible and with the shortest possible interruption to the data transmission and reception. Ensuring that a terminal is always connected to a suitable cell is a tremendous task, involving parameter setting and tuning of different dependent functions in the terminal and the network. This becomes even more of a challenge when considering the large number of terminals, variations in the speed and direction of terminal movement, the unpredictable nature of the surrounding environment, diverse service requirements, and the need to maintain those services irrespective of the cell serving the terminal.

Thus, proper mobility management has been essential for every generation of cellular network technology.

Handover can lead to large signaling overhead and long connection interruption time, sometimes as high as 90 ms. A new way of doing handover with short connection interruption time was introduced in Release 18 of the 3GPP 5G Advanced in December 2023. It’s called L1/L2 Triggered Mobility (LTM).

But before discussing the principles and details of LTM, let’s look at how the baseline handover procedure works.

What is handover?

Mobility in connected mode – which is when a mobile terminal is active – is controlled by the network, assisted by the mobile terminal. The baseline handover procedure in cellular network standards, such as 4G LTE and 5G New Radio (NR) is illustrated in Figure 1. The baseline handover procedure uses layer 3 (L3) signaling, and to distinguish it from other mobility features, such as LTM, we here use the term L3 handover.

In 3GPP terminology, the mobile terminal is named User Equipment (UE). The base station the UE is connected to before the handover is referred to as the source gNodeB (gNB). The base station the UE is connected to after the handover is referred to as the target gNB. Before the handover, the UE is in the source cell, and the target cell is where the UE will be after handover.

Figure 1. Baseline layer 3 handover of a UE from a source cell to a target cell

While the UE uses the connection to the source gNB for user data transfer, it also performs measurements (such as quality or signal strength) on neighboring cells as well as on the source cell and transmits measurement reports (1) to the source gNB. The source gNB uses received measurement reports to decide whether to perform a handover of the UE to one of these neighbor cells. A typical example of when handover needs to be performed is when the measured quality or signal strength of a received signal of the source cell is below a certain threshold while, at the same time, a measured quality or signal strength of a received signal of a neighbor cell is above a certain threshold.

The source gNB selects one of the neighboring cells as the target cell for handover. It then sends a Handover request message (2) to the base station controlling this target cell to prepare a configuration containing all the information the UE needs to connect to the target cell.

This configuration is encapsulated into a signaling message, the handover command message that is encapsulated into a Handover Request Ack message (3) back to the source gNB.

The source gNB then transmits this handover command message (4) to the UE.

The UE switches to the target cell and transmits a message, commonly known as the “handover complete” message (5), in the target cell to the target gNB.

In the baseline L3 handover, the handover command message is sometimes large. This leads to signaling overhead and delays, due to necessary retransmissions as the UE moves away from the source gNB, which sometimes causes the handover to be triggered too late. This, in turn, may result in the radio link to the source gNB being lost before handover can be performed, and then the UE would perform a recovery procedure that causes a long service interruption.

Figure 2 illustrates the timeline:

Figure 2. Interruption during baseline L3 handover

When the UE has successfully received a handover command message, it immediately breaks the connection with the source cell and then initiates all necessary actions to establish the connection in the target cell, including UE reconfiguration, downlink synchronization, and uplink synchronization. Once the UE has established the connection, it resumes the user-plane transmission and reception in the target cell.

During the resulting handover interruption time, the UE cannot transmit or receive user data. In the baseline L3 handover, this time is typically between 50-90 ms. The exact number depends on factors like radio conditions and the UE implementation. This interruption normally does not impact commonly used services like telephony, video streaming, web surfing or video calls, as they have been designed to tolerate moderate packet loss and delays (also known as latency). But 50-90 ms is often too long for services with requirements on short latency, such as time critical communication (TCC) or ultra-reliable low latency communication (URLLC). Examples are Extended reality (XR) applications in cloud, remote control, or mobility automation.

Standardized features to reduce handover interruption time

Dual Active Protocol Stack (DAPS) handover was standardized during 3GPP Rel-16, with the aim to reduce the handover interruption time. The main principle of DAPS is that the UE keeps the connection with both the source node and target node active, until the overall handover procedure towards the target cell has been completed.

The requirement to keep two active connections simultaneously in DAPS has turned out to be quite challenging. Also, in practice, DAPS is limited to low bitrate services and cannot be used for the highest frequencies when beamforming is used. This led 3GPP to investigate alternative solutions to reduce the handover interruption time.

Inspired by beam management solutions for intra-cell mobility and the idea to preconfigure multiple target cell candidates before the actual handover decision is made, as with conditional handover, 3GPP has now developed and specified L1/L2 Triggered Mobility, where both measurements and handover signaling reuse paradigms from beam management procedures.

Introducing L1/L2 triggered mobility

The main idea with LTM is to enable mobility using lower-layer signaling to reduce overhead and interruption time. LTM supports the use of dual connectivity and carrier aggregation which makes it useful for services requiring high bitrate and short mobility interruption. LTM can also be used for the higher frequency bands above 7 GHz, together with beamforming. In a split radio access network architecture, the gNB is separated into a gNB Central Unit (gNB-CU) and one or more gNB Distributed Units (gNB-DU). LTM supports both intra-DU and inter-DU mobility without changing the gNB-CU. A a single gNB-DU may control many cells and a gNB-CU may be connected to many gNB-DUs. Therefore, in practice, LTM works for fairly large areas with many cells even with this intra-gNB-CU limitation.

The mobility procedure in LTM is performed in three steps, illustrated in Figure 3.

Figure 3. Overview of the procedure for L1/L2 Triggered Mobility (LTM)

Step 1 LTM preparation: The UE is prepared with LTM candidate cells. The UE measures and transmits L1 measurement reports for those cells.

Step 2 UL/DL pre- synchronization: An LTM candidate cell is identified as the target cell. While the UE keeps the connection with the source cell, the network triggers the UE to synchronize to that cell.

Step 3 LTM execution: The network triggers LTM cell switch towards the target cell and UE applies the new RRC configuration.  

LTM steps in detail

In Figure 4, we illustrate the different steps of LTM in more detail as a message sequence chart with the signaling between the UE and the gNB.

Figure 4. A message sequence chart of the procedure for configuring and executing LTM

LTM uses a mix of Radio Resource Control (RRC), Medium Access Control (MAC) and physical layer (L1) signals and messages. LTM is a network-controlled procedure, meaning that the decisions to configure and execute LTM are taken by the gNB. We also see that it is assisted by the UE, as the UE sends measurements reports to the network. Both these principles are also used in the baseline L3 handover. The use of L1 measurement reports for LTM are optional, and the network may instead use L3 measurement reports received by the UE to trigger DL/UL sync and/or LTM execution.

Figure 5 illustrates how the different signals and messages in LTM are used to reduce the time window when the user plane connection is interrupted. The example illustrates the case when the UE performs both downlink pre-synchronization and uplink pre-synchronization.

Figure 5. Interruption during L1/L2 Triggered Mobility

Compared to the baseline L3 handover, in LTM, most of the handover procedure steps are performed without breaking the connection with the source cell. Consequently, the interruption time is limited to the very last step, where reception of the cell switch command triggers the UE reconfiguration.

Downlink pre-synchronization (DL pre-sync) is triggered by the network. The source cell transmits a new type of Medium Access Control (MAC) Control Element (CE), named Candidate Cell TCI state activation, to the UE. The UE will add the indicated beam of the LTM candidate cell and start synchronizing to it in the downlink.

Uplink pre-synchronization (UL pre-sync), in this example, is triggered by the network in a similar way. The source cell transmits physical layer signaling, the physical downlink control channel (PDCCH) order, to the UE. The UE then transmits a random-access preamble in the LTM candidate cell indicated by the PDCCH order. When the network receives the random-access preamble, it calculates a timing advance (TA) value. The calculated TA value is kept by the network and will not be provided to the UE until the actual LTM execution.  

The network transmits the LTM cell switch command to the UE. The LTM cell switch command includes the TA value that was calculated by the network during the uplink pre-synchronization. As the UE received a TA value, it is aware of the exact timing for its uplink transmissions, and therefore, the UE can avoid random access. Now the UE breaks the connection and performs the remaining UE reconfiguration. Thereafter, the user data transmission and reception of the connection can resume in the target cell. Thanks to the downlink and uplink pre-synchronization,  the interruption is typically between 20-30 ms, which is much shorter than in baseline L3 handover.

Evolution of LTM

In Release 18 of the 3GPP 5G Advanced specifications issued December 2023, LTM is limited to mobility within a base station, gNB.

This initial version will be further enhanced in 3GPP Release 19. For example, 3GPP discussions have started to add support also for inter-gNB mobility using LTM and enhancements in the measurement reporting for LTM.

More reading

Everything you need to know about 5G Advanced