This paper deals with the problem of triggering handoff procedure at an appropriate point of time to reduce the ping-pong effect problem in the long-term evolution advanced (LTE-A) network. In the meantime, we also have studied a dynamic handoff threshold scheme, named adaptive measurement report period and handoff threshold (AMPHT), based on the user equipment’s (UE’s) reference signal received quality (RSRQ) variation and the moving velocity of UE. AMPHT reduces the probability of unnecessarily premature handoff decision making and also avoids the problem of handoff failure due to too late handoff decision making when the moving velocity of UE is high. AMPHT is achieved by two critical parameters: (1) a dynamic RSRQ threshold for handoff making; (2) a dynamic interval of time for the UE’s RSRQ reporting. The performance of AMPHT is validated by comparing numerical experiments (MATLAB tool) with simulation results (the ns-3 LENA module). Our experiments show that AMPHT reduces the premature handoff probability by 34% at most in a low moving velocity and reduces the handoff failure probability by 25% in a high moving velocity. Additionally, AMPHT can reduce a large number of unnecessary handoff overheads and can be easily implemented because it uses the original control messages of 3GPP E-UTRA.

To determine a more accurate handoff triggering point of time is a critical point to greatly reduce the failure probability of handoff [

In the 3GPP Evolved Universal Terrestrial Radio Access (E-UTRA) standard [

Researchers were devoted to developing the queueing model for investigating the performance of handoff procedures in the last decades [

The aforementioned handoff drawbacks and problems motivate us to investigate a better handoff scheme taking a variable RSRQ-based handoff threshold and a variable report period related to RSRQ variation. We observed that the RSRQ variation is highly related to the moving velocity of UE and the environments the UE stays. That is, the RSRQ variation of UE is highly related to its moving velocity and diverse environments. Based on these, we propose a signal-aware and signal-variation-aware (velocity-aware) handoff scheme called

The rest of this paper is organized as follows. Section

In this section, we simply introduce the handoff mechanism of 3GPP E-UTRA [

Event list predefined in 3GPP E-UTRA standard.

Event | Triggering situation |
---|---|

A1 | RSRQ of serving cell becomes better than |

A2 | RSRQ of serving cell becomes worse than |

A3 | RSRQ of neighbor cell becomes offset better than SeNB |

A4 | RSRQ of neighbor cell becomes better than |

A5 | RSRQ of PCell becomes worse than |

A6 | RSRQ of Neighbor cell becomes offset better than TeNB |

Although the 3GPP E-UTRA provides the guideline of handling the time point of handoff triggering, the two RSRQ threshold values are fixed. These fixed threshold values are not feasible for implementing a dynamic handoff scheme, which may reflect the condition of the variation of RSRQ values (i.e., caused by moving velocity) and the environment the UEs stay. We investigate the relation between two consecutive

The system is modeled as

Figure

Diagram of distance versus

If the moving velocity of UE is moderate, denoted by

If the moving velocity of UE is fast, denoted by

If moving velocity of UE is slow, denoted by

In Cases

The SeNB has to keep monitoring

The handoff scheme must be able to recognize both

As shown in Figure

In this section, we present the proposed scheme, AMPHT, in a step-by-step manner. The flowchart of AMPHT is shown in Figure

Flowchart of AMPHT.

Initially,

The SeNB gets the latest

The SeNB passively receives the MRM sent by the UE when the event A2 occurs. In this case,

In the 3GPP E-UTRA standard, the SeNB can obtain the TrackingAreaCode, which is a set of predefined codes to represent each area in a cell, by the MRM from the UE. AMPHT uses the TrackingAreaCode to determine whether the UE approaches the boundary of the serving cell. Once the SeNB obtains the rough moving direction and

Based on

Here, we illustrate three cases for AMPHT design.

After the estimation of

When

After obtaining

To avoid the handoff failure, AMPHT uses (

In this section, we present the performance of AMPHT. The performance of AMPHT is validated by comparing the numerical experiments (MATLAB tool) with the simulation results (ns-3 LENA simulator). Related system parameters are given as follows. The serving cell is surrounded with 6 neighboring cells and the radius of each cell is 1000 m. The distance between two adjacent eNBs is 1800 m, which means there exists an overlapping area between any two adjacent cells. We assume an UE is always in radio resource control- (RRC-) connected mode; that is, AMPHT is always on when the UE is RRC-connected. Each numerical experiment and simulation result are obtained by averaging 10000 samples. In each sample, UEs are randomly distributed within the coverage of serving cell. The UE moves randomly in a fixed velocity until triggering handoff procedure and moving into the neighboring cell. In our analysis, the network response time of handoff procedure is 500 ms. The predefined handoff measuring period is 200 ms. All parameters are listed and shown in Table

Parameters of system scenario.

Parameter | Value |
---|---|

Number of eNB | 7 |

Power of eNB | 46 dBm |

Radius of eNB | 1000 m |

Distance between adjacent eNB | 1800 m |

Noise figure | 5 dB |

Network response time of handoff procedure | 500 ms |

Predefined handoff measurement period | 200 ms |

To simplify the numerical experiments, we set noise figure as 5 dB to represent additive white Gaussian noise (AWGN) and other possible interference in analysis scenario. We assume our analysis scenario is a free space and

After these parameters of system scenarios and the situations of handoff have been determined. Our numerical experiments use four levels of moving behavior, each velocity level of which is shown in Table

Analysis parameters of numerical experiments.

Parameter | Value |
---|---|

Walk velocity | 2–5 km/h |

Bicycle velocity | 10–20 km/h |

Vehicular velocity | 80–120 km/h |

High speed railway (HSR) velocity | 250–300 km/h |

Analysis times | 10000 times |

Also, we have compared the proposed AMPHT algorithm and the original 3GPP E-UTRA handoff algorithm by simulation. Comparing with numerical experiments, to obtain the results that are closer to the reality, we use ns-3 LENA simulator to simulate the metropolitan area. By using ns-3 LENA simulator, UE’s RSRQ value is affected by buildings and obstacles in the simulated area. Except for the accuracy of handoff triggering, we further compare the event-triggered handoff probability and the MRM transmission times by simulation results. The comparison of numerical experiments and simulation results are shown below.

The numerical experiments of premature handoff probability are shown in Figure

Effect of moving velocity on premature handoff probability by MATLAB tool and ns-3 LENA simulator.

MATLAB tool

ns-3 LENA simulator

Next, we examine the simulation results of premature handoff probability that is shown in Figure

Based on the analysis of premature handoff probability, AMPHT is more effective than standard handoff procedure to reduce the premature handoff probability. The premature handoff takes an important part in the cause of ping-pong effect.

The numerical experiments of handoff failure probability are shown in Figure

Effect of moving velocity on handoff failure probability by MATLAB tool and ns-3 LENA simulator.

MATLAB tool

ns-3 LENA simulator

Next, the simulation results of the handoff failure probability are shown in Figure

Based on the analysis of handoff failure probability, we show that the AMPHT is more effective than standard handoff procedure to reduce handoff failure probability. The handoff failure takes an important part in the cause of service interruptions.

Since our proposed AMPHT can adjust

Effect of moving velocity on event-triggered handoff probability by ns-3 LENA simulator.

When moving velocity of UE is low and the RSRQ value is close to handoff triggering point

Since our proposed AMPHT can adjust the period of UE’s RSRQ checking by RSRQ variation, we can reduce more unnecessary MCM and MRM transmission times to decrease the handoff overheads. Here, we compare the MRM transmission times between 3GPP E-UTRA handoff procedure and AMPHT by ns-3 LENA simulator. The simulation results of MRM transmission times are shown in Figure

Effect of moving velocity on MRM transmission times obtained by ns-3 LENA simulator.

When the moving velocity of UE is relatively high, the residential time of UE in a cell will be shorter. As a result, the times of MRM transmission will also decrease. When the moving velocity of UE is 10 km/hr, standard handoff procedure transmits MRM 3528 times to check the UE’s RSRQ value periodically. On the other hand, the adjustable period feature of AMPHT can reduce the MRM transmission times to 1453. It is worth noting that most of MRM are used for replying MCM to SeNB. In this way, AMPHT reduces about 4000 times to transmit unnecessary message in the moving velocity of UE of 10 km/hr.

In this paper, we have investigated an adaptive measurement report period and handoff threshold (AMPHT) handoff triggering scheme for SeNB. AMPHT periodically monitors the UE’s RSRQ when UE is moving. AMPHT also calculates the variance of RSRQ to adjust the period of transmitting MCM in advance. Numerical experiments and simulation results have shown that the premature handoff probability and the handoff failure probability can be decreased. Here, we sort out some situations that AMPHT are recommended to apply.

To achieve AMPHT, the SeNB has to maintain every UE’s RSRQ value at least three times to calculate the next MCM transmission period. The UE’s RSRQ value maintenance and next MRM transmission period prediction will increase handoff overheads of SeNB. To reduce SeNB handoff overheads, we strongly recommend that AMPHT should be applied when the moving velocity of UE is high, since service interruptions are more unacceptable compared with ping-pong effect.

By monitoring the RSRQ variation of UE periodically and estimating the moving velocity of UE to change the interval time to time, the SeNB is able to determine an accurate handoff triggering point of time smartly for reducing the MCM and MRM transmission times. To reduce the message exchange times between SeNB and UE, we strongly recommend that AMPHT should be applied when the moving velocity of UE is low, since low movement or static causes amounts of messages to monitor UE’s RSRQ for standard handoff procedure.

Based on the discussion listed above, system operator should apply the AMPHT by different situation. Finally, considering the current service application of UEs, AMPHT can be investigated further for supporting real-time QoS or QoE among eNB for moving UE.

The authors declare that there is no conflict of interests regarding the publication of this paper.

This work was supported in part by the Ministry of Science and Technology, Taiwan, under Contract MOST-103-2221-E-182-042.