Optimizing the balance between different handover parameters for network selection is one of the challenging tasks for seamless communication in heterogeneous networks. Traditional approaches for network selection are mostly based on the Receive Signal Strength (RSS) from the Point of Attachment (PoA) of a network. Most of these schemes are suffered from high handover delay, false handover indications, and inappropriate network selection for handover. To address these problems, we present an optimized network selection scheme based on the speed of a mobile node. A mechanism based on two different thresholds on the speed of a Mobile Node (MN) is integrated in the proposed scheme. If the speed of an MN is greater than any of the threshold, the MN performs handover to a particular network. We employ Grey Relational Analysis (GRA) in the proposed scheme to select the best PoA of the selected network. Similarly, to deal with false handover indications, we proposed an optimized handover triggering technique. We compare our proposed scheme with existing schemes in context of energy consumption for scanning, frequent and failed handovers, packet loss ratio, and handover delay. The proposed scheme shows superior performance and it outperforms existing schemes used for similar purpose. Moreover, simulation results show the accuracy and performance of the proposed scheme.
The aim of 4G networks is to provide generic connectivity among heterogeneous networks. With the advancement of new networks such as Long Term Evaluation (LTE), WIFI ac draft and WiMAX rel 2 provide high data rate and better connectivity to the end users. To access multiple networks an MN must be equipped with multiple interfaces. To ensure continuous connection among heterogeneous networks an MN must perform seamless switching from an Access Point (AP) or Base Station (BS) of a network to another AP or BS. This seamless transfer from one network to another can be possible if either an MN is already registered with all of the networks or there is a central system responsible for the registration of an MN. A seamless transfer of an ongoing session from one network to another network is called Vertical Handover (VHO). The VHO enables an MN to move inside heterogeneous networks and perform handover to any network regardless of the breaking of connection.
There are a number of parameters affecting a network selection process during VHO; these are velocity of an MN, RSS from the current PoA, energy required for scanning different networks through different interfaces, network connection time, and so forth. A handover process can be categorized in three stages,
The IEEE 802.21 Media Independent Handover (MIH) provides seamless mobility between all families of IEEE technologies and 3GPP [
In last couple of years, various network selection schemes have been proposed. Most of the schemes do not consider the current context and preferences of a user. Similarly, these schemes perform handover on the basis of single parameter. Considering more than one parameter for network selection leads to two-dimensional cost functions. The first dimension enables a user to request different services from a network and second dimension represents the total cost of the network against the requested services. The cost based network selection can be further categorized in three different parameters: weighting parameters, QoS parameters, and network priority parameters. The context of all of these three types of cost is different and it varies according to network situation and availability. A network selection method can be based on one of the types from these parameters. In last decade, researchers introduce a number of schemes based on these parameters [
To deal with the aforesaid constraints, a network selection scheme can be based on the requirements of an MN. These requirements consist of communication cost, data rate, QoS, and so forth. In last decade, various schemes proposed optimized network selection schemes based on user preferences [
Parameters used for network selection.
In order to optimize the working of a network selection process, we proposed a model based on the speed of an MN. The proposed scheme efficiently selects an appropriate network for handover while considering MN’s current speed. Moreover, the proposed scheme adopts GRA decision mechanism for the selection of best PoA of the selected network. The scheme significantly reduces the number of false handover indications and packet loss ratio. Similarly, the proposed scheme successfully reduces the frequent handovers problems present in recent literature. Moreover, we also introduce a handover triggering mechanism which efficiently reduces the number of false handover indications.
The rest of the paper is structured as follows. Section
In this section, we first present detailed study of handover triggering techniques. Then, we discuss network selection models that support our assumptions, focusing on showing the differences from our approach.
Mobility robustness optimization provides an MN with the support to detect and correct three types of triggering issues, that is, too early, too late, and to a wrong cell. Researchers proposed various techniques to enhance the working of handover triggering, avoid these three types of issues, and reduce the false handover indications. If a handover is triggered too early it uses the network resources in an inefficient way and an MN does not succeed to connect to the target network. Similarly, in case of too late handover the MN moves far away from the current network and hence it disconnects from the current network during handover to a target network. All of these three types of issues are explained in Figure
Too early, too late, and wrong cell handover.
There are other techniques which periodically scan the available networks. If the RSS of available networks becomes greater than current network the MN triggered the handover [
Recently, various techniques have been proposed for network selection during handover in heterogeneous wireless networks. Most of the schemes are based on the optimization of different parameters necessary for handover. The optimization of these parameters reduces the handover time and latency. With the passage of time the numbers of new access networks are increased rapidly and thus produced signaling overhead and other issues related to a handover phenomenon. Similarly, the new access technologies such as LTE-Advanced and Bluetooth 4.0 low energy were introduced to save communication time and energy. All of the recent technologies try to provide their customer with the best QoS. The QoS of a network can be enhanced if a customer is provided with a continuous connection among different networks.
In order to assure required QoS for various applications running by MN and to avoid frequent handover in heterogeneous networks, an Analytic Hierarchy Process (AHP) method for network selection is introduced in [
A scheme based on the optimization of MIH standard has been proposed in [
The energy consumption during selection of networks is an important parameter and it needs to be considered carefully. The energy consumption by an MN is directly depending on the scanning of available networks during a handover process. Different schemes based on the energy efficient network selection for multimedia based applications have been proposed in [
In last decade, researchers work hard to design energy efficient scanning techniques. Traditional techniques were mainly based on scanning through all interfaces. With the passage of time two different techniques become famous for scanning of available networks, that is, periodic and adaptive scanning. These techniques are optimized by different researchers and most of the scanning techniques were based on them. An energy efficient adaptive scanning algorithm is proposed in [
All of the above network selection techniques are based on different parameters. These schemes are good in a particular direction but not good for a generic network selection scheme in heterogeneous wireless networks. Next generation networks are growing rapidly and hence network selection on particular parameter and objective will not be enough for future use. Therefore, an MN should be provided with an appropriate network for handover in heterogeneous wireless networks.
The proposed model operates in two stages where in stage one a handover triggering mechanism is developed to reduce false handover indications and in stage two network selection scheme is used to provide an MN with appropriate network and PoA.
A handover triggering mechanism facilitates MN to initiate handover process when it is required. The proposed model adopts threshold based mechanism for handover initiation session. In other words, handover is initiated if RSS from the current network drops below a predefined threshold. The advantages of the threshold based triggering mechanism are reducing the number of false handover indications and hence reducing total number of handover failures to a network with overloaded APs/BSs. The time
Diameter is equal to the double of the radius
The radius of the coverage area of AP/BS is not exactly circular and thus the propagation of signals at one direction is different from another direction. Thus to avoid changes of RSS in one direction from another we use an index
The total handover time must be less than the time required by MN to remain connected to an AP/BS. Therefore, the handover triggering point must be set as a distance equal to
We put a threshold of RSS level on the boundary of coverage area of a PoA of a network. In our scheme, to avoid false handover indications, the threshold
When MN is moving at a high speed in a coverage area of WIFI network, it requires frequent switching from one AP to another since the radio coverage of WIFI AP is small. This frequent switching leads to much energy consumption, high packet loss, and breaking of connection. To deal with such situation, an optimized network selection scheme is proposed for handover based on the speed of MN. We assume that MN moving in heterogeneous networks has three different interfaces: WIFI, WiMAX, and Cellular. When the RSS from the current AP/BS drops below a particular threshold, it will initiate handover by selecting a network on the basis of its speed. If the speed is slow, the MN will scan the available networks through WIFI interface and the rest of the interfaces will be in standby mode. Similarly, if the speed is medium or high, the MN will select Cellular or WiMAX networks, respectively. To support energy efficient handover in this way, we implement a monitoring index based on the speed of MN using
We use two thresholds for network selection based on the value of
As shown in Figure
Selection of available networks based on MN’s speed.
Once MN decides a network for handover based on its speed then it is important to select best PoA of the selected network. There are different decision functions available for this purpose like SAW, WPM, MEW, AHP, GRA, and TOPSIS. AHP and GRA are the well-known mechanisms used for selecting a PoA on the basis of different parameters. We used GRA in our scheme to select the best available PoA. GRA analyzes the relationship rank between discrete sequences. We take one of the sequences as a user defined sequence. A user will first obtain the highest values of each objective and combine them in the user defined sequence. After calculating the user defined sequence, MN will compute the Grey Relational Coefficient (GRC) of each sequence and user defined sequence. MN compares the GRC value of user defined sequence and another comparative sequence of available PoAs. The PoA with highest GRC value is selected for handover. To clearly demonstrate the working of GRA used in proposed scheme, we use
A sequence is more preferable if its GRC value is greater. We computed GRC for every sequence using
Once the GRC of each sequence is calculated, the MN then selects the PoA having maximum GRC for handover. The working of the GRA in proposed scheme is illustrated in Figure
Working of GRA in proposed scheme.
The working of the best PoA selection among available PoAs is illustrated in Figure
Selection of best PoA.
The proposed scheme is compared with existing schemes used for reducing false handover indications and network selection. Extensive simulations were performed to test the accuracy and performance of the proposed scheme. The proposed scheme is implemented in C++ language. The random waypoint mobility model and random movement trajectory are adopted for MN’s movement across heterogeneous networks. We assumed the movement of an MN across three different networks, that is, WIFI, WiMAX, and cellular networks. The network selection is tested for different speeds of MN. The MN selects a network during handover on the basis of the speed. The values of two thresholds
Simulation parameters.
Parameter | WIFI | Cellular | WiMAX |
---|---|---|---|
Cell radius (m) | 100 | 500 | 500 |
Frequency (Hz) |
|
|
|
Path loss exponent |
4.0 | 3.0 | 3.5 |
Transmission power (dBm) | 15 | 27 | 25 |
Threshold ( |
−60 | −64 | −64 |
Speed limit ( |
>0 & ≤3 | ≥4 & ≤6 | >6 & ≤10 |
The proposed scheme requires less energy compared to existing schemes [
Figure
Energy consumption.
The proposed scheme selects a network on the basis of the speed of an MN. If the speed of MN is higher than
Frequency of handovers.
The optimal handover triggering scheme efficiently reduces the number of failed handovers. The proposed handover triggering scheme is not affected by too early and too late handover issues. The proposed scheme triggered handover on exact time and hence the MN successfully selects and handover to an appropriate network. The existing scheme is highly affected by too early and too late handover issues because of the poor triggering scheme [
The probability of blocking of new connections on an AP/BS is given by
The comparison of failed handovers in proposed scheme and existing scheme [
Number of failed handovers.
The proposed scheme is compared with the existing scheme in context of packet loss [
Packet loss ratio.
The existing scheme requires higher handover delay because of the selection of inappropriate network for handover [
Handover delay at different speeds.
In this research work, we proposed an optimized network selection scheme based on the speed of an MN in heterogeneous wireless networks. We proposed two thresholds on the speed of MN. The MN checks its speed against the predefined threshold and according to its current speed it selects the appropriate network. A WIFI network cannot be used for fast MN’s movement, because of its smaller coverage area and frequent handover problem. Moreover, the MN only scans a particular network using a single interface. Thus the energy consumption during scanning through all interfaces is significantly reduced. The proposed scheme is compared with the periodic and adaptive scanning in context of energy consumption. The proposed scheme outperforms periodic and adaptive scanning techniques in consumption of energy. We integrate the functionality of GRA in our scheme to select the best PoA for better handover performance. Moreover, the proposed handover triggering mechanism significantly minimized the number of false handover indications, failed handover attempts, packet loss ratio, and handover delay. The simulation results reveal that the proposed scheme achieved 10 to 15% performance gain over existing schemes.
The authors declare that there is no conflict of interests regarding the publication of this paper.
This work was supported by the IT R&D Program of MSIP/KEIT (10041145, Self-Organized Software platform (SoSp) for Welfare Devices). This study was supported by the BK21 Plus project (SW Human Resource Development Program for Supporting Smart Life) funded by the Ministry of Education, School of Computer Science and Engineering, Kyungpook National University, Korea (21A20131600005).