Sage Journals: Discover world-class research

Abstract

IEEE 802.15.4 has become the de facto standard in many areas of wireless sensor network (WSN) such as body area networks, industrial automation, and healthcare domains. In many WSN applications, the sensor nodes could be mobile, and indoor communications are limited in terms of signal propagation. Therefore, a large number of access points need to be deployed to cover large areas. To maintain the sensor connectivity, sensor node should frequently change the serving access point by performing a mechanism known as a handover. We observed that the amount of time required for the association process is the key reason IEEE 802.15.4 is unable to handle the mobility. In this paper, we show what is required for node mobility support and propose three strategies for the support. First, the proactive algorithm is developed to anticipate the future link breakage. Second, a new greedy scanning technique is presented, which prevents nodes from scanning multiple channels. Third, the coordinator selection algorithm is developed which chooses the best coordinator that gives the longest connectivity time reducing the handover frequency. Experimental results have verified that our schemes work well in the mobile sensor network environment.

1. Introduction

A wireless sensor network (WSN) is a network of self-organizing low-powered devices having sensing and communication capabilities [1]. The need for Low Rate Wireless Personal Area Networks (LR-WPANs) has been driven by the large number of emerging applications such as home automation, factory automation, healthcare monitoring, and environmental surveillance [2]. IEEE 802.15.4 standard for LRWPAN has been widely accepted as the de facto standard for WSN that focuses on short-range wireless communications. The goal of the IEEE 802.15.4 LRWPAN is to support low data rate connectivity between wireless sensors with low complexity, cost, and power consumption [3]. In many WSN applications, the sensor nodes need to be mobile. Healthcare wireless sensor networks (HWSNs) are one such area where node mobility is dominant [4]. A number of smart physiological sensors can be integrated into a wearable wireless network, which monitor vital body signs such as heart rate, temperature, blood pressure, ECG, and EEG. If an emergency is detected, the physicians will be immediately informed through the computer system by sending appropriate messages or alarms. Furthermore, HWSN (connected to the Cloud) can also be deployed in home environment which monitors not only human health but also human activities to provide low-cost, high-quality healthcare and social network services to users [5, 6]. HWSNs used to monitor patients should offer mobility support of the sensor nodes carried by the patients. However, supporting mobility in IEEE 802.15.4 brings lots of new challenges and issues.

IEEE 802.15.4 has small coverage area. Thus, large numbers of access points are deployed to cover large areas. Each access point covers a limited area, also called its personal area network (PAN). Hereafter we use the term PAN coordinator or just coordinator to refer to an access point. Mobility of devices causes frequent loss of connection. To maintain the sensor connectivity, sensor node should frequently change their access point by performing a mechanism known as a handover. However, providing IP connectivity to mobile devices means that the devices need to be empowered with sophisticated mobility related IP protocols like MIPv6, and HMIPv6 [7]. Thus, it is not feasible to run complex and sophisticated mobility protocol on a mobile node. Thus, in this paper, we proposed light weighted handover mechanism that can be handled solely by IEEE 802.15.4. In beacon-enabled IEEE 802.15.4, a node is considered to lose connectivity if it is unable to receive beacons from the coordinator for the certain amount of time. Node reassociation defines the procedures for searching new PAN and joining it. There are 16 channels in the 2.4 GHz band, 10 in the 915 MHz band, and 1 in the 868 MHz band allocated for IEEE 802.15.4 [3]. The association starts with the beacon scanning in every channel from the neighboring PANs. Since every channel has to be scanned, it takes certain time, where duration is proportional to the number of channels scanned. Furthermore, the connection is lost once the mobile node moves away from the transmission range of the coordinator. Therefore, the service continuation depends on how and which coordinator is selected in the mobile WSN environments. To support mobility, the association procedure should be modified in such a way that the node's loss of connectivity should be realized quickly, the channel scanning time should be minimized, and the coordinator connectivity time should be maximized.

In this paper, it is shown that handover cannot be handled promptly and energy efficiently in the mobile WSN without proper methods. Thus, the contribution of this paper is threefold. First, using link quality indicator (LQI) history, a proactive algorithm is developed which tries to anticipate if the node is going to lose connectivity before it really happens. Second, we develop a fast coordinator discovery scheme that prevents a node from scanning all available channels. Third, we develop an intelligent algorithm to choose the best coordinator to prolong the node connectivity time as long as possible. All three updates are designed considering the low computational power of the sensor node. In a beacon-enabled network, nodes have to scan every single available channel to discover coordinators around. Here, we present a novel association scheme called greedy channel scan (GCS) that prevents node from scanning all available channels. The main idea of GCS is to scan only first few channels. However, the network setup algorithm has been developed in such a way that scanning only few channels is sufficient for the efficient neighbor discovery and association. The first algorithm tries to anticipate the link breakage by analyzing the LQI history. Thus, the algorithm decreases the time required by node to realize the link breakage from the coordinator. The second algorithm enables mobile node to scan only few channels and acquires network information about all the coordinators in the vicinity. The third algorithm increases the node connectivity time with a coordinator by performing handover to the coordinator that gives the longest connectivity time. The algorithm tries to anticipate if a node is moving away from or toward a coordinator by examining the LQI of the multiple beacons received from the same coordinator. The proposed association scheme provides support for mobility without involvement of any higher layer. Furthermore, the proposed modifications to IEEE 802.15.4 are simple and limited.

The remainder of this paper is organized as follows. In Section 2 existing research efforts have been surveyed. Section 3 presents the brief description on association procedure of beacon-enabled IEEE 802.15.4 followed by the network model in Section 4. Section 5 shows the operation and the main features of the proposed association scheme. The numerical and simulation analysis of the proposed scheme have been described in Sections 6 and 7. Finally, we conclude the paper in Section 8.

2. Related Works

There are some efforts done to minimize the association duration in IEEE 802.15.4. In IEEE 802.15.4e [8], optional fast association (FastA) is defined, which allows a device to associate in a reduced duration. However, most of the efforts are limited to mobility management, decreasing the duration of association message exchange [9], increasing connectivity [10] or coordinator discovery [11], whereas the channel scanningpart has been left untouched. The authors in [12] have presented an interesting solution to decrease the scanning duration by using dedicated channel for the beacon transmission. Although the results presented were impressive, using dedicated channel for beacon may be the waste of available bandwidth. Similarly, there are several works on multichannel solutions, but they are limited to throughput improvement or beacon collision avoidance [13–15]. To our best knowledge, we are the first to propose the handover solution in IEEE 802.15.4 considering channel scanning also.

In [9], Zhang et al. proposed an improved association scheme called Simple Association Process (SAP) that eliminates the redundant primitives, thus decreasing the packet collisions and the association delay.

A fast association mechanism [16] is proposed for real-time WPAN applications. Delay caused by scanning multiple channels is reduced because the scanning process is stopped as soon as a beacon is received. Although this scheme prevents nodes from scanning all available channels, nodes still need to scan multiple channels before finding a coordinator. Furthermore, the first beacon received may not always be the suitable coordinator.

Similarly, there are other works which focus on the neighbor discovery for quick association. In [11], algorithms are proposed for the optimized discovery of IEEE 802.15.4 static and mobile networks operating in multiple frequency bands and with different beacon intervals. In [10], a scheme to increase coordinator connectivity time with mobile nodes is presented for IEEE 802.15.4 beacon-enabled networks. Nodes use time-stamp of received beacons during the scan, along with link quality to determine the appropriate coordinator for association. Other mobility management schemes for cluster tree based WPAN have been proposed by Chaabane et al. [17] and Bashir et al. [18]. These approaches use the speculative algorithm for node association based on LQI. Based on LQI value, the mobile node anticipates cell change based on LQI before the loss of connection and tries to associate with the next coordinator. However, in all the cases, nodes have to scan multiple channels to find coordinators.

3. Beacon-Enabled IEEE 802.15.4

The IEEE 802.15.4 standard supports three kinds of topology: star, peer-to-peer, and cluster tree topologies, which can operate on beacon and nonbeacon-enabled modes. In this paper, we consider beacon-enabled mode only. Therefore, in the remaining of this section, we explain the beacon-enabled mode and node association process. In the beacon-enabled mode, communication is synchronized and controlled by a PAN coordinator, which transmits periodic beacons. The beacon contains information related to PAN identification, synchronization, and superframe structure. The period between two consecutive beacons is called a superframe and can have an active portion and an inactive portion as shown in Figure 1. The structure of superframe is determined by coordinators using two parameters: superframe order (SO) and beacon order (BO). SO determines the length of active portion of superframe, whereas BO defines the beacon interval. The relationship between BO and SO can be expressed as $0 \leq SO \leq BO \leq 14$ . The active portion of each superframe is divided into 16 equally spaced slots. The first slot is used for beacon transmission, whereas the rest of the superframe duration is further subdivided into the contention access period (CAP) and an optional contention-free period (CFP). All communications must take place during the active part, and devices can sleep to conserve energy in the inactive part.

Figure 1

The superframe structure of IEEE 802.15.4.

PAN coordinator should perform the energy detection (ED) to detect the peak energy of a channel and choose an appropriate channel for data transmission. In each channel, ED scan is performed for the duration of $t_{scan}$ symbols. Then, coordinator carries out an active scan to locate any coordinator transmitting beacon frames within its personal operating space (POS). During the active scanning, coordinator first sends out a beacon request command and waits for the duration of $t_{scan}$ symbols. If a beacon could not be detected during $t_{scan}$ , the coordinator can construct its own PAN by broadcasting its periodic beacons. After a PAN has been initialized, the other devices in the POS of the PAN can communicate with the coordinator and associate with this PAN.s.

3.1. Coordinator Discovery and Association

In order to start association, a sensor node needs to know the PAN's physical channel, coordinator ID, addressing mode, and PAN ID. Thus, the passive scan is performed for the coordinator discovery. As shown in Figure 2(a), during the passive scan, a device searches for beacon frame in each channel for the duration of $t_{scan}$ symbols and records the beacon frames received in each channel. If no beacon is detected, the device starts another passive scan after a period of time. During a passive scan the node discards all frames received that are not beacon frames. IEEE 802.15.4 maintains a list called PANDescriptor that records all the beacons received. However, the standard protocol does not specify in detail on the coordinator selection procedure, and we assume the node will mostly choose the nearest coordinator. Once the node selects the suitable coordinator, it starts association procedure by sending a request for association as shown in Figure 3.

Figure 2

Channel scanning mechanism in beacon-enabled IEEE 802.15.4.

Figure 3

Association procedure in IEEE 802.15.4.

3.2. Loss of Connectivity and Reassociation

Node loses connectivity if it moves beyond the coverage of PAN. A node considers itself as orphan node if it cannot receive beacon from the coordinator for aMaxLostBeacons times. Orphan scan allows a device to attempt to relocate its coordinator following a loss of connectivity. The device shall first send the orphan notification command frame and waits for coordinator realignment command frame for at most macResponseWaitTime symbols as shown in Figure 2(b). This procedure is repeated until it receives the coordinator realignment frame or all the available channels are scanned. If the orphan scan is unsuccessful, the device looks for a new parent by performing the passive scan. During an orphan scan, the node discards all frames received that are not coordinator realignment MAC command frames. For a mobile node, the connection with coordinator can frequently break and the reassociation procedure is time consuming because the mobile node has to go through the orphan scan first followed by the passive scan and then the association message exchange. We use the term reassociation to explicitly denote the association procedure used by the mobile node. From our study, it is observed that IEEE 802.15.4 can be used in the mobile sensor network applications if somehow this whole reassociation duration be shortened to some tolerable level.

4. Network Model

One obvious choice for IEEE 802.15.4 to cover a large areas is by using cluster tree topology. However, a cluster tree topology presents two major problems. First, the collision probability is high since all nodes transmit on the same channel. Second, the IEEE 802.15.4 standard does not specify how to synchronize a cluster tree network [17]. Thus, we select the start topology to balance the network load. But, communication between different PAN coordinators is not possible unless they belong to the same channel, or if they define a common transmission channel. In this paper, we consider that different wiredly connected PANs can form a unique heterogeneous network composed of star PANs. All PANs are assumed to be connected with the base station through the wired connection. Messages between nodes that do not belong to the same PAN Id can then be routed through the base station.

The basic network design is shown in Figure 4, which models a typical HWSN. The construction of a HWSN comprises three main elements, namely, (i) a base station or gateway that acts as a bridge between the HWSN and Internet, (ii) PAN coordinator that support communication to/from the sensor nodes, and (iii) the sensor nodes themselves that collect body parameters and send them wirelessly over the network. Due to the limited coverage area of each PAN in indoor environments, several PAN coordinators are deployed to cover the monitored zone as shown in the figure. Furthermore, to prevent interference, each PAN may operate in a different transmission channel. Mobile node changes the point of attachment as it pass from a PAN to another in the network. For the maximum coverage of the area, the coordinators are assumed to be septated by the distance of transmission range. However, adjacent PAN does not interfere each other as they operate in different channels.

Figure 4

Network model.

In this work, LQI value is used to detect the movement of a sensor node. In IEEE 802.15.4, every MAC frame contains LQI value that ranges from 0 to 255. The LQI measurement is a characterization of the strength and/or quality of a received packet. However, the calculation of the LQI is not specified in the IEEE 802.15.4 standard. However, receiver ED, signal-to-noise ratio (SNR), or combination of these methods can be used. In order to observe the change in LQI with respect to distance, mobility, and background traffic, we conducted a simple experiment in NS-2. In PAN1 of Figure 4, node 1 is moved toward node 5 at the speed of 1 m/s. Transmission range of PAN was set to 10 meters. The effect of interference is also accounted by introducing background traffic from nodes 2, 3, and 4. LQI values of beacon received from the coordinator as node 1 moves are shown in Figure 5. In NS-2, the LQI is calculated based on the received signal strength and the signal-to-noise ratio. A packet is only received if its LQI is equal or greater than 128. Thus, LQI obtained ranges from 128 to 255 because the packets whose LQI is below 128 are dropped. Also, the LQI value of 255 is observed when the distance between node and coordinator is less than 7 m and the gradual decrease in LQI is observed once the distance from the coordinator exceeds 7 m. With movement, some false LQI values are observed, and the LQI drops as the sensor node moves away from the coordinator. LQI is also affected by the background traffic because of interference and, especially, when the node is moving, where sudden drops in LQI values are observed. However, in this work, the normalized value of LQI is used to detect node movement. In [10, 18], authors also obtained similar LQI result.

Figure 5

Change in LQI with distance and background traffic.

5. Proposed Scheme

From the detailed study of IEEE 802.15.4, it is observed that mainly three updates are possible to enable IEEE 802.15.4 to efficiently handle mobility. The first update is the early detection of loss of connectivity. The detection of loss of connectivity by counting loss of beacons for aMaxLostBeacons times is time consuming. A proactive method can significantly reduce the detection time of link breakage. The second update is the reduction of neighbor discovery time. The channel scanning for neighbor discovery is the most time consuming procedure. If somehow the nodes are prevented from scanning multiple channels, the mobility can be handled efficiently. And, the third update is to retain the node connectivity as long as possible. IEEE 802.15.4 does not specify how to select a coordinator. However, if the coordinator that gives longest connectivity time is selected, the frequency of handover can be reduced. In our proposed scheme, we implement all above mentioned updates to handle the mobility efficiently. Also, all the three updates are simple and can be easily implemented.

5.1. Initialization of PAN Coordinator

IEEE 802.15.4 has 16 working channels from channel 11 to 26 in 2.4 Ghz band. Among them, channels 15, 20, 25, and 26 are interference free from Wi-Fi networks [19]. The set of these channels ${15, 20, 25, 26}$ is called clear channels. PAN coordinator performs the energy detection to detect the peak energy of all channels. Then, coordinator carries out an active scan to locate its neighbors and records them. Based on the results of energy detection and the active scan, the coordinator will choose an appropriate channel for data transmission such that the selected channel has no interference with the neighboring PANs. However, the first priority is given to clear channels.

5.2. Beacon Frame

Once the PAN is initialized, it starts transmitting periodic beacon. However, 1 more parameter is transmitted in the beacon. The occupied bandwidth (OBW) of the PAN is transmitted in the beacon payload field. The OBW represents the volume of traffic handled by the PAN. In every superframe, each PAN obtains the bandwidth occupied for data transmission during one superframe period which is expressed as

\begin{matrix} T_{com} = Beacon + \sum_{1}^{n} ‍ (Dat a_{n} + ACK + IFS), \end{matrix}

(1)

where

T_{com}

is the superframe duration used for communication and n is the number of data packets handled. Beacon, Data

_{n}

, and ACK are the time duration of beacon, data, and ACK frames, respectively, and IFS is the interframe space duration. OBW is then converted to the number ranging from 0 to 255 as follows:

\begin{matrix} OBW = \frac{T_{com}}{BI} \times 255, \end{matrix}

(2)

where BI is the beacon interval of the PAN coordinator.

5.3. Proactive Reassociation Decision

LQI is an important metric available in IEEE 802.15.4 for detection of link quality between two communicating nodes. Make-before-break approach requires either sensor node or coordinator to monitor the LQI for triggering the handover. However, the frequency of beacons can be low or high depending on BO of the coordinator. At higher BO, beacons are received after longer intervals resulting in delayed detection of link quality. In our proactive reassociation decision (PRD) scheme, sensor node monitors all packets from the coordinator and are used for analyzing the LQI. PRD algorithm shown in Figure 6 is used to take handover decision. The algorithm first observes the LQI value of each packet received from the coordinator. The node anticipates the link breakage by analyzing the LQI history and missing of a beacon. But, the link quality can be degraded due to increasing in distance, interference, and collision of packets. Thus, instantaneous LQI alone is not a reliable parameter for taking handover decision. Furthermore, observed LQI values can fluctuate at any instance of time. Therefore, successive LQI readings are considered. A counter shown in Figure 6 keeps track of how many times the LQI is less than previously recorded LQI. In addition to LQI drop, missing of beacon is used to ensure the link breakage. If LQI drops continuously for MaxCounter number of times followed by missing of a beacon, the reassociation decision is taken. The algorithm is designed in such a way that the fluctuation in the reading is compensated by decreasing the counter as shown in the flow chart.

Figure 6

Flow chart of proactive reassociation decision.

5.4. Greedy Channel Scan

Channel scanning is the most time consuming part of the association procedure. The time spent on association for various values of BO is shown in Figure 7. For each value of BO, the time required for association as the number of channels to scan varies from 1 to 16 is shown in the figure. The figure illustrates that the association time increases as the number of channel increases. Thus, to prevent the node from scanning every channel, greedy channel scan (GCS) scheme is proposed. In the GCS scheme, a mobile node scans the clear channels first. However, if the last used channel is also a clear channel, then it will scan clear channels excluding the last used one. For example, if the last used channel is 20, then it will scan channels 15, 25, and 26 only. The idea behind omitting its last used channel from scanning is because the adjacent PANs will not use the same channel. However, if somehow the greedy algorithm is unable to find PAN, then all channels are scanned again starting from channel 11. Thus, in an average mobile node scans only 3 channels.

Figure 7

Time required for the association as BO and number of channels are varied.

5.5. PAN Selection and Association

The PAN selection algorithm (PSA) is explained with an example. Figure 8 shows an intersection hall way in a hospital where 4 PAN coordinators W, X, Y, and Z are deployed for network coverage. The mobile node M (initially associated with X) moves toward coordinator Y at the speed of 1 m/s. The circle of coordinator is the coverage region of a PAN. IEEE 802.15.4 maintains a list called PANDescriptor that records all the beacons received and contains fields such as PAN ID, coordinator address, logical channel, and LQI (LQI of beacon received). In IEEE 802.15.4, the multiple beacons received from the same coordinator are ignored. However, in PSA, we analyze LQI of the multiple beacons from the same coordinator to anticipate the node's direction of movement. Two more fields are added in PANDescriptor as shown in Table 1. At the point P2 (in Figure 8), M starts the passive scan. Table 2 shows the instance of PANDescriptor maintained by M during the passive scan. If a beacon is received for the first time, LQI_Prev and LQI are made the same. If the beacon is received again from the same coordinator, observations are done such that LQI holds the current LQI and LQI_Prev field holds the average value of LQI of beacons received. Direction field (initially zero) is decreased by one if the current LQI is less than the LQI_Prev, and otherwise, it is increased. The positive value of Direction field indicates that the node is moving towards that PAN and negative indicates that node is moving away. But, getting associated with the PAN that gives longest connectivity is not sufficient for a reliable and stable handover. The new PAN should also be able to support the continuous transmission service even after the handover. Thus, the occupied bandwidth of PAN that is broadcast in beacon is also considered for the coordinator selection. Using the observed average LQI (LQI_Prev) and OBW, the PAN weightage (PANW) is calculated as follows:

\begin{matrix} PANW = α \times OBW + (1 - α) \times LQI_Prev, \end{matrix}

(3)

where α (

0 \leq α \leq 1

) and

1 - α

are the weights for occupied bandwidth and link quality, respectively. Based on above values, node uses PAN selection algorithm (PSA) (Algorithm 1) to select the optimal PAN for the association. In the example of Figure 8, coordinator Y is selected because coordinators W and Z have negative values of Direction (assuming equal OBW for all).

Table 1

Added fields of PANDescriptor in GCS.

Field	Description
LQI_Prev	LQI value of last received beacon
Direction	Counter to detect increase or decrease in LQI

Table 2

An example of PANDescriptor maintained by the mobile node M.

Time	Coordinator W/Z			Coordinator Y
Time	LQI_Prev	LQI	Direction	LQI_Prev	LQI	Direction
$T 1$	200	200	0	255	255	0
$T 2$	197	193	−1	255	255	1
$T 3$	192	187	−2	255	255	2
$T 4$	187	181	−3	255	255	3

Algorithm 1: PAN selection algorithm.

(1) Compare the Direction and PANW of every PAN

(2) Select that PAN which has positive Direction and lowest PANW

(3) If all Directions $\leq$ 0, select the PAN with the lowest PANW

Figure 8

Selection of PAN.

6. Numerical Analysis

Let aBaseSuperFrameDuration be the number of symbols forming a superframe when SO = 0. In IEEE 802.15.4, it takes equal duration in performing ED, active, and passive scan on a channel which is given by

\begin{matrix} t_{scan} = a B a s e S u p e r F r a m e D u r a t i o n \times (2^{BO} + 1) . \end{matrix}

(4)

6.1. Association

The total time spent for association is the sum of time spent in the channel scan and the time spent in the association message exchange (Asso $_{msg}$ ). In IEEE 802.15.4, sensor node performs the passive scan in all available n channels, but only s ( $s < n$ ) channels are scanned in the case of GCS. The total time spent for association in a newly joining node with both protocols are given by

\begin{matrix} t_{802.15.4_asso} = n \times t_{scan} + Ass o_{msg}, \end{matrix}

(5)

\begin{matrix} t_{gcs_asso} = s \times t_{scan} + Ass o_{msg} . \end{matrix}

(6)

From (5) and (6), GCS is able to decrease the association time of newly joining node by almost the factor of

n / s

6.2. Detection of Connectivity Loss

In IEEE 802.15.4, the beacon interval (BI) is given by the following equation:

\begin{matrix} BI = a B a s e S u p e r F r a m e D u r a t i o n \times 2^{BO} . \end{matrix}

(7)

In IEEE 802.15.4, mobile node starts the orphan scan, if it misses the beacon for aMaxLostBeacons times. However, in the case of GCS, reassociation is performed if the beacon is missed just once. Thus, the time required for the detection of loss of connectivity (DLC) can be calculated as

\begin{matrix} DL C_{802.15.4} = BI \times a M a x L o s t B e a c o n s, \\ DL C_{gcs} = BI . \end{matrix}

(8)

6.3. Reassociation

Once a node realizes that it has lost connectivity with the PAN, the node performs the orphan scan on each channel for duration of $m a c R e s p o n s e W a i t T i m e (32 \times a B a s e S u p e r F r a m e D u r a t i o n)$ symbols until its parent PAN is found or all n channels are scanned. Upon the failure of orphan scan, a new PAN is searched through passive scan as mentioned above. Thus, the total time spent for reassociation is given by

\begin{matrix} t_{orphan_scan} = n \times m a c R e s p o n s e W a i t T i m e, \\ t_{802.15.4_reasso} = t_{orphan_scan} + t_{802.15.4_asso} . \end{matrix}

(9)

However, in the proposed scheme, (6) also gives the reassociation duration since there is no orphan scan.

6.4. Connectivity Time

To evaluate the device connectivity time with its coordinator, we consider a four-way intersection, which models the corridor of a hospital as shown in Figure 8. We assume that all coordinators have the same transmission range and are uniformly spaced. The figure shows a situation in which the mobile node M can take any direction at the intersection. In this kind of situation, choosing the furthest coordinator will give the longest connectivity provided that the mobile node is moving toward it. Let $C_{1}$ be remaining distance in meter before the mobile node M moves out of the transmission range of coordinator W and Z. Similarly, $C_{2}$ be the remaining distance before M moves out of the transmission range of Y. Suppose the mobile node moves towards coordinator Y at the speed of v m/s. As can be seen from the figure, the connectivity time for the coordinators W or Z is $C_{1} / v$ and for the coordinator Y is $C_{2} / v$ . Since $C_{2}$ is larger than $C_{1}$ , choosing $C_{2}$ will give the longest connectivity time. PSA algorithm also select coordinator Y as explained before. PSA algorithm selects that coordinator which gives the longest connectivity time irrespective of moving direction of mobile node.

6.5. Numerical Example

Assuming network parameters of Table 4, we get aBaseSuperFrameDuration of 15.36 ms and macResponseWaitTime and Asso_msg of 0.49 s [3]. Thus, using these values and above equations the association time and DLC duration for both IEEE 802.15.4 and proposed scheme is calculated and shown in Table 3. For GCS, ( $s = 3$ ) is used. From the values obtained in the table, we can conclude that GCS reduces the association time of IEEE 802.15.4 by significant amount. Also, as can be seen from the table DLC is also an important parameter to decrease the association delay.

Table 3

Total time spent for association.

BO/SO	Beacon interval	$t_{scan}$	Association		DLC		Reassociation
BO/SO	Beacon interval	$t_{scan}$	802.15.4	Proposed	802.15.4	Proposed	802.15.4	Proposed
0	15.36 ms	30.72 ms	0.98 s	0.58 s	0.06 s	0.02 s	8.82 s	0.58 s
1	30.72 ms	46.08 ms	1.22 s	0.63 s	0.12 s	0.03 s	9.07 s	0.63 s
2	61.44 ms	76.8 ms	1.72 s	0.72 s	0.25 s	0.06 s	9.56 s	0.72 s
3	122.88 ms	138.24 ms	2.7 s	0.90 s	0.49 s	0.12 s	10.54 s	0.90 s
4	245.76 ms	261.12 ms	4.66 s	1.27 s	0.98 s	0.25 s	12.51 s	1.27 s
5	491.52 ms	506.88 ms	8.6 s	2.01 s	1.97 s	0.49 s	16.44 s	2.01 s
6	983.04 ms	998.4 ms	16.46 s	3.49 s	3.93 s	0.98 s	24.30 s	3.49 s

Table 4

Network parameters and values.

Parameter	Value
Assumed power supply	3 V
Reception power	56.5 mW
Transmission power	48 mW
Idle power	2.79 mW
Sleep power	30 μW
Transition time	192 μs
Transmission range	10 m
aMaxLostBeacons	4
α	0.5
Routing	AODV
Frequency band	2.4 GHz
Radio data rate	250 kbps
Number of channels	16
LQI_THRESHOLD	150
Traffic	CBR
Buffer size	10 packets
Packet size	100 B

7. Performance Evaluation

The proposed scheme is implemented on the NS-2 network simulator [20]. The simulated nodes were configured by using the parameters listed in Table 4. The parameters of Table 4 are taken from CC2420 datasheet [21]. 16 PANs were deployed for the coverage of 50 by 50 m area as shown in Figure 9. At the simulation time of 50 s, the mobile node (node 1) starts to transmit data and also moves to the direction as indicated by the arrows and simulation stops when it comes to its end position (left bottom position). The mobile nodes switches PAN as it moves. The total distance travel by the mobile node is 86 m. The mobile node transmits data at the rate of 4 Kbps to the base station (black node at the center). Coordinators choose their operational channel giving priority to clear channels. The communication link between a PAN and the base station is wired. Coordinators transmit beacons at the rate corresponding to $BO = 2$ . The OBW is equal in all PANs.

Figure 9

Network topologies used in the simulation.

7.1. Association Time and Energy

The total time spent for DLC and the re/association in the various values of BO is obtained from numerical analysis and NS-2 simulations as shown in Figure 10. On an average, GCS performed the passive scan on 3 channels, thus $s = 3$ is used for analytical calculation. In all cases, the analytical results match well with the simulation results. As shown in the figure, the time required by GCS for re/association is much lower because only 3 channels were scanned for the association procedure. However, in the case of IEEE 802.15.4, it scans all available 16 channels spending significant amount of time and energy. Furthermore, as shown in the figure, PRD algorithm successfully decreased DLC duration. At BO = 3, the proposed scheme is able to decrease the DLC duration by 4 times, the node association by 3 times, and the reassociation duration by 11.71 times. The energy spent by the mobile node for re/association under various BO is shown in Figure 11. Since the time spent by the proposed scheme is much lesser, it has a direct proportional impact in the energy consumption. At BO = 3, proposed scheme is able to decrease the node association energy by 7.5 times and the reassociation energy by 17.75 times. Thus, from above observations, we can conclude that proposed scheme is able to save both energy and time by the significant amount.

Figure 10

Total time spent for association at different beacon intervals.

Figure 11

Total energy spent for association at different beacon intervals.

7.2. Connectivity and Connection Time

Percentage of connectivity is calculated as the ratio of the time that a node is associated with the network to the total simulation time. In the proposed scheme the total time spent on passive discovery of a coordinator and the detection of link breakage is less. Furthermore, nodes do not perform orphan scan. Thus, the node quickly gets associated with a new PAN as soon as it lost connection with the old one. The percentages of node connectivity at different BO(s) and node speeds are shown in Figure 12. As shown in the figure, nodes in the proposed scheme are able to be connected to the network during the most of the time, whereas IEEE 802.15.4 has low connectivity. We observed that even a slight mobility has a significant negative impact on the node connectivity in the case of IEEE 802.15.4. At the human walking speed of 1.5 m/s, IEEE 802.15.4 had poor connectivity even at the lower values of BO and nodes were unable to connect when $BO = 5$ . It is observed that the poor connectivity of IEEE 802.15.4 is due to the long duration required for reassociation (takes 10.54 s at $BO = 3$ ). Furthermore, the time required for the detection of link breakage also decreases the node connectivity. The average duration of time for the connection with a PAN is shown in Figure 13. As seen in the figure, the connection time is longer in the proposed scheme as compared to IEEE 802.15.4. The connection time decreases as the node speed is increased. In our simulation, there were total 8 handover possible before mobile nodes come to end. In an average IEEE 802.15.4 was able to perform 5 handovers, whereas the proposed scheme was able to perform 8 handovers. At the speed of 1.5 m/s and $BO = 3$ , IEEE 802.15.4 was able to perform just 3 handovers, whereas proposed scheme performed all 8 handover successfully. The time spent by the mobile node within the transmission range of a coordinator might not be enough to complete the association, which accounts for poor association rate in IEEE 802.15.4.

Figure 12

Node connectivity at different BO using proposed scheme and IEEE 802.15.4 at 0.5, 1, and 1.5 m/s.

Figure 13

Connection time with coordinator at different BO using proposed scheme and IEEE 802.15.4 at 0.5, 1, and 1.5 m/s.

7.3. Throughput

The effect of handover on throughput of node 1 is shown in Figure 14. The figure shows the throughput of node 1 calculated every second. The mobile node is using superframe duration corresponding to BO = 3 and speed of 1 m/s. Throughput of proposed scheme drops while performing handover because packets cannot be transmitted in the periods of passive discovery and association. However, nodes quickly get associated with a new coordinator and restart data transmission. Nodes can buffer packets. Therefore, buffered data are also transmitted after new association is completed, resulting into increased throughput immediately after the association. The total transferred rate was 3600 bps for the proposed scheme whereas 2200 bps for IEEE 802.15.4. As can be seen from the figure, during handover, throughput of the proposed scheme never dropped to zero, whereas it dropped to zero in every handover in case of IEEE 802.15.4. Also, the packet delivery ratio was observed 90% for proposed scheme and 55% for IEEE 802.15.4, respectively.

Figure 14

Throughput observed in the PAN coordinator at BO = 3, speed = 1 m/s, and data rate of 4 Kbps.

8. Conclusion

In this paper, three strategies are proposed for the efficient handover in IEEE 802.15.4. The first algorithm (PRD) anticipates if the node is going to lose connectivity by analyzing the LQI history before it really happens. Thus, helps in early detection of future link breakage and handover decision. Second algorithm called GCS is presented that can decrease both time and energy required for association. GCS algorithm scans the clear channels first preventing nodes from scanning all the available channels and looking for beacon. The third algorithm anticipates the nodes direction of movement with respect to the coordinator and selects that coordinator which has maximum available bandwidth and towards which the node is moving. Our analytical and simulation results demonstrated that our scheme is highly efficient in terms of both energy and time. With the implementation of our scheme, we give IEEE 802.15.4 the new ability to handle mobility.

Footnotes

Conflict of Interests

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

Acknowledgment

This study was supported by research funds from Chosun University, 2012.

References

Sthapit

Pyun

J.-Y.

Medium reservation based sensor MAC protocol for low latency and high energy efficiency

Telecommunication Systems 2013 52 4 2387 2395

2-s2.0-79960866820

10.1007/s11235-011-9551-z

Zen

Habibi

Rassau

Ahmad

Performance evaluation of IEEE 802.15.4 for mobile sensor networks

Proceedings of the 5th IEEE and IFIP International Conference on Wireless and Optical Communications Networks (WOCN '08)

May 2008

Surabaya, Indonesia

1 5

2-s2.0-51049120288

10.1109/WOCN.2008.4542536

IEEE 802.15.4

Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specifications for Low-Rate Wireless Personal Area Networks (LR-WPANs)

IEEE, September 2006

Caldeira

Rodrigues

Lorenz

Intra-mobility support solutions for healthcare wireless sensor networks-handover issues

IEEE Sensors Journal 2012 13 11 4339 4348

10.1109/JSEN.2013.2267729

Hung

L. X.

Lee

Truc

P. T. H.

The Vinh

Khattak

A. M.

Han

Dang

V. H.

Hassan

M. M.

Kim

Koo

K.-H.

Lee

Y.-K.

Huh

E.-N.

Secured WSN-integrated cloud computing for u-life care

Proceedings of the 7th IEEE Consumer Communications and Networking Conference (CCNC '10)

January 2010

Las Vegas, Nev, USA

1 2

2-s2.0-77951268185

10.1109/CCNC.2010.5421618

Jung

J. J.

Knowledge distribution via shared context between blog-based knowledge management systems: a case study of collaborative tagging

Expert Systems with Applications 2009 36 7 10627 10633

2-s2.0-67349103070

10.1016/j.eswa.2009.02.052

Gargi

Saif Shams

S. M.

Akhbar

A. H.

Raza H.M.

M. T.

Kim

K.-H.

Yoo

S.-W.

Network assisted mobility support for 6LoWPAN

Proceedings of the 6th IEEE Consumer Communications and Networking Conference (CCNC '09)

January 2009

Las Vegas, Nev, USA

1 5

2-s2.0-63749083899

10.1109/CCNC.2009.4784871

IEEE P802.15.4e-2012

Part 15.4: Low-Rate Wireless Personal Area Networks (WPANs), Amendment 1: MAC sub-layer

February 2012

Zhang

Wang

Dai

Performance evaluation of IEEE 802.15.4 beacon-enabled association process

Proceedings of the 22nd International Conference on Advanced Information Networking and Applications (AINAW '08)

March 2008

Okinawa, Japan

541 546

10.1109/WAINA.2008.241

10.

Zen

Habibi

Ahmad

Improving mobile sensor connectivity time in the IEEE 802.15.4 networks

Proceedings of the Australasian Telecommunication Networks and Applications Conference (ATNAC '08)

December 2008

Adelaide, Australia

317 320

2-s2.0-64049113167

10.1109/ATNAC.2008.4783343

11.

Karowski

Viana

A. C.

Wolisz

Optimized asynchronous multi-channel neighbor discovery

Proceedings of the IEEE INFOCOM

April 2011

Shanghai, China

536 540

2-s2.0-79960867045

10.1109/INFCOM.2011.5935221

12.

Sthapit

Choi

Y. S.

Kwon

G. R.

Hwang

S. S.

Pyun

J. Y.

A fast association scheme over IEEE 802.15.4 based mobile sensor network

Proceedings of the 9th International Conference on Wireless and Mobile Communications (ICWMC '13)

July 2013

Nice, France

179 184

13.

Osterlind

Dunkels

Approaching the maximum 802.15.4 multi-hop throughput

Proceedings of the 5th ACM Workshop on Embedded Networked Sensors (HotEmNets '08)

June 2008

Charlottesville, Va, USA

1 6

14.

Shih

B.-Y.

Chang

C.-J.

Chen

A.-W.

Chen

C.-Y.

Enhanced MAC channel selection to improve performance of IEEE 802.15.4

International Journal of Innovative Computing, Information and Control 2010 6 12 5511 5526

2-s2.0-78650291998

15.

Toscano

Bello

L. L.

Multichannel superframe scheduling for IEEE 802.15.4 industrial wireless sensor networks

IEEE Transactions on Industrial Informatics 2012 8 2 337 350

2-s2.0-84859905227

10.1109/TII.2011.2166773

16.

Meng

Han

A new association scheme of IEEE 802.15.4 for real-time applications

Proceedings of the 5th International Conference on Wireless Communications, Networking and Mobile Computing (WiCom '09)

September 2009

Beijing, China

1 5

2-s2.0-73149091656

10.1109/WICOM.2009.5301369

17.

Chaabane

Pegatoquet

Auguin

Jemaa

M. D.

Energy optimization for mobile nodes in a cluster tree IEEE 802.15.4/ZigBee network

Proceedings of the IEEE Computing, Communications and Applications Conference (ComComAp '12)

January 2012

Hong Kong, China

328 333

2-s2.0-84860464628

10.1109/ComComAp.2012.6154866

18.

Bashir

Baek

W. S.

Sthapit

Pandey

Pyun

J. Y.

Coordinator assisted passive discovery for mobile end devices in IEEE 802.15.4

Proceedings of the IEEE Consumer Communications and Networking Conference (CCNC '13)

January 2013

Las Vegas, Nev, USA

601 604

10.1109/CCNC.2013.6488506

19.

Tas

N. C.

Sastry

Song

IEEE 802.15.4 throughput analysis under IEEE 802.11 interference

Proceedings of the International Symposium on Innovationsand Real Time Applications of Distributed Sensor Networks

2007

Shreveport, Louisiana

20.

Ns2, http://nsnam.isi.edu/nsnam/index.php/Main_Page

21.

Texas Instrument

CC2420 Data Sheet, http://www.ti.com/lit/ds/symlink/cc2420.pdf

Handover Strategies in Beacon-Enabled Mobile Sensor Network

Abstract

1. Introduction

2. Related Works

3. Beacon-Enabled IEEE 802.15.4

3.1. Coordinator Discovery and Association

3.2. Loss of Connectivity and Reassociation

4. Network Model

5. Proposed Scheme

5.1. Initialization of PAN Coordinator

5.2. Beacon Frame

5.3. Proactive Reassociation Decision

5.4. Greedy Channel Scan

5.5. PAN Selection and Association

Algorithm 1: PAN selection algorithm.

6. Numerical Analysis

6.1. Association

6.2. Detection of Connectivity Loss

6.3. Reassociation

6.4. Connectivity Time

6.5. Numerical Example

7. Performance Evaluation

7.1. Association Time and Energy

7.2. Connectivity and Connection Time

7.3. Throughput

8. Conclusion

Footnotes

Conflict of Interests

Acknowledgment

References