Sourav Dhar et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5509-5525
DESIGN AND SIMULATION OF VERTICAL HANDOVER ALGORITHM FOR VEHICULAR COMMUNICATION Sourav Dhar, Amitava Ray, Rabindranath Bera Sikkim Manipal Institute of Technology, Majitar, Sikkim, INDIA- 737136 Email:
[email protected] Abstract: The primary objective in this work is to design and simulate a novel handover algorithm in heterogeneous radio network for intelligent transportation system (ITS). Analytic hierarchy process (AHP) is used to solve decision problem in multiple constraint environment. The user preferences are taken into consideration in this algorithm. A case study is illustrated to demonstrate the effectiveness of this model. Several network contexts like received signal strength (RSS), initial delay for connection establishment, network traffic load and bandwidth offered, service context like usage cost, user contexts like type of application and terminal context like speed of the vehicle are considered as different attributes of this algorithm. RSS is considered as the triggering factor, i.e, a network will be considered as an alternative only if its RSS is above threshold. The proposed method decides the priority of radio access network that is most suitable for user’s application at a particular vehicular speed in the constraint resource environment. Sensitivity analysis is done to justify the design of the algorithm. This algorithm is specific to vehicular communication system and hence variation in network selection with vehicle speed is presented. The results show that the presented model not only realistically optimizes the best available network on the move but also avoids unnecessary handovers. Simulation work has been carried out in the MATLAB environment. Key Words: Intelligent Transportation System (ITS), Analytic Hierarchy Process (AHP), Dedicated Short Range Communication (DSRC), Software Defined Radio (SDR), Sensitivity Analysis. 1.
Introduction
Intelligent Transportation System (ITS) converges remote sensing and communication technologies to improve safety and to make the journey enjoyable. This is now part of the national strategy, for improving the security, safety, efficiency and comfort, of every nation [1]. Association for Intelligent Transport Systems (AITS), INDIA and Intelligent Transport Systems Center of Excellence (India) are involved to promote the concept of road safety in India. The government of INDIA has shown their interest to take up research, policy recommendations, programs and projects on ITS with AITS. AITS India is an umbrella organization which represents a broad spectrum of members including industry, government, consumer organizations and academia in fostering the development and deployment of ITS in India. Such systems often include components used in electronic commerce, and related applications. AITS India also has strong links internationally with similar organizations including ITS America, ITS Australia and ITS Asia-Pacific. ITS applications can be broadly classified in two categories: safety applications and non safety application. Adaptive cruise control (ACC), Driver assistance systems (DAS) and Collision avoidance and warning systems (CAWAS) are some examples of safety application. On the other hand, internet access from car, vehicle to vehicle (V2V) and vehicle to infrastructure (V2I) communication [2, 3], satellite television from car, radio taxi, and automatic toll collection are categorized as non-safety applications. Intelligent Transportation System (ITS) needs an “always best connected (ABC)” communication to ensure vehicular safety. Intelligent Transportation System (ITS) is an application of 4G which is characterized by heterogeneous radio access networks. The heterogeneous co-existence of access technologies with largely different characteristics creates a decision problem for determining best available network at a point of time. An intelligent, Network independent handoff which is also known as vertical handoff, decision algorithm is an important issue for a seamless “ABC” vehicular communication. A review of vertical handoff decision algorithms are given in [4,5]. A robust algorithm should ensure good quality of service (QoS) irrespective of physical environment [6]. QoS depends on the type of application, i.e., conversational, streaming and interactive applications require different bandwidth, delay, jitter etc. A right vertical handoff decision algorithm by determining the “best”
ISSN: 0975-5462
5509
Sourav Dhar et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5509-5525 network at “best” time among available networks based on dynamic factors such as “Received signal strength (RSS)” of network and “velocity” of mobile station simultaneously with static factors like usage expenses, link capacity and power consumptions is presented in [7]. Importance of mobile station velocity and movement pattern is considered in [8]. Competitive and cooperative relationships among the major ITS communication technologies, WiMax, WLAN and UMTS, are considered in [9]. An analytic hierarchy process (AHP) based network selection algorithm for UMTS and WLAN is presented in [ 3,10]. Along with WCDMA and WLAN, DVB-H which supports both data and broadcast services correspondingly, is explained in [11]. In [12], the vertical handover decision problem is formulated as a Constrained Markov Decision Process (CMDP). A two step, i.e., a pre-handoff decision algorithm followed by a handoff decision algorithm is presented in [13]. An AHP based comparison of different vertical handoff algorithms is presented in [14]. An AHP based multi-criteria decision making (MCDM) tool is designed for vertical handover among WLAN, UMTS and GPRS and is presented in [15]. Measurements of application perceived throughput in DAB, GPRS, UMTS and WLAN is done for automatic network selection in [16]. A 3-step network selection strategy for new cell arrival in a road condition is shown in [17]. AHP and grey relational analysis (GRA) based network selection mechanism for UMTS and WLAN is presented in [18] for next generation networks. A framework for both horizontal and vertical handover in wireless network based on mobile IPv6 has been proposed in [19]. Handoff triggering and network selection algorithm in CDMA-WLAN integrated networks is proposed and QoS performance against velocity of mobile terminal is discussed in [20]. A load balancing vertical handover algorithm, which will maximize the collective battery lifetime of Mobile Nodes, is proposed in [21]. After a thorough review of the literature, we conclude with the following criticisms: 1) The traditional vertical handover algorithms consider constraint resources, but it does not provide the optimal solution for ABC communication for ITS in the context of multiple constraints. 2) The mathematical formulation of vertical handover is rigorous and need longer execution time. 3) Selection of best network against vehicular speed variations is not shown in the “traditional vertical handover models” except in [3]. 4) Sensitivity analysis, in the field of vertical handover, was not done in any of these research papers except [39]. The novelty of our study is to address the challenge of vertical handover in road transportation where vehicle speed is the major influencing factor. For an appropriate vertical handover the decision makers must define the decision hierarchy, define criteria weights, and evaluate radio access technologies against the criteria. Once the priorities for each network alternatives have been determined, they are incorporated in the network optimization process of the methodology where they represent a network priority. The ability to conduct sensitivity analysis is designed into the methodology, providing the decision maker with additional insight regarding the robustness of a model. In the present work, we propose a model that analyzes the case in which a vehicle is kept at different positions of a highway, and the network ranking optimizes the “best” network in multiple constraint environments. This evaluation technique requires knowledge of vehicular speed, received signal strength (RSS), type of application (bandwidth requirements), network traffic load, usage cost of service and initial delay for connection establishment. The rest of this paper is organized as follows: problem definition is given in section 2, proposed solution to this problem is presented in section 3, analytic hierarchy process is explained in section 4, algorithm design is presented in section 5, simulation is discussed in section 6, result analysis is done in section 7, and sensitivity analysis is done in section 8. Finally conclusions are drawn in section 9. 2. Problem Definition Unlike horizontal handoff, Vertical handoff does not only take place in cell edges. There are many other factors that should be considered for a proper vertical handoff. Authors have taken an initiative to develop a cognitive radio [22] based embedded system which will perform both remote sensing and communication for ITS. Such a system based on Software defines radio (SDR) is already presented in [23]. Adaptive algorithms are required to upgrade this SDR based system to a Cognitive Radio based system. The aim of this work is to design and simulate an application specific vertical handover among WiMAX, UMTS and WLAN for ITS using Analytic Hierarchy Process (AHP) [24]-[34]. We are taking the following criteria into consideration while selecting the technology on entering a cell having two or more of the above radio access networks:
ISSN: 0975-5462
5510
Sourav Dhar et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5509-5525 Triggering factor Critical factor Influencing factors
Received Signal Strength (RSS) (Availability of the network) Terminal Mobility (Speed of the vehicle) Bandwidth (depends on Application type), Initial Delay, Network Traffic Load, Usage cost
Suppose the user is moving from a cell to another as shown in Figure 1. In typical highway traffic, we have considered five different situations. At position A, the vehicle has passed the toll plaza and is about to enter in a city. Thus all three radio networks are available here. Point B is almost similar to point A; here the vehicle is in a suburban area. Point-C is far from city and only UMTS network is available here.
Figure 1: A Typical Radio Network Availability Scenario for ITS [39]
Point-D is near to toll plaza, and UMTS along with dedicated short range communication (DSRC) network is available here. Currently, the IEEE standard proposed for DSRC, known as 802.11p, is based upon the IEEE 802.11a standard [37]. The physical Layer of the 802.11p standard is the same as the physical Layer of the 802.11a standard, except for the used sample rate. The main application of 802.11p is vehicle-to-vehicle (V2V) communication. V2V is also a synonym for dedicated short range communication (DSRC) [36]. DSRC is already in use in USA, Europe and Japan for electronic toll collection [23]. Point- E indicates the suburban area where WLAN network is not available and the vehicle is in the edge of the WiMAX coverage area. Handover needs to take place when either RSS is critical or QoS is poor for current radio access network The applications taken into consideration are Conversational, video streaming and normal file transfer (data streaming) [14]. Each of these applications requires a different bandwidth. All the three technologies do not support these applications with equal QoS. Each wireless technology has a limit on the maximum amount of mobility it can support. As we are considering a transportation system, speed of the vehicle (mobility) is an important factor that is to be taken into consideration. WiMAX supports the highest mobility among the three and WLAN has the lowest support for mobility and the least coverage area [23]. Another parameter taken into consideration is the Initial delay, which is the setup time for a connection. Initial delay is more in some systems compared to others. According to Fiedler et al. [35], delay could be up to seven seconds for UMTS. WLAN connectivity, on the other hand, is perceived as responding instantaneously. WiMAX response could be faster than UMTS but slower than WLAN. 3.
Assumptions and Proposed Solution The following assumptions are made for solving the problem defined in section 2. i)
Situations of point “A” and point “B” of figure 1, where all the three networks are available, are considered for simulation. At point “C” only UMTS is available. At point “D” UMTS and WLAN are available and at
ISSN: 0975-5462
5511
Sourav Dhar et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5509-5525 point “E” WiMAX and UMTS are available networks. However, this handover scheme will be valid for any point on the highway. ii) All the radio access networks will be equally loaded at a given instant. Random network traffic load distribution will be considered in future works. iii) Two types of applications are considered, viz., conversational/ voice communication and streaming applications (audio, video streaming, internet access, data transfer etc.). iv) Speed of the vehicle is an objective factor (which can be compared on the basis of measured value) and other influencing factors are subjective factors.
Figure 2: Functional Architecture for Vertical handover
Figure 2 gives our proposed vertical handover functional architecture containing the following given modules. The context repository module basically a database which collects all the contextual information, through monitoring and measurements, required to identify the need for handover and to apply handover decision. These data are monitored periodically and updated accordingly. All these data are supplied to the adaptability manager. Network context repository: This holds all the network related contexts. Static parameters like offered bandwidth, initial delay for connection establishment and dynamic data like network availability (RSS (Received Signal Strength) and current network traffic load. Terminal context repository: This holds the up to date speed of the vehicle, battery power, location information, etc. User context repository: This accepts the inputs from the user. User preferences are basically the type of application (data rate/ bandwidth required) and affordable cost for the service. Service context repository: This stores mainly the static information like monetary cost of service, service capabilities etc. The Adaptability manager module is responsible for providing transparent switching between networks. So, it encloses the main phases of a handover process: handover initiation, Handover decision (i.e., network selection) and Handover execution. Working of the adaptability manager is explained in the flowchart (figure 3). Handover initiation is a continuous process of RSS and QoS measurement. If either of these two is found critical then spectrum sensing will be started. If any other network is found available (network availability detector is
ISSN: 0975-5462
5512
Sourav Dhar et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5509-5525 responsible for this work.) then handover process will be initiated else the wireless device will continue the communication through the current radio access network. Once the handover process is initiated, the gathered context information in context repository will be supplied to the adaptability manager. The optimum network will be selected based on the AHP evaluation technique. In handover execution phase, the adaptability manager will first check whether the selected network is different from the current network. If so, then it will issue a vertical handover command and will direct the control unit of SDR to reconfigure the hardware according to the selected network. 4. Analytic Hierarchy Process The Analytic Hierarchy Process (AHP) is a technique used for dealing with problems which involve the consideration of multiple criteria simultaneously. It is unique in its ability to deal with intangible attributes and to monitor the consistency with which a decision maker makes his decisions AHP, developed by Saaty [27-32], addresses how to determine the relative importance of a set of activities in a multi-criteria decision problem. The process makes it possible to incorporate judgments on intangible qualitative criteria alongside tangible quantitative criteria. The AHP method is based on three principles: i) Structure of the model; ii) Comparative judgment of the alternatives and the criteria; iii) Synthesis of the priorities. In the literature, AHP has been widely used in solving many complicated decision-making problems [24-34, 38]. The steps of AHP are as follows: Step-I: A complex decision problem is structured as a hierarchy (Figure 4). AHP initially breaks down a complex multi-criteria decision-making problem into a hierarchy of interrelated decision criteria, decision alternatives. With the AHP, the objectives, criteria and alternatives are arranged in a hierarchical structure similar to a family tree. A hierarchy has at least three levels: overall goal of the problem at the top, multiple criteria that define alternatives in the middle, and decision alternatives at the bottom. Step-II: The comparison of the alternatives and the criteria. Once the problem has been decomposed and the hierarchy is constructed, prioritization procedure starts in order to determine the relative importance of the criteria within each level. The pairwise judgment starts from the second level and finishes in the lowest level, alternatives. In each level, the criteria are compared pairwise according to their levels of influence and based on the specified criteria in the higher level. In AHP, multiple pairwise comparisons are based on a standardized comparison scale of nine levels (Table 1). Let C = {Cj |j = 1, 2,...., n} be the set of criteria. The result of the pairwise comparison on ‘n’ criteria can be summarized in an (n x n) evaluation matrix ‘A’ in which every element ‘aij’ (i,j = 1,2,..., n) is the quotient of weights of the criteria, as shown:
a11 a12 a 21 a22 A an1
ISSN: 0975-5462
a1n a2 n a33 where aij 1, for , i j, and ann
aij 1
a ji
for aij 0 (1)
5513
Sourav Dhar et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5509-5525
No
No
Is the RSS of the selected Network critical?
Is QoS Critical?
Continue with the selected network
Yes Yes
Start Spectrum sensing
No
Any other network is available?
Handover Initiation
Yes Feed the handover information to the handover decision block
Optimum Network is selected from the available networks using AHP evaluation technique.
Is the selected network different from the current network?
Handover Decision
No Handover Execution
Yes Execute Vertical Handover Issue hardware reconfiguration command to SDR
SDR
Figure 3: Working of the adaptability manager
ISSN: 0975-5462
5514
Sourav Dhar et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5509-5525 Step III: Objective factors and the subjective factors are to be identified. Objective factors are those which could be measured directly. Thus pair wise comparison can be achieved directly taking the ratio of ith factor measured value (scaled between 1 to 9) to jth factor measured value (scaled between 1 to 9). Thus equation (1) could be rewritten for objective factors as follows:
m1 / m1 m1 / m2 m / m m / m 2 2 2 1 A mn / m1 mn / m2
m1 / mn m2 / m3 m2 / mn m3 / m3 (2) mn / mn
Where mk= Measured value (scaled between 1 to 9) of the objective factor for kth alternative. It is observed that for the matrix given above the following relation holds [50]: Aw = nw Where, w is the priority vector and n is the number of elements being compared. This is the case for a perfectly consistent comparison matrix whose elements satisfy aij = aikaij for all ( i , j ,k) Subjective factors are those which can be quantified from human knowledge. Since the comparisons are made subjectively based on human knowledge, the consistency of judgment needs to be verified through evaluating the consistency ratio (CR). Table 1: The Nine -Point Scale Of Pair-Wise Comparison [27] Weights 1 3 5 7 9 2,4,6,8 Reciprocals
Verbal Scale Equal Importance of both elements. Moderate importance of one element over another. Strong importance of one element over another. Very strong importance of one element over another. Extreme importance of one element over another. Intermediate values
Explanation Two elements contribute equally. Experience and judgment favor one element over another. An element is strongly favored. An element is strongly dominant.
An element is favored by at least an order of magnitude. Used to compromise between two judgments. Reflecting dominance of second alternative compared with the first.
Step-IV: Develop a normalized matrix (A’) by dividing each element in a column of the pair-wise comparison matrix A by its column sum. The sum of every column of a correctly normalized matrix will always be 1.
a11 a i1 n a 21 a i1 n A' a n1 a i1 n
ISSN: 0975-5462
a12 ai 2
a 22 ai 2
a 33 ai3
n
n
n
a1 n n a in a2n n a in a nn n a in
(3)
5515
Sourav Dhar et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5509-5525 Step-V: Average each row of the normalized matrix. These row averages form the priority vector (P) of alternative preferences with respect to the particular criteria. These values will also add up to 1.
p1 p P 2 where, pn
pk Avg (k th row of A' ) (4)
Step VI: Now find the Eigen vector (Λ) as,
1 2 Λ= (A x P)/ P = --------- (5) n The largest Eigen value (λmax) can be found as,
max
1 2 n n
(6)
Figure 4: Decision hierarchy
Step-VII: To determine whether or not a level of consistency is reasonable, we need to develop a quantifiable value for the comparison matrix. If w is the column vector of the relative weight wi (i = 1, 2, 3 . . .n), comparison matrix A is consistent if Aw = nw. Step-VIII: Using AHP, compute the consistency ratio as CR = CI/RI where CI =consistency index of comparison matrix = (λmax − n)/ (n − 1) and RI = Random inconsistency = 1.98(n-2)/n (listed in Table 2).
ISSN: 0975-5462
5516
Sourav Dhar et. al. / International Journal of Engineering Science and Technology Vol. 2(10), 2010, 5509-5525 Table 2: Random inconsistency for different set of alternatives Number of alternatives (n) 3 4 5 6 7 8
Random inconsistency (RI) 0.58 0.90 1.12 1.24 1.32 1.41
Step-IX: AHP does tolerate some inconsistency and according to Saaty [26-31], if the consistency ratio is
