Connectivity Probability of Wireless Ad Hoc Networks: Denition, Evaluation, Comparison

Tatiana K. Madsen, Frank H. P. Fitzek and Ramjee Prasad

Center for TeleInFrastructure, Deptartment of Communication Technology, Aalborg University, Denmark Email: [tatiana | | prasad]@kom.aau.dk

Gerrit Schulte
Technical University of Berlin, Berlin, Germany Email: gerrit@schulte.cc Abstract. The paper presents a new approach investigating mobile ad hoc network connectivity. It is shown how to dene and evaluate the connectivity probability of a mobile network where the position of the nodes and the link quality changes over time. The connectivity probability is a measure that can capture the impact of the node movement on the network connectivity. A number of mobility models is considered ranging from the classical Random Direction model to the Virtual World model based on the mobility measurements of a multiplayer game. We introduce an Attractor model as a simple way to model nonhomogeneous node distribution by incorporating viscosity regions in the simulation area. Methods of ergodic theory are used to show the correctness of the approach and to reduce the computational time. Simulation results show how the node density distribution aects the network connectivity. Keywords: ad hoc network, mobility models, simulations, connectivity probability

1. Introduction The movement pattern of users has a signicant impact on performance of mobile and wireless communication systems and networks. To incorporate user mobility in the simulation evaluation of e.g. routing protocols, a variety of models have been proposed (see [1, 2, 3] and references therein). The mobility models dene how the nodes move within the network. Trustworthy mobility models should satisfy some criteria. Models used for cellular systems should provide realistic values for such parameters as the cell residence time, number of handos during a call, etc. In ad hoc networks the emphasis is shifted and other issues play the key role. These issues are connectivity of a network [4], path duration (also called longevity of routes [5]) and stability of routes [3]. To achieve a connected ad hoc network, there must be a multihop path from each node to any other node. Observing a network at a given moment in time, we can determine if some of the nodes are isolated from other nodes or not. Assuming some initial node distribution in
c 2005 Kluwer Academic Publishers. Printed in the Netherlands.

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.1

Madsen, Fitzek, Schulte, Prasad

a simulation area, we can guarantee with some probability that the network is connected at that moment in time [4]. Now the question arises what will happen when the nodes start moving around. Some links will be broken and new links will be established. It is desirable to have an estimation whether or not the network will remain connected and it would be practically convenient to have a measure of a network being connected under mobility of nodes. To estimate a network connectivity in the presence of mobility, we introduce a measure, called connectivity probability. This parameter is dened as a measure of time intervals during which a network is connected. It gives the probability that observing a network in randomly chosen point in time we will nd it connected. Connectivity probability can capture the important properties of a network when position of nodes and states of the links changes over time. We dene a general framework for evaluating the impact of node movement on the network connectivity. To incorporate the nodes dynamics in the theoretical analysis of the connectivity properties of a network, we use some results of the theory of dynamical systems. Ergodic properties of mobility models based on deterministic modelling of node movement, allow us to derive a simple expression for the connectivity probability. For example, this can be done for Random Direction (RD) model and its extensions. A number of dierent mobility models are considered in the paper. We illustrate how the connectivity probability can be estimated with the examples of Random Waypoint [2], Attractor, Virtual world model [3], and Model with obstacles [14, 17]: The Random Waypoint (RW) model is widely used for protocol performance evaluation and is implemented in the network simulator ns2. We introduce the Attractor model as a simple enhancement of any mobility model when regions of slow movements are incorporated in the simulation area. This leads to nonhomogeneous node distribution. The Virtual World model is based on the mobility patterns obtained from the measurements of the virtual world scenarios. The mobility measurements of a multiplayer game, such as Quake II, can be used for investigation of the impact of mobility on the performance of multihop protocols. The mobility model with obstacles incorporates the rectangularshaped obstructions in the simulation region to make the simulation scenarios more realistic. It is especially actual when simulating

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.2

Connectivity Probability of Wireless Ad Hoc Networks

indoor scenarios where the inuence of walls and other objects can not be ignored. The remainder of the paper is organized as follows. Section 2 provides the motivation for our work and gives a short overview of the related work. Section 3 introduces a formal denition of the connectivity probability. Section 4 explains how the connectivity probability can be estimated for dierent mobility models. Section 5 presents simulation results. Section 6 oers some concluding remarks.

2. Motivation and related work A problem of connectivity of ad hoc networks has got a lot of attention during last years. For multihop networks the connectivity is one of the fundamental and important properties. The majority of research considers only the static situation, when the nodes do not move. For example, in [4] connectivity dependence on the node density and the transmission radius is evaluated in the static scenario. Another problem that was tackled is dening the critical transmission range, a threshold transmission range that is required for connectivity (see e.g. [6, 7]). Including the mobility of nodes complicates analysis of network characteristics, though many applications of ad hoc networks requires that the nodes are mobile but still connected. It has been realized that the mobility pattern of the nodes has a big impact on the performance of ad hoc protocols [2, 8]. But for the correct interpretation of the results, one should be aware of the properties of the network under consideration, including the connectivity property. Let us consider the situation when at the initial moment of time nodes are distributed uniformly on the simulation area and the network is connected. Then the nodes start moving to any of the corners of the area and after some time the network will become and remain partitioned. This example is an extreme case when an initially connected network becomes unconnected, but it illustrates the importance of understanding how mobility of the nodes inuences the connectivity. Under or overestimation of network connectivity can potentially lead to misinterpretation of the results, when falls in the protocol performance would be attributed to the specics of a protocol design and not to the characteristics of the network. The main tools used for investigation of network connectivity come from the graph theory and the probability theory. For example, the theory of geometric random graphs (GRG) deals with a set of n points distributed according to some density and focuses on the properties of

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.3

Madsen, Fitzek, Schulte, Prasad

the resulting node placement. Using GRG theory some conclusions can be made on the transmission range in a dense ad hoc network [9]. Other results for ad hoc network connectivity can be obtained making use of the percolation theory [10] (see also [11] for the connectivity of hybrid ad hoc networks). In [6] the problem of connectivity is approached from the probabilistic angle and the transmission range is dened as a random variable allowing statistical modelling of the connectivity. Probabilistic approach can be also used for the investigation of sparse ad hoc networks, namely to determine the critical transmission range which generates communication graphs that are connected with high probability [7]. Unfortunately, these instruments are only amiable for the analysis of static scenarios, since they fail to capture situations when edges of the connectivity graph changes over time. To incorporate the dynamics of a network topology into the analysis of the connectivity problem, we resort to the methods provided by the theory of dynamical systems. Dynamical systems are exactly the mathematical eld that is developed and used for description of evolution in time of objects of dierent nature. Therefore, it is also natural in our situation, for discussion of the qualitative estimation of the network connectivity under mobility of nodes, to use ideas and methods of the modern theory of dynamical systems.

3. Denition of Connectivity Probability We start with giving some insights in the problem of network connectivity in a static environment. Then we show how connectivity can be dened when the nodes are mobile and introduce connectivity probability as a measure of network connectivity in a dynamic environment. Let D be a bounded domain on the Euclidean plane R2 = {x, y} with piecewise smooth borders. Inside domain D there are n nodes. In the initial moment t = 0 nodes are somehow placed and then they start moving. Let ri = (xi , yi ) be the radiusvector of node i. Further, we assume that every node has a transceiver with communication range r: if the distance between two nodes is larger than r, they cannot establish a direct communication link. Nodes can transmit information by using multihop connections. Denition 1. A network is connected (or fully connected ) if for every pair of nodes there exists a path between them. Note that for a network to be connected we require existence of a path from source to destination at any moment in time. As it is

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.4

Connectivity Probability of Wireless Ad Hoc Networks

shown in [12] it might not be necessary, but it imposes design of specic routing protocols. We do not include this special case in the scope of our consideration. Connectivity probability is used to quantify the network connectivity. In the static environment (when the positions of nodes are xed), the probability for a network to be connected depends on the density of the nodes and their transmission range. In a typical static simulation scenario, in order to evaluate the connectivity probability a number of nodes are placed at random in the simulation area. A random variable is introduced which is equal to one if the network is found connected and zero if it is disconnected. The average of this random variable over several trials gives the connectivity probability. Now we will show how to extend this approach to the case of dynamic environment. Under motion of nodes a semiaxis of time R+ is divided into intervals
1 , 2 , 3 , . . . + where k (k ) denotes a time interval during which the network is connected (disconnected). Lets introduce the function f+ (t) such that + f+ (t) = 1 if t k , and f+ (t) = 0 if t k .Time intervals k can be considered as randomly distributed; thus, f+ (t) is a stochastic process. Denition 2. In the dynamic environment, connectivity probability P+ is dened by P+ = E[f+ (t)] (1)

where E[] stands for the expected value (if the expected value exists). Analogously the function f (t) can be introduced: f (t) =
1, t k + 0, t k

The probability that a network is disconnected is given by P = E[f (t)]. Since f+ (t) = 1 f (t), then P+ + P = 1. One can see that in general case the probability P+ is time-dependent: P+ = P+ (t). For stationary stochastic processes P+ (t) = const. If stationary process is ergodic, then equality (1) can be substituted by the following: with probability one 1 lim
0

f+ (t)dt = P+

This equality is equivalent to the following: P+ = lim

1 mes(T +

[0, ])

(2)

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.5

Madsen, Fitzek, Schulte, Prasad

+ where T + = k 1 . That is, the probability that a network is connected can be dened as a density of set T + on the semiaxis R+ = {t > 0}. Therefore, the problem of connectivity of a network consisting of moving nodes is reduced to the problem of existence and estimation of the expected value (1). If the mobility model is stationary and ergodic, then formula (2) can be used for estimation of connectivity. Detailed analysis of the estimation of the connectivity probability is presented in the next section.

4. Analysis 4.1. Note on dynamical systems and stochastic processes Dynamical systems theory can be useful for modelling the movement of nodes in an ad hoc network. In homogeneous networks, where the properties and capabilities of the nodes are the same, we can reasonably assume that the movement of a single node can be described by one and the same system of dierential equations. If we introduce some form of randomization in the movement pattern of a node, then it would require construction of stochastic dierential processes. If the considered process is stationary, a system of autonomous dierential equations can be used, where the rightsides of equations do not depend explicitly on time and dierent nodes dier only by their initial conditions. In the theory of dynamical systems a phase ow is introduced that is a group of shifts along the trajectories during one and the same time interval. The phase ow generates a dynamical system. The system can be described by dierential equations of the following form: x = g(x), x (3)

where is the phase space, x is a set of coordinates in (usually it is position and velocity), dot means time dierentiation. Let n be a number of nodes and x(1) , . . . , x(n) are their phase coordinates. Then these coordinates satisfy the following dierential equations: x(k) = g(x(k) ), k = 1, . . . , n (4)

In this way dynamics of n nodes is completely dened by dynamical system (4), that is a direct product of n copies of original dynamical sys tem (3). Its phase space = . . . = ()n is a direct product of n copies of initial phase space, and phase coordinates x = (x(1) , . . . , x(n) )
mes stands for measure of a domain. In this case it is the total length of the intervals T + [0, ]
1

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.6

Connectivity Probability of Wireless Ad Hoc Networks

are a set of coordinates of individual nodes. Note that if the system (3) has an invariant measure in , then the system (4) also has an invariant measure in , that is a direct product = 1 n . In the problem of the nodes connectivity, the phase space can be divided into two domains D and D = \ D in the following way: when x D, then every node from n nodes x1 , . . . , xn can communicate with any other node; when x D, then some of the nodes can not reach some of the others. Following dynamical systems approach, the connectivity probability can be estimated as a fraction of time t when x(t) D. The problem of estimation of the connectivity measure can be sim plied signicantly if dynamical system (4) is ergodic in . By denition, the system is ergodic if a measure of any invariant subdomain of the phase space is either equal to zero or is equal to the measure of the whole space. The key result of the ergodic theory is Birkho ergodic theorem [13]. Here we provide its formulation using our notations. Let f () be a x Then for almost all solutions of measurable integrable function on . the ergodic system (3) the following equality takes place: 1 T T lim where mes =
T 0

f ((t))dt = x

f ()d x , mes

(5)

d = ()

is the measure of the whole phase space. Let, for example, f be a characteristic function of a measurable domain D: f () = 1 if x D and f () = 0 if x D. Due to the x x the fact that f is limited and D is measurable, function f : R is integrable. In this case, the leftside of (5) is equal to the fraction of time interval 0 t T when x(t) lies in the domain D. Then the connectivity probability of a network will be equal to the rightside of (5), that in this case is mes D/mes (6) This approach can be also interpreted in terms of the theory of stochastic processes. One can introduce the probabilistic measure in the phase space . Probability for a system to be in the measurable domain D is determined by formula (6). Let f (x) be a characteristic function of the domain D introduced above and x(t) be the solution of system (3). Then the function f (x(t)) can be interpreted as a stochastic process. Its realization will be dependent on the choice of the initial conditions. Let E[f ( )] be the expected value of the function f (t) at

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.7

Madsen, Fitzek, Schulte, Prasad

the moment . If the rightsides of the equation (3) do not depend on time, then the stochastic process is stationary. In particular, it means that E[f ( )] does not depend on . If, in addition to being stationary, the system is ergodic, the expected value can be calculated by using formula (6): 1 T T lim
T 0

f ((t))dt = E[f ] = mes D/mes x

This is true for almost all realization of the stochastic process. Therefore, the problem of calculation of the expected value (1) is reduced to the geometrical problem of calculation of domain volumes in a phase space if the process is ergodic. As we show in the following, using formula (6) is computationally more ecient than evaluating the expected value. Thus, it should be preferred when possible, i.e. when the system is ergodic. At this point a natural question comes: can we expect that one or another widelyused mobility model is ergodic? The problem of ergodicity has been intensively studied, but it is far from being completely resolved due to its complexity. There exist a number of wellknown examples of ergodic systems from the theory of dynamical systems. For example, consider a system of nodes within a rectangular domain moving along the straight lines. Once the boundary is reached, a new direction of movement is chosen such that the trajectory forms equal angles with the boundary. One can imagine a trajectory as a stroke of light reected by a mirror. These kind of systems are called billiards. One can easily notice that they correspond to a particular case of Random Direction mobility model (in RD model a new direction is chosen uniformly towards the inside of the simulation area). Billiards, restricted on the integral manyfold, are ergodic. Their ergodicity property can be established with the help of Weyls theorem on homogeneous distribution (see e.g. [13]). One should note, though, that not every statinary process is ergodic and in realistic scenarios ergodicity is more the exception than the rule. If the rightside of (3) explicitly depends on time, then the process f (t) = f (x(t)) is nonstationary. In general, formula (6) is not applicable and the expected value depends on time. However, it turns out that under a large range of conditions, the expected value of f (t) calculated over a long interval of time gives suciently good approximation of the expected value E[f (t)] over semiaxis 0 t < . Thus, if E[f (t)] changes slowly over time, then this value can be estimated as the average value of f (t) over long (but not innitely-long) time interval [t, t + ]: with probability one 1 T T lim
T 0

f (t)dt = lim

1 T T

T 0

E[f (t)]dt

(7)

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.8

Connectivity Probability of Wireless Ad Hoc Networks

If we consider time interval where f (t) oscillates but E[f (t)] changes slowly, then E can be approximated as a constant and the rightside of (7) is equal to E(t). Thus, for slow stochastic processes formula (7) can be used for estimation of the connectivity probability. 4.2. Estimation of connectivity probability for different mobility models In [14] we present a detailed analysis and derivation of formula (6) for Random Direction model. In this case mes is just the area of a simulation domain: mes = ab where a and b are lengths of the sides if rectangular region is considered. mes D is area of a domain consisting of all possible positions of the nodes such that the network is connected. It seems to be dicult to nd a close form expression for the measure of D as a function of communication radius r. Instead, MonteCarlo simulations can be used: nodes are placed randomly, independently and uniformly in a simulation area and the connectivity of the network is veried. Then P+ M/N , where N is a total number of simulation trials and M is a number of successful trials (when the network is connected) if N is chosen suciently large. Formula (6) can be used for the range of other mobility models that generalize RD model: when speeds of nodes are dierent or when borders of the area consist of horizontal and vertical lines, or speaking more generally, when rectangularshaped obstacles are placed in a simulation area. The rst case is illustrated by an example of Attractor model in Section 5.3. The realization of the second case by the model with obstacles can be found in [15, 16]. It is discussed in Section 5.5. Using formula (6) for estimation of the connectivity probability, any xed moment of time can be considered, since the process is stationary. In principle, t = 0 is as good choice as any other, and by using t = 0 the complexity of the simulation is of the same order as for the static case evaluation. However, in practise, it is recommendable to consider t = t0 > 0 to make the result less correlated with the initial conditions. Considering other stationary mobility models, it is more dicult to answer if the limit (2) exists. If a collection of mobile nodes comprises a dynamical system with some particular properties (namely, a system has an invariant measure), then for almost all initial conditions the limit exists. It can happen that for dierent initial conditions (i.e. dierent initial positions and velocities of the nodes) the limiting values will be dierent. In this situation averaging over initial node distribution should be done, and the obtained average connectivity probability can be used as an estimate that a network is connected in any point in time.

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.9

10

Madsen, Fitzek, Schulte, Prasad

Table I. Calculation approaches for dierent models RD Formula (6) RW Formula (1) Attractor Formula (6) VW Formula (7) Obst Formula (6)

For mobility models based on RD pattern, we have two ways to calculate the connectivity probability: either evaluating the expected value (1) or by applying formula (6). Comparing these two approaches, we have done some measurements of cpu time. To obtain connectivity probability for one set of parameters, using formula (6) it takes approx. 132 sec, whereas if we calculate directly by using (1), cpu time is more than 2000 sec. The big dierence in cpu times can be explained by the fact that in the second case modelling in time is required and at each time stamp we have to recalculate the positions of the nodes. Moreover, using MonteCarlo approach, we can estimate how many simulation runs are required in order for results to lie within a specied condence interval. Applying direct calculation, it is not possible by any analytical means to estimate how long in time we have to perform the simulation, it can dier in each particular case. Therefore, we have chosen to calculate connectivity by using formula (6) when possible, since it is computationally less expensive. To summarize this section, Table I is given showing what approach is used for estimation of the connectivity probability for the results in Section 5.

5. Simulation results Dierent application scenarios of ad hoc networks would require different connectivity probability. For example, in a network used for safetycritical or lifecritical operations, the connectivity probability should be more than 90%. If temporary network disconnections can be tolerated, then the connectivity probability can go down to 50% or even lower. 5.1. Random Direction model In Random Direction model, nodes select a direction in which to move, uniformly between 0 and 2, and speed from a given distribution. Subsequently motion starts. Once the simulation boundary is reached, a node pauses for a specied period of time, then it chooses a new

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.10

Connectivity Probability of Wireless Ad Hoc Networks

11

direction uniformly toward the inside of the simulation domain and the process continues. Especial interest has a case when a new direction is chosen such that the trajectory of movement forms equal angles with the boundary. We refer to it as Billiard model. One of the important characteristics of the Billiard model is a uniform node density distribution. This can be shown directly [17] or by using the following wellknown fact from the theory of dynamical systems: in a billiard system for almost every initial conditions a trajectory of a node covers a region of motion everywhere dense.
1

0.8 Connectivity probability

0.6

0.4

5 nodes 10 nodes 50 nodes

0.2

0 0 20 40 60 Communication radius 80 100

Figure 1. Connectivity probability as a function of communication radius. Simulation area 100x100 m2

Figure 1 presents the connectivity probability dependence on a communication radius where the simulation area is a square of 100x100 m2 and the number of nodes is equal to 5, 10 and 50. When the density of nodes is high, in order to have a connected network with probability 95% at any moment of time, we should require the communication radius of 20 m. One can observe that for the dense network the line presenting connectivity probability is steeper than the one for the sparse network. One should note that in the case when all nodes move with the same speed, the connectivity probability would not depend on the actual value of the speed. 5.2. Random Waypoint model Another widely used model for protocol performance evaluation is the Random Waypoint model. A node chooses a random destination and a travelling speed and travel towards the destination along a straight line. Upon arrival, it pauses and repeats the process. Table II shows connectivity probability for the following set of parameters: simulation area is 150x150 m2 , number of nodes n is 100

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.11

12

Madsen, Fitzek, Schulte, Prasad

Table II. Connectivity Probability for Billiard, Random Waypoint and Attractor models B r = 30, n = 100 r = 40, n = 80 0.92 0.017 0.99 0.006 RW 0.85 0.007 0.97 0.03 Att 0.55 0.003 0.90 0.018

and communication radius r is 30 m, and n = 80 and r = 40 m. For RW model the results are obtained as the average of 50 simulation runs. For Billiard model 1000 trials are performed. The 95% condence interval is given in Table II. Comparing results for Billiard and Random Waypoint, one can observe that the connectivity probability for RW model is lower, though the same set of parameters were considered. This result can be explained if we consider the node distribution in the simulation area for these two models. As it was already noticed (see e.g. [18]), the node density distribution for RW is not uniform, but has a convex shape. Figure 2 shows a histogram obtained through simulation on a 150x150 m2 area with 80 nodes. The area is divided into 15x15 subareas and the number of nodes in the subarea at every time step is counted. A higher node density in the middle of the area and a signicantly lower density at the area edges is due to the fact that nodes at the borders are more likely to move back to the middle by randomly choosing a new direction of movement. The node density has a direct impact on the connectivity probability. The existence of areas with low density leads to the decrease in the connectivity, as we can see from Table II. If a sparse network is considered with low number of nodes, then the opposite eect can be observed. Table III shows results when the number of nodes is equal to 13. The probability to nd a network fully connected under Random Waypoint is higher than for Random Direction model. In this case the density of nodes is so low even in the high density region, that the number of nodes might not be enough to form a connected network. Therefore, here the fact that for RW model in the central part of the simulation area the node density is higher plays crucial role. 5.3. Attractor model The considered above models result in almost uniform spatial node distribution. Tough we can imagine situations when mobile users tend to gather in one subarea (shopping mall or trac jam due to a road construction). These scenarios can not be captured by group mobility

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.12

Connectivity Probability of Wireless Ad Hoc Networks

13

0.0036 0.0034 0.0032 0.003 0.0028 0.0026 0.0024 0.0022 0.002 0.0018 12 14 16

6 X/10, m

10

12

14

16 0

10

Y/10, m

Figure 2. Node density distribution for Random Waypoint model

models when a group of users exhibit similar behavior. To reect individual trajectories of users movement that willingly or due to some circumstances spend some time in one subarea, we introduce a mobility model called Attractor model. The main feature of this model is the presence of gravity points or areas that attract mobile users. One can think of many ways how attractors can be simulated. For example, one can specify points of attraction and how the users move from one point to another by borrowing methods of uid dynamics. Here we propose another, very simple way introducing areas of viscosity. We can take any movement pattern as underlying mobility model and enhance it with regions of slow movement.

0.12 0.1 0.08 0.06 0.04 0.02 0 12 14 16

6 X/10, m

10

12

14

16 0

10

Y/10, m

Figure 3. Node density distribution for Attractor model

As an example, we consider 80 nodes moving in the 150x150 m2 area according to Random Direction model with v = 5m/s. Two attractors

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.13

14

Madsen, Fitzek, Schulte, Prasad

are introduced: one viscosity region is a disk with center in (50, 50) and radius 30 m, the second is a disk with center located in (100,100) and radius 20 m. When coming to the rst attractor, the users slow down to 1 m/s and in the second region the speed of movement is 0.5 m/s. This results in nonuniform density distribution (see Figure 3). High density corresponds to the attractor areas. One can observe that the node density is inverse proportional to the speed of movement: density in the second attractor is approx. 0.1 nodes/m2 , in the rst 0.048 nodes/m2 , in the rest of the area 0.001 nodes/m2 .
1 0.9 0.8 Connectivity Probability 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 1000 2000 3000 4000 time 5000 6000 7000 r = 85 m r = 70 m r = 50 m

Figure 4. Connectivity probability for Virtual world model

0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 12 14 16

6 X/10, m

10

12

14 0

10

Y/10, m

Figure 5. Node density distribution for Virtual World model

As expected, the probability to have a fully connected network for Attractor model is signicantly lower than for Billiard and Random Waypoint models if the same number of nodes are considered (Table II). For Attractor model nodes tend to spend more time in the viscosity

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.14

Connectivity Probability of Wireless Ad Hoc Networks

15

Table III. Connectivity Probability for Quake, Billiard and Random Waypoint models Quake r = 85, n = 13 r = 70, n = 13 r = 50, n = 13 0.94 0.8 0.15 B 0.98 0.008 0.84 0.022 0.21 0.0025 RW 0.99 0.019 0.97 0.033 0.61 0.095

regions where nodes are located close enough to be able to communicate directly or where there are enough nodes to support multihop connections. In the rest of the area the node density will be low; this explains why the connectivity probability is low. 5.4. Virtual world model Presented above models are generally used for simulations because they provide an easy way to simulate movements of nodes. It is a common concern among researchers in the eld of selforganizing networks how realistic these models are and if the results obtained with the help of these models reect the reallife situation. From the other hand, obtaining realistic measurements of mobility is complex and expensive. To create a more realistic model, multiplayer games can be used since they can measure the mobility of the players in the games virtual world [3]. The software for multiplayer games is no longer limited to games, but it is also used to emulate the real world. Therefore, it is possible to create new virtual world and adopt them to specic needs. Following the approach presented in [3], we use virtual world scenarios to explore mobility issues. To obtain mobility patterns, a multi player game, Quake II, is used. The traces of 13 players moving in the square area were recorded. The position of each player was logged every time the server updated a frame, that is approx. every 100 ms. The data was collected for about 12 min (for more details see [3]). For postprocessing the data, the area of players movement, 4000 x 4000 units of the map editor, was scaled to 150x150 m2 . Using the measurement data, we can try to estimate the connectivity probability in this case. Since nothing can be said about stationarity of the obtained process in advance, we have chosen to apply formula (7) and observe if the limiting value exists. Presented analysis is made based on one record of the movement pattern. We plot the expression of the rightside of (7) assuming dierent communication radius

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.15

16

Madsen, Fitzek, Schulte, Prasad

(Figure 4). One can observe that the limiting value exists (curves asymptotically approaches constant lines). The diculty of using virtual traces is that limited material is at dispose for the analysis. We can not predict how the users will move in the future, at the times when we stop recording traces. The conclusion to be made here is if the users would have the similar type of behavior that the one observed, then at any point in time we would nd the network fully connected with the probability 80% if communication radius is 70 m. The connectivity probabilities of Quake and Billiard models (Table III ) are close, but for the Billiard it is slightly higher. Though the density histograms looks quite dierent for these two models. In Quake the players were given the mission to explore the area. In principle, it could mean that a user would try to visit every subarea at least once, and if we would observe the users behavior for sucient long interval of time and record the density distribution, we could expect it to be close to uniform. In reality, the density graph (Figure 5) diers from RD model. The area along the right wall was almost unvisited, whereas in the rest of the territory small density picks are almost uniformly spread. Some player were more lazy than others and they preferred to spend more time in one place. This and may be the presence of some interesting objects contributed to the nonuniform distribution. 5.5. Mobility models with obstacles As an example of a mobility model with obstacles we consider a situation pictured in Figure 6. Six users move in a room 30m by 20 m is divided by a 20 meter wall (e.g. a museum or exhibition hall). For some positions of the nodes the wall is blocking the communication and no line-of-sight (LOS) path exists.

Figure 6. Simulation area with an obstacle

Line C (Figure 7) is plotted under the assumption that when a signal is propagated between a pair of nodes and there is an obstacle obstructing the direct transmission path, the signal is completely blocked by

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.16

Connectivity Probability of Wireless Ad Hoc Networks

17

the obstacle. We can observe that the probability of having a connected network is not higher than 0.9 regardless of the communication range/ transmission power of the device. This is not very realistic. For comparison the connectivity probability is shown when no wall is present (line A). This is equivalent to the situation when the wall is made of such light material that it does not seriously obscure the communication. To make the model more realistic, some assumptions should be made about the propagation characteristics. The signal strength decay depends on materials of walls and ceiling, size of the room and the size and form of other objects. A proper number should be chosen in each particular case. Line B is constructed under the assumption that in non-LOS case the signal decay is 20 dB. From Figure 7 one can see that in contrast with the line C, for the line B there exist parameters of the system where the connectivity probability is almost one.
1 0.9 0.8 0.7 Connectivity Probability 0.6 0.5 0.4 0.3 0.2 0.1 0 0 line A line B line C 5 10 15 20 25 Communication radius 30 35 40

Figure 7. Connectivity probability as a function of communication radios. Number of nodes = 6. Line A: LOS always exists, Line B: signal is decayed in NLOS cases, Line C: signal is completely blocked in NLOS cases

6. Conclusion We present a new method of describing the connectivity of an ad hoc network under mobility of nodes. This method is applicable to both dense and sparse networks. We show that some tools of the theory of dynamical systems can be used to conduct the analysis of the connectivity property and to speed up the connectivity evaluation when the underlying mobility model possesses ergodic properties. The results of the paper consist of a number of theoretical insights and proofs, which also can be of interest to researchers in the area

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.17

18

Madsen, Fitzek, Schulte, Prasad

of ad hoc networking, since they highlight the important issue of mobile ad hoc network connectivity. Extensive simulation results show evaluation of network connectivity for a range of models. The connectivity probability can be also estimated if some measurement data is available, e.g. traces of players in Quake game. The relationship between the connectivity probability and the node density distribution is underlined. In this study the focus was mainly on the stationary mobility models. It would be interesting to observe how the connectivity probability depends on time for nonstationary processes. A nonstationary model can be created by e.g. moving the position of the gravity points in the Attractor model or changing continuously the lengths of the rectangular sides of the simulation area in RD model. This topic is a matter of further investigation.

Acknowledgements The authors would like to thank L. Badia and T. Henderson for their help in collecting the measurement data of the multi player game. Finally, we would like to thank the anonymous reviewer for insightly and detailed comments that have helped to improve the quality of the paper.

References
1. C. Bettstetter. Smooth is Better than Sharp: A Random Mobility Model for Simulation of Wireless Networks. In Proceedings of the 4th ACM International Workshop on Modelling, Analysis, and Simulation of Wireless and Mobile Systems, Italy, July 2001. T. Camp, J. Boleng and V. Davies. A Survey of Mobility Models for Ad hoc Network Research. WCMC: Special Issue on Mobile Ad Hoc Networking, 2(5): 483-502, 2002. F. Fitzek, L. Badia, M. Zorzi, G. Schulte, P. Seeling, T. Henderson, Mobility and Stability Evaluation in Wireless Multi-Hop Networks Using Multi-Player Games. In In Proceedings of NETGAMES03, May 2003. C. Bettstetter. On the minimum node degree and connectivity of a wireless multihop network. In Proceedings of the 3rd ACM international symposium on Mobile ad hoc networking & computing, June 2002. D. Turgut, S.K. Das, M. Chatterjee, Longevity of routes in mobile ad hoc networks, In Proceeding of IEEE 53rd Vehicular Technology Conference, May 2001. H. Koskinen. Statistical model describing connectivity in ad hoc networks. In Proceedings of WiOpt04, March 2004.

2.

3.

4.

5.

6.

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.18

Connectivity Probability of Wireless Ad Hoc Networks

19

7.

8.

9. 10.

11. 12.

13. 14.

15. 16.

17. 18.

P. Santi, M.D. Blough. The critical transmitting range for connectivity in sparse wireless ad hoc networks. IEEE Transactions on Mobile Computing, 2(1):25-39, January 2003. F. H. P. Fitzek, T. K. Madsen, R. Prasad, M. Katz. Impact of Node Mobility on the Protocol Design of Self-Organizing Networks. In Proceedings of Wireless World Research Forum 11, June 2004. P. Panchapakesan, D. Manjunath, On the Transmission Range in Dense Ad Hoc Radio Networks. In Proceedings of IEEE SPCOM01. 2001. P. Gupta, P.R. Kumar. Critical Power for Asymptotic Connectivity in Wireless Networks. Stochastic Analysis, Control, Optimization and Applications. Boston: Birkhauser, pp. 547-566, 1998. O. Dousse, P. Thiran, M. Hasler. Connectivity in ad hoc and hybrid networks. In Proceedings of INFOCOM 2002, 2002. B. Bui-Xuan, A. Ferreira, and A. Jarry. Evolving graphs and least cost journeys in dynamic networks. In Proceedings of WiOpt03 Modeling and Optimization in Mobile, Ad-Hoc and Wireless Networks, Sophia Antipolis, March 2003. V. Arnold and A. Avec, Ergodic problems of classical mechanics. Benjamin, New York. 1968. T. K. Madsen, F. Fitzek, R. Prasad. Impact of Dierent Mobility Models on Connectivity Probability of a Wireless Ad Hoc Network. In Proceedings of International Workshop on Wireless Ad Hoc Networks. June 2004. T. K. Madsen, F. Fitzek, R. Prasad. Simulating Mobile Ad Hoc Networks: Estimation of Connectivity Probability. In Proceedings of WPMC04. 2004. A. Jardosh, E.M. Belding-Royer, K.C. Almeroth, S. Suri, Towards realistic mobility models for mobile ad hoc networks, In Proceedings of the 9th annual international conference on Mobile computing and networking, September 2003. D. Yu, H. Li. Inuence of Mobility Models on Node Distribution in Ad Hoc Networks. In Proceedings of ICCT03. 2003. J. Yoon, M. Liu, B. Noble. Random Waypoint Considered Harmful. In Proceedings of INFOCOM 2003, April 2003.

Authors Vitae

Tatiana K. Madsen was born in Moscow, Russia, in 1975. She is an Assistant Professor in the Department of Communication Technology, University of Aalborg, Denmark. She received her M. Sc. and Ph.D. degrees in Mathematics from Moscow State University, Russia in 1997 and 2000, respectively. The area of her PhD studies was dynamical systems, their ergodic properties and integrability. Her current research interests lies within the areas of ad hoc networks with the focus on realistic mobility models and ad hoc routing and 4G wireless communications including QoS support for multimedia services and IP header compression techniques for wireless networks.

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.19

20

Madsen, Fitzek, Schulte, Prasad

Frank H. P. Fitzek was born in Berlin, Germany in 1971. He is an Associate Professor in the Department of Communication Technology, University of Aalborg, Denmark heading the Future Vision group. He received his diploma (Dipl.-Ing.) degree in electrical engineering from the University of Technology - Rheinisch-Westflische Technische a Hochschule (RWTH) - Aachen, Germany, in 1997 and his Ph.D. (Dr.Ing.) in Electrical Engineering from the Technical University Berlin, Germany in 2002. As a visiting student at the Arizona State University he conducted research in the eld of video services over wireless networks. He co-founded the start-up company acticom GmbH in Berlin in 1999. In 2002 he was Adjunct Professor at the University of Ferrara, Italy giving lectures on wireless communications and conducting research on multi-hop networks. His current research interests are in the areas of 4G wireless communication, QoS support for multimedia services, access techniques, security for wireless communication, and the integration of multi hop networks in cellular systems. Dr. Fitzek serves on the Editorial Board of the IEEE Communications Surveys & Tutorials. He is the program chair for the International Conference on Advances in Computer Entertainment Technology (ACE2004) and serves in the program committee for VTC2003, VTC2004, ACE2004, and IEEE MWN2004.

Jan Gerrit Schulte is the CEO of acticom mobile networks GmbH in Berlin, Germany. He studied electrical engineering at the Technical University Berlin from 1989 until 2001, while the main focus were telecommunication networks, the project alpha core dealt with Internet Measurements and Characterists, and the Diploma thesis was about routing and stability in wireless ad-hoc networks. From 1996 until 2000 he worked as student researcher at the Telekommunications Network Group at TU-Berlin. From there Gerrit Schulte changed to SIEMENS AG WM where he worked within Bluetooth Development and Design. In 2001 Gerrit Schulte began to work as CEO of acticom, a company that he also co-founded in 1999 and which is developing software for wireless telecommunication networks and devices, such as mobile phones and cellular networks.

Ramjee Prasad received a B.Sc. (eng) degree from Bihar Institute of Technology, Sindri, India, and M.Sc. (Eng) and Ph.D degrees from Birla Institute of Technology (BIT), Ranchi, India, in 1968, 1970 and 1979, respectively. During February 1988 May 1999 he has been with the

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.20

Connectivity Probability of Wireless Ad Hoc Networks

21

Telecommunications and Trac-Control Systems Group of Delft University of Technology (DUT), The Netherlands, where he was actively involved in the area of wireless personal and multimedia communications (WPMC). He was head of the Transmission Research Section of International research Centre for Telecommunications Transmission and Radar (IRCTR) and also Program Director of the Centre for Wireless Personal Communications (CEWPC). As from June 1999 Ramjee Prasad joined as the Wireless Information Multimedia Communications chair and co-director of Center for PersonKommunikation at Aalborg University, Denmark. From January 2004 he is Director of the Center for Teleinfrastruktur (CTIF). He has published over 500 technical papers, and authored and co-edited a series of book about Wireless Multimedia Communications (Artech House). His research interest lies in wireless networks, packet communications, multiple access protocols, adaptive equalizers, spread-spectrum CDMA systems and multimedia communications. He is a fellow of the IEE, a fellow of IETE, a senior member of IEEE, a member of NERG, and a member of the Danish Engineering Society (IDA).

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.21

madsen_fitzek_schulte_prasad_revised.tex; 13/04/2005; 15:22; p.22