Real-Time Coordination and Routing in Wireless Sensor and Actor Networks.
ABSTRACT In Wireless Sensor Actor Networks (WSAN), sensor nodes perform the sensing task and actor nodes take action based on the sensed
phenomena in the field. To ensure efficient and accurate operations of WSAN, new communication protocols are imperative to
provide sensor-actor coordination in order to achieve energy-efficient and reliable communication. Moreover, the protocols
must honor the application-specific real-time delay bounds for the effectiveness of the actors in WSAN.
In this paper, we propose a new real-time coordination and routing (RCR) framework for WSAN. It addresses the issues of coordination among sensors and actors and honors the delay bound for routing
in distributed manner. RCR configures sensors to form hierarchical clusters and provides delay-constrained energy aware routing (DEAR) mechanism. It
uses only cluster-heads to coordinate with sink/actors in order to save the precious energy resources. The DEAR algorithm
integrates the forwardtracking and backtracking routing approaches to establish paths from source nodes to sink/actors. In
the presence of the sink in WSAN, it implements the centralized version of DEAR (C-DEAR) to coordinate with the actors through
the sink. In the absence of sink or ignoring its presence, there is a distributed DEAR (D-DEAR) to provide coordination among
sensors and actors. Cluster-heads then select the path among multiple alternative paths to deliver the packets to the actors
within the given delay bound in an efficient way. Simulation experiments prove that RCR achieves the goal to honor the realistic application-specific delay bound.
- SourceAvailable from: edu.tr
Conference Paper: Exploiting Energy-aware Spatial Correlation in Wireless Sensor Networks[Show abstract] [Hide abstract]
ABSTRACT: Wireless sensor networks (WSNs) promise fine-grain monitoring in a wide variety of applications, which require dense sensor nodes deployment. Due to high density of nodes, spatially redundant or correlated data is generated. Redundancy increases the reliability level of information delivery but increases the energy consumption of the nodes too. Since energy conservation is a key issue for WSNs, therefore, spatial correlation can be exploited to deactivate some of the nodes generating redundant information. In this paper, we present an energy-aware spatial correlation based on a clustering protocol. In this approach, only the cluster-heads are responsible of exploiting spatial correlation of their member nodes and selecting the appropriate member nodes to remain active. The correlation is based on the distortion tolerance and the residual energy of member nodes. Each cluster-head divides its clustered region into correlation regions and selects a representative node in each correlation region which is closer to the center of correlation region and has the higher residual energy. Hence, the whole field is represented by a subset of active nodes which perform the task well. Simulation results prove that the required reporting rate can be achieved with lesser number of nodes by exploiting spatial correlation and eventually conserving the nodes energy.Communication Systems Software and Middleware, 2007. COMSWARE 2007. 2nd International Conference on; 02/2007
- [Show abstract] [Hide abstract]
ABSTRACT: Wireless sensor and actuator network (WSAN) are composed of a large number of heterogeneous sensors and actuators. In the automated WSAN, sensors are collaborated to monitor the physical phenomenon in the surveillance field, while the actuators are to collect sensing data, process the data and perform appropriate actions without the existence of central controller. Most WSANs are used in the real-time sensing and reaction systems towards physical environment. In this paper, we propose a real-time architecture for automated WSAN to bind the latency in applications. In the architecture, we present distributed mechanisms for sensor-actuator event reporting and self-aware coordination to maintain the delay bound sensor-actuator communication. And then we present mechanism for ordered multi-event task assignment, and the acting coordination mechanism to provide efficient reaction and execution of the event task in time. Preliminary simulation results are presented to demonstrate the advantages of our solutions.01/2009;
- [Show abstract] [Hide abstract]
ABSTRACT: In wireless sensor and actor network research, the commonly used mobility models for a mobile actor are random walk model, random waypoint mobility model, or variants thereof. For a fully connected network, the choice of mobility model for the actor is not critical because, there is at least one assured path from the sensor nodes to the actor node. But, for a sparsely connected network where information cannot propagate beyond a cluster, random movement of the actor may not be the best choice to maximize event detection and subsequent action. This paper presents static and dynamic intelligent mobility models that are based on the inherent clusters’ information of a sparsely connected network. Simulation results validate the idea behind the intelligent mobility models and provide insights into the applicability of these mobility models in different application scenarios.Telecommunication Systems 03/2009; 40(3):141-149. · 1.16 Impact Factor
Real-Time Coordination and Routing in
Wireless Sensor and Actor Networks
Ghalib A. Shah, Muslim Bozyi˘ git1,¨Ozg¨ ur B. Akan, and Buyurman Baykal2
1Department of Computer Engineering,
Middle East Technical University, Ankara, Turkey 06531
2Department of Electrical and Electronics Engineering,
Middle East Technical University, Ankara, Turkey 06531
Abstract. In Wireless Sensor Actor Networks (WSAN), sensor nodes
perform the sensing task and actor nodes take action based on the sensed
phenomena in the field. To ensure efficient and accurate operations of
WSAN, new communication protocols are imperative to provide sensor-
actor coordination in order to achieve energy-efficient and reliable com-
munication. Moreover, the protocols must honor the application-specific
real-time delay bounds for the effectiveness of the actors in WSAN.
In this paper, we propose a new real-time coordination and routing
(RCR) framework for WSAN. It addresses the issues of coordination
among sensors and actors and honors the delay bound for routing in dis-
tributed manner. RCR configures sensors to form hierarchical clusters
and provides delay-constrained energy aware routing (DEAR) mecha-
nism. It uses only cluster-heads to coordinate with sink/actors in order to
save the precious energy resources. The DEAR algorithm integrates the
forwardtracking and backtracking routing approaches to establish paths
from source nodes to sink/actors. In the presence of the sink in WSAN,
it implements the centralized version of DEAR (C-DEAR) to coordinate
with the actors through the sink. In the absence of sink or ignoring its
presence, there is a distributed DEAR (D-DEAR) to provide coordina-
tion among sensors and actors. Cluster-heads then select the path among
multiple alternative paths to deliver the packets to the actors within the
given delay bound in an efficient way. Simulation experiments prove that
RCR achieves the goal to honor the realistic application-specific delay
Recent advances in the field of sensor networks have led to the realization of
distributed wireless sensor networks (WSN). A WSN is composed of large num-
ber of sensor nodes, which are densely deployed in the sensor field in a random
fashion with a sink node. The task of sensor nodes is to detect the events in
the sensors field and route them to the sink node, which is responsible for the
monitoring of the field.
Y. Koucheryavy, J. Harju, and V.B. Iversen (Eds.): NEW2AN 2006, LNCS 4003, pp. 365–383, 2006.
c ? Springer-Verlag Berlin Heidelberg 2006
366G.A. Shah et al.
Recently, the capabilities of the WSN are extended to include the actor nodes
responsible with taking action against the detected events . Such architecture
is called a wireless sensor and actor networks (WSAN), where a small numbers
of actors, as compared to sensors, are spread in the sensor field as well. Actors
are mostly mobile and resource-rich devices and can be thought to form a mobile
ad hoc network of their own. This paradigm of WSAN is capable of observing
the physical world, processing the data, making decisions based on the sensed
observation and performing appropriate actions. Typically, the architecture of a
WSAN consists of sensors which sense the phenomena, a sink that collects the
data from the sensors to process and actors that act upon the command sent
by the sink. In the literature, such architecture is known as semi-automated ar-
chitecture. An architecture in which sensor nodes send information to the actor
nodes directly without the involvement of sink node is called an automated archi-
tecture . It is apparent that the communication path in a semi-automated ar-
chitecture introduces significant delay, which is not acceptable for delay-sensitive
applications. For example, consider a military application where sensors in the
battlefield will detect the movement of red forces and send the information to
the sink which is situated in a remote command and control station. The sink
then triggers an action through an actor to counter in the threat area. In this
case unnecessary delay is introduced due to sensor-sink communication which
could be removed if the actors can take localized actions without the involve-
ment of the sink; depending on the data sent by the sensors. Hence, the most
challenging task in WSAN is the coordination between sensors and actors to
provide real-time response.
In WSAN, the effective sensor-actor coordination requires the sensors to know
the right actors and the routes to reach them. Moreover, it requires delay es-
timates for all possible routes. In addition to energy constraints as in WSN,
WSAN also imposes timing constraints in the form of end-to-end deadlines.
Clearly, there is a need for real-time communication protocols for WSAN, which
provide effective sensor-actor coordination while consuming less energy.
There have been considerable efforts to solve the routing problem in wireless
sensor networks , , , , , , , . However, these protocols
do not consider the heterogeneity of WSAN. Moreover, none of these protocols
provide sensor-actor coordination and real-time routing. A coordination frame-
work  for WSAN has been proposed that is an event-based reactive model of
clustering. Cluster formation is triggered by an event so that clusters are created
on-the-fly to optimally react to the event itself and provide the required relia-
bility with minimum energy expenditure. Reactive cluster formation algorithms
have the disadvantage that they consume precious time on event occurrence for
cluster formation. Hence for real-time coordination such an architecture is not
suitable. Moreover, cluster to actor routing in  is done using greedy geographi-
cal approach. A packet forwarding node finds the next hop node according to the
greedy approach failing to do so results into a packet loss as the packets enters
into a void region. Since the work assumes that the network is dense therefore it
does not propose any void region prevention or recovery mode implementation.
Real-Time Coordination and Routing in WSAN367
Consequently, there exists no unified solution which addresses the real-time co-
ordination as well as routing problem for the heterogeneous WSANs.
In this study, we propose a real-time coordination and routing (RCR) frame-
work, which addresses the sensor-actor coordination with real-time packet deliv-
ery in the semi-automated architecture as well as automated architecture. RCR,
incorporates the two components, namely DAWC and DEAR. DAWC is our
heuristic clustering protocol used to dynamically configure the sensor nodes in
the form of clusters to achieve energy efficiency. Whereas, RCR achieves the real-
time demand τ of packet delivery through our delay-constrained energy aware
routing (DEAR) protocol that is the first and foremost aim of the protocol. The
DEAR protocol, described in Section 3, establishes a backbone network by inte-
grating the forward tracking and backtracking mechanism that provides all the
possible routes towards target nodes (sink/actors). The path selection criterion
is based on the packet delay as well as the balance consumption of energy of sen-
sor nodes. In the presence of the sink in WSAN, it implements the centralized
version of DEAR (C-DEAR) to coordinate with the actors through the sink.
On the other hand, when there is no sink or central node in WSAN, it provides
the distributed version of DEAR (D-DEAR) for coordination among sensors and
actors. Performance evaluation study reveals that RCR addresses the real-time
coordination and routing requirements of WSANs.
The remainder of the paper is organized as follows. In Section 2, we present
the cluster formation procedure. We discuss the route path computation and
sensor-actor coordination in Section 3. Performance evaluation and results are
considered in Section 4. Finally, the paper is concluded in Section 5.
2Hierarchical Configuration of Sensor Nodes
To provide a real-time coordination among sensors and actors in WSAN, RCR
configures the sensor nodes hierarchically. The operations of the routing pro-
tocols are discussed in Section 3. The configuration of sensor nodes to achieve
real-time coordination is discussed in this section. We propose so called Dy-
namic Weighted Clustering Algorithm (DAWC). The operations of DAWC con-
sist of cluster formation of sensor nodes, delay budget estimation for forwarding
a packet from the cluster-heads and to guarantee the packet delivery within the
given delay bound τ.
There are many clustering algorithms , , ,  proposed in the literature
but unlike these studies, DAWC is neither periodical clustering procedure nor the
cluster size is fixed in terms of hops. It adapts according to the dynamic topology
of the sensor and actor networks. Our work is motivated from the previous
work “A Weight Based Distributed Clustering” . However, unlike , DAWC
provides the cluster formation procedure to cope with the dynamic number of
hops in a cluster and provides support for real-time routing. Cluster formation is
based on the weighting equation formulated in the Section 2.2, which sets weight
to different application parameters according to the applications need. DAWC
adapts to the variation in the sensors field and can be optimized accordingly.
368G.A. Shah et al.
Once the cluster has been formed, cluster-heads get estimates of delay budget1
of their member nodes. The delay budgets of member nodes help to build the
delay-constrained energy efficient path.
When the sensor nodes are not uniformly deployed in the sensor field, the
density of nodes could be different in different zones of the field. Choosing an
optimal number of clusters kopt, which yield high throughput but incur latency
as low as possible, is an important design goal of DAWC.
In this section, we evaluate the optimal number of clusters koptand, hence, the
optimal size of clusters. We assume the uniform deployment of sensor nodes
and devise a formula to find out kopt. Later, we see the implication of it to
non-uniform deployment for optimal configuration in Section 2.1.
Fig.1. Network model to formulate the optimal clusters. Fig 1(a) represents the model
to find the probability of DN nodes. Fig 1(b) illustrates the model of routing packets
from cluster-heads to the sink node.
Optimal Clusters in the Field. Each member node transmits its data packet
to the cluster-head. Let r be the transmission radius of each node regardless
of its functioning. In the clustering process, there is some probability that a
number of dangling nodes2(DN ) may exist due to the density of nodes or cov-
erage of the elected cluster-head. Let us first find out the probability of such
nodes. To do that, we map the sensor field (M × M) to non-overlapping cir-
cles of radius r as shown in Fig 1(a) and assume that the nodes lying outside
the boundary of the circle are DN nodes and the others are member nodes.
These DN nodes require affiliating to cluster-head through the nodes insides
the circle (member nodes). The squared field M2can be packed by M2/(2r)2
non-overlapping circles of radius r. Thus, the probability γ of a multi-hop
1delay budget is the time to deliver the data packet from the cluster-head to the
2Nodes which have not joined any group or cluster are referred as dangling nodes.
Real-Time Coordination and Routing in WSAN369
Let Eelec be the energy consumed by the electronic circuitry in coding, mod-
ulation, filtration and spreading of the signal. Whereas, ?ampr2is the energy
consumed for signal amplification over a short distance r. Thus, the energy con-
sumed by each member node in transmitting a packet of size l is
EMember= l(Eelec+ ?ampr2(1 + γ))
The above equation can be simplified by taking the area as circle given in Eq.
16 of .
EMember= l(Eelec+ ?ampM2(1 + γ)
Let us assume that the sensory field is covered by a circle of radius R, where the
sink node lies at the center of this circle as shown in Fig 1(b). This assumption is
made for routing packets from cluster-heads to the sink. The assumption is rea-
sonable because it is less likely that a node lying outside the boundary of circle
will be elected as cluster-head due to the low weight than the nodes inside the
circle. Cluster-heads do not extend their transmission range to transmit pack-
ets directly to the sink node and, therefore, has the same radius r as member
nodes. We adapt the multi-hop model proposed by  to route packets from
cluster-head to the sink.
In the model, a circle is divided into concentric rings with the thickness r.
The energy spent to relay the packet from outside ring towards inside ring is
l(2Eelec+ ?ampr2). The number of hops Γ require to route packet from cluster-
head to sink node can be calculated byR
the cluster-head is close enough to the sink to directly transmit packets. This
probability can be calculated by using the nodes distribution in the rings given
r(1−¯ h). Where ¯ h is the probability that
¯ h =r
Since packets from the cluster-heads far from the sink node are relayed through
intermediate nodes. Therefore, if Λ(i) is the number of neighbors of a node i then
Λ(i) × Eelecis the energy consumed by the electronic circuitry of the neighbors
of forwarding node i during the propagation of packet. The number of neighbors
(Λ) of any node can be found as
Λ = nπr2
Hence, the energy consumed in routing data from cluster-head to sink is mea-
ECH−Sink= l(ΛEelec+ Eelec
+ (2Eelec+ ?ampr2+ ΛEelec)Γ).