Clustering of wireless sensor and actor networks based on sensor distribution and connectivity

https://doi.org/10.1016/j.jpdc.2009.02.004Get rights and content

Abstract

Wireless Sensor and Actor Networks (WSANs) employ significantly more capable actor nodes that can collect data from sensors and perform application specific actions. To take these actions collaboratively at any spot in the monitored regions, maximal actor coverage along with inter-actor connectivity is desirable. In this paper, we propose a distributed actor positioning and clustering algorithm which employs actors as cluster-heads and places them in such a way that the coverage of actors is maximized and the data gathering and acting times are minimized. Such placement of actors is done by determining the k-hop Independent Dominating Set (IDS) of the underlying sensor network. Basically, before the actors are placed, the sensors pick the cluster-heads based on IDS. The actors are then placed at the locations of such cluster-heads. We further derive conditions to guarantee inter-actor connectivity after the clustering is performed. If inter-connectivity does not exist, the actors coordinate through the underlying sensors in their clusters to adjust their locations so that connectivity can be established. The performances of the proposed approaches are validated through simulations.

Introduction

Wireless Sensor and Actor Networks (WSANs) have started to receive growing attention from the research and engineering communities in recent years. Potential applications of WSANs include detecting and countering pollution in coastal areas, performing oceanic studies of bird/fish migration and weather phenomena, detection and prevention of forest fires, deterring terrorist threats to ships in ports, destruction of mines in different environments facilitating/conducting urban search and rescue (USAR), detecting suspiciously active chemical/biological agents, etc., [3]. Such networks employ a large number of miniaturized low-cost sensing nodes that are responsible for measuring ambient conditions and reporting such measurements to some actor nodes over wireless communication links.

Coverage is one of the most important design goals in most applications of WSANs [8]. It is often required for the network to provide services at every part of the deployment area. In addition to coverage, actors’ responsiveness is usually desired in order for the network to be effective. For example in forest monitoring applications, actors such as fire trucks and flying aircrafts need to be engaged as rapidly as possible to control a fire and prevent it from spreading. Similarly for scientific studies or space applications actors should respond instantaneously to record rare phenomena, e.g., capture an image or record a weird behavior of a habitat.

In order to provide such services over large areas, clustering of the WSAN is often pursued given that each actor can also act as a sink [3], [5]. In a clustered WSAN architecture, each actor acts as a cluster-head and takes certain actions based on the received data from the sensors within its cluster. However, before forming the clusters it is important to determine the location of the cluster-heads in order to maximize the coverage of all actors. A good coverage should minimize the overlap among the action ranges of the actors and include all the sensors deployed within the monitored region. For instance, there is no need to maintain actors at locations where no sensors exist to sense the nearby objects. Therefore, we need to position the actors as evenly as possible within the sensor network for maximizing coverage. Note that this is a different interpretation of coverage compared to previous works which only considered area for coverage. Based on this motivation, for example in a forest monitoring application where robots capable of sprinkling water are used as actors, the positions of such actors should be decided based on the distribution of sensors in the forest. Given that sheer number of sensors is deployed, deterministic placement of sensors may not be possible and thus the sensors need to self organize and coordinate in order to determine the number and/or the best locations of actors needed in extinguishing the fire.

In order to achieve such good distribution of actors within the sensor network, we propose to determine the k-hop independent dominating set (k-IDS) of the sensor network and position the actors next to the location of the nodes in this set. Utilizing this positioning technique, this paper provides a clustering approach for WSANs where each actor leads a cluster consisting of the sensor nodes dominated by a dominator in k-IDS. Here, the parameter k shows the number of hops for a sensor to reach its dominator and thus its cluster-head. The value of k can be determined based on the available number of actors, action range of the actors and/or application level needs such as delay and throughput. Placing the actors at the locations of the nodes in k-IDS will provide two things: First, it will place actors in all parts of the sensor network which boosts the coverage of the actor network. Second, it will guarantee for each sensor to relay its data within k hops to the cluster-head and thus help to reduce the decision time for the actors when a collaborative action is planned.

Since determining IDS and thus k-IDS of a network is an NP-Hard problem [17], [23], [15], one of the goals of this paper is to provide a heuristic approach to solve this problem. We provide a probabilistic and distributed heuristic which considers a neighborhood of k-hop radius for each node when picking the dominators. Once the dominators are determined, each sensor joins to the closest dominator’s cluster by unicasting a message to the dominator node. Each dominator keeps a list of the nodes it dominates and then shares this cluster list with the actor which will be placed next to it.

Once the actors are placed and the network is clustered around the actors, one of the other important metric to be achieved is inter-actor connectivity. Such connectivity is crucial in some applications where actor need to do a collaborative action. In such a case, they need to talk to each other to decide which actors to be picked for the best response in a distributed manner. Given the time criticality of the actions in WSANs, such a coordinated action should also be planned as fast as possible. To enable convenient and fast interaction among the actors, a connected inter-actor network topology is desirable. Therefore, we further extended our actor placement algorithm in order to establish inter-actor connectivity. We first derived the necessary conditions to guarantee such connectivity by adjusting the parameters such as the transmission range. Later, we provided a general solution which will enable connectivity regardless of the size of the actor transmission range. The basic idea is to determine the partitions of actors within the WSAN and then let them communicate through the border sensors within each cluster. Once the locations for the partitions are determined, they apply a contraction algorithm around the primary partition (i.e., the one which has the actor with the highest node ID) to ensure inter-actor connectivity at the expense of some messaging cost and slight degradation in the total coverage.

Both k-IDS-based actor clustering and connectivity establishment approaches were validated through extensive simulations. k-IDS-based actor placement and coverage is shown to be both effective in achieving high coverage and reduced delay and number of actors employed. On the connectivity side, the inter-actor connectivity was achieved with constant messaging overhead on both actors and sensors without significantly reducing the total actor coverage. The contributions of this paper can be listed as follows:

  • A novel clustering algorithm for WSANs based on k-IDS in order to determine the actor locations.

  • A novel distributed heuristic for k-IDS problem.

  • A message efficient and distributed relocation algorithm to establish connectivity among actors which serve as cluster-heads.

This paper is organized as follows. Section 2 summarizes the related work and distinguishes our work from the previous research. In Section 3 we provide the system model and then define the actor placement and clustering problem and present our approach. Section 4 includes the simulations to evaluate the performance of our clustering approach. In Section 5, we provide the analysis and algorithms to provide inter-actor connectivity with the placement mechanism used in Section 3 along with performance evaluation. Finally, Section 6 concludes the paper.

Section snippets

Related work

In this section, we summarize the related work regarding coverage, connectivity and clustering separately and distinguish our approaches from the works presented in the literature.

Clustering of WSANs around actors based on the distribution of sensors

In this section, we first describe the system model and assumptions and provide the problem definition without considering inter-actor connectivity. Then we describe our actor placement approach before the clustering is performed.

Experimental evaluation of IDS-based clustering

The effectiveness of the k-IDS based clustering approach is validated through simulation. This section describes the simulation environment, performance metrics, and experimental results.

Establishing connectivity among actors

Once the actors are placed, an important consideration is to determine whether the actors form a connected topology or not so that they can communicate with each other for collaboration purposes. Note that in WSANs, collaboration of actors is often required in order to decide which actors will be performing the required tasks. Although this can be done through utilizing sensors as relays, the decision making process in general usually requires a lot of message exchanges and thus may put a lot

Conclusion

Wireless sensor and actor networks (WSANs) are gaining popularity in a number of civil and military applications such urban search-and-rescue, border protection, etc. Actors collected sensor’s data and collaboratively perform tasks in response to detected events/targets. In this paper, assuming a coverage metric based on the covered sensor count, we presented a distributed coverage-aware placement and clustering mechanism for WSANs. Our approach utilizes k-IDS of the underlying sensor network

Acknowledgment

This work is partially supported by Southern Illinois University Carbondale Faculty Seed Grant.

Kemal Akkaya received his B.S. and M.S. degrees in Computer Science from Bilkent University, Ankara, Turkey in 1997 and Middle East Technical University (METU), Ankara, Turkey in 1999 respectively. In 2005, he received his Ph.D. in Computer Science from University of Maryland Baltimore County. Currently, he is an assistant professor in the Department of Computer Science at Southern Illinois University Carbondale. His research interests include energy aware routing, security and quality of

References (27)

  • I.F. Akyildiz et al.

    Wireless sensor and actor networks: Research challenges

    Elsevier Ad hoc Network Journal

    (2004)
  • K. Akkaya et al.

    Maximizing connected coverage via controlled actor relocation in wireless sensor and actor networks

    Elsevier Computer Networks Journal

    (2008)
  • M. Younis et al.

    Strategies and techniques for node placement in wireless sensor networks: A survey

    Elsevier Ad Hoc Network Journal

    (2008)
  • A. Abbasi, K. Akkaya, M. Younis, A distributed connectivity restoration algorithm in wireless sensor and actor...
  • K. Akkaya, et al. Distributed recovery of actor failures in wireless sensor and actor networks, in: Proceedings of IEEE...
  • K. Akkaya et al.

    Coverage and delay aware actor placement mechanisms for wireless sensor and actor networks

    International Journal of Sensor Networks

    (2008)
  • P. Alimonti et al.

    Improved approximations of independent dominating set in bounded degree graphs

  • A.D. Amis, et al. Max–min D-cluster formation in wireless Ad Hoc networks, in: Proceedings of IEEE INFOCOM, March...
  • M. Batalin, G. Sukhatme, The analysis of an efficient algorithm for robot coverage and exploration based on sensor...
  • M. Batalin, G.S. Sukhatme, Spreading out: A local approach to multi-robot coverage, in: Proceedings of the...
  • M. Batalin, G.S. Sukhatme, Coverage, exploration and deployment by a mobile robot and communication network, in:...
  • S. Bandyopadhyay, E. Coyle, An energy-efficient hierarchical clustering algorithm for wireless sensor networks, in:...
  • B. Basu et al.

    Movement control algorithms for realization of fault-tolerant Ad hoc robot networks

    IEEE Networks

    (2004)
  • Cited by (63)

    • Agent-based broadcast protocols for wireless heterogeneous node networks

      2018, Computer Communications
      Citation Excerpt :

      Each cluster contains a set of things managed by a Cluster-head. Broadcasting is performed locally in each cluster and getaways are used for the communication between cluster heads to send the items to their final destinations [17,18,25]. The problem of permutation routing is involved when each node in the network needs to receive information from other nodes.

    • Distributed connectivity restoration in networks of movable sensor nodes

      2016, Computers and Electrical Engineering
      Citation Excerpt :

      In schemes proposed by Vaidya et al. [12] and Akkaya et al. [13], the redundant nodes are assigned to each critical node say cut-vertex node as backup nodes in order to tolerate failures of cut-vertex nodes. In [11] Akkaya et al. propose a distributed clustering algorithm which employs actors as cluster-heads and places them in such a way that the connectivity and coverage provided by actors is maximized. Such placement of actors is done by determining the k-hop Independent Dominating Set (IDS) of the underlying sensor network.

    • Efficient event handling in Wireless Sensor and Actor Networks: An on-line computation approach

      2016, Journal of Network and Computer Applications
      Citation Excerpt :

      In all the above discussed cases, actors should cooperate so that the most suitable actors finally handle the on-going event. A cluster-based approach for organizing SNs around actors has also been followed in Akkaya et al. (2009). Specifically, a dominating set is determined in the sensor network such that each of the remaining SNs are at most k hops away from one of the dominators and any two dominators are at least one hop away of each other in this set.

    • Restricted randomness DBSCAN : AA faster DBSCAN Algorithm

      2021, ACM International Conference Proceeding Series
    View all citing articles on Scopus

    Kemal Akkaya received his B.S. and M.S. degrees in Computer Science from Bilkent University, Ankara, Turkey in 1997 and Middle East Technical University (METU), Ankara, Turkey in 1999 respectively. In 2005, he received his Ph.D. in Computer Science from University of Maryland Baltimore County. Currently, he is an assistant professor in the Department of Computer Science at Southern Illinois University Carbondale. His research interests include energy aware routing, security and quality of service issues in ad hoc wireless and sensor networks.

    Fatih Senel received his B.S. degree from the department of Computer Science at Bilkent University, Turkey, in 2004. He is currently a graduate student in the department of Computer Science at Southern Illinois University Carbondale. His research interests include clustering, relocation and fault-tolerance in wireless sensor and actor networks.

    Brian McLaughlan received a B.S. and M.S. in Computer Engineering from the University of Arkansas. He worked for a decade as an IT manager and software developer before pursuing a Ph.D. in Computer Science in the department of Computer Science at Southern Illinois University. His current interests include command and control of massive agent systems, ubiquitous computing, and emergent swarm behavior.

    View full text