Reliable and Energy-efficient Multi-hop Wireless Networks

Open Access
Das, Arnab
Graduate Program:
Electrical Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
August 26, 2009
Committee Members:
  • George Kesidis, Dissertation Advisor
  • George Kesidis, Committee Chair
  • Constantino Manuel Lagoa, Committee Member
  • John Metzner, Committee Member
  • John F Doherty, Committee Member
  • Dr Jun Shu, Committee Member
  • Christopher H Griffin, Special Member
  • Multi-hop Wireless Networks
  • WMN
  • IWSN
  • reliability
  • energy-efficient
In this thesis, we study multi-hop versions of two classes of wireless networks: (a) Mobile Ad Hoc Networks (MANETs) and (b) Wireless Mesh Networks (WMNs). Our primary focus is to develop reliable structures for MANETs and energy-efficient implementations for WMNs. We study a MANET setting where each node in the MANET engages in deliberate and discreet packet-dropping at a certain rate, while relaying on behalf of other nodes. Assuming the nodes to be non-cooperative, we develop a scheme in which each source node would blame its nearest neighbor for any packet dropped along the multi-hop path vector. This is simple, but elegant, as it still reveals the true relative dropping rates of each packet-dropping node in the MANET. We then propose a novel scheme for improving the overall performance or robustness of such a MANET by adopting a routing strategy which considers the recent packet-dropping performance (reputation) of intermediate relaying nodes in determining the best route from a source to a destination. For deciding the best route, we propose the use of mechanisms based on the modification of the Ad Hoc On-Demand Distance Vector (AODV) routing protocol. As opposed to the AODV scheme, which uses a broadcast mechanism for route discovery, the proposed scheme (called Robust AODV or RAODV) considers the reputation levels of neighboring nodes to determine the node(s) to which the route discovery packet should be sent. In connection with WMNs, we study a WMN structure deployed for residential broadband Internet access and WMN applications for Wireless Sensor Networks (WSNs). Our WMN model deployed for residential broadband Internet access has a specific topology of three tiers comprising WMN infrastructure nodes or Mesh Access Points (MAPs), fixed, in-range end-users or end-user Stations (STAs), and potentially mobile, out-of-range end-users or Mobile-STAs (M-STAs). Our interest is only at the edge as our primary focus is the range extension of the WMN to M-STAs via the relaying services of STAs. We develop an iterative game model in which the relaying STAs engage in deliberate packet-dropping of M-STA packets for their own throughput gain. We next extend the model to include the scenario in which the MAP and the STAs are network service providers in respectively a primary and a secondary market. The STAs do not engage in deliberate packet-dropping, since they are now wholly responsible for providing traffic to the M- STAs, and as a result will lose revenue. Obviously, the second model is better for mobile end-users (M-STAs) as they receive better service since STAs do not drop packets in this scenario for their own selfish reasons. We then investigate WMNs used in Advanced Metering Infrastructure (AMI), which is a form of WSN. These WMNs usually consist of a collection of Neighborhood Area Networks (NANs), which include smart, wireless-enabled meters or sensor nodes. We present an elegant approach to reducing the cost of the AMIs by increasing the ratio of the number of sensors/meters in Neighborhood Area Networks (NANs) to the number of Access Points (APs) or gateways. Finally, we propose schemes for energy conservation in Industrial Wireless Sensor Networks (IWSNs), which are variants of WMNs. Our contribution integrates new approaches to partitioning an IWSN into energy-efficient clusters with a graph-theoretical framework for energy-efficient routing within a cluster. The concept of betweenness centrality is demonstrated to be a useful metric to optimize the placement of the cluster head within an IWSN cluster in our clustering approach. In addition, our strategy combines a novel scheme for duty cycling of the nodes in a cluster to further minimize energy expenditure in an IWSN. Betweenness centrality is also used as a metric for developing a duty-cycling approach to further minimize the energy consumption in an IWSN. Most of the previous contributions have dealt with only one particular approach to minimizing energy consumption in an IWSN. The emphasis in some of them has been on efficient partitioning of the IWSN into clusters, while others have dealt with energy-efficient routing or effective approach to duty cycling of sensor nodes. In contrast, this contribution integrates a new graph-theoretical framework for partitioning an IWSN into energy-efficient clusters, a minimum-cost (i.e., energy-efficient) routing scheme within a cluster, and a novel duty-cycling scheme to further minimize the cluster’s energy consumption.