Catálogo de publicaciones - libros

Tremendous advances in embedded systems, sensors and wireless communications technology have made it possible to build large-scale wireless sensor networks (WSNs). Due to their potential applications,WSNs have attracted significant attention in the industry, academic, and government organizations. In particular, by commanding a large number of distributed and coordinated sensor nodes, WSNs can effectively act as the human-physical world interface in future digital world through sensing and actuating. However, the inherent characteristics of WSNs typified by extremely scarce resources (e.g., bandwidth, CPU, memory and battery power), high degree of uncertainty, and lack of centralized control pose significant challenges in providing the desired quality, information assurance, reliability, and security. This is particularly important for mission critical applications that involve information intensive WSNs including video sensors. In this talk, we will examine the uncertainty-driven unique challenges and key research problems in information intensive wireless sensor networks in the areas of aggregation, clustering, routing, data dissemination, coverage and connectivity, and security. We will present our novel solutions to some of these problems and conclude with future directions.

The recent researches in wireless networks prompt the opportunistic transmission that exploiting channel fluctuations to improve the overall system performance. In wireless ad hoc networks, nodes may have packets destined to multiple neighboring nodes. We consider an opportunistic scheduling that takes advantage of time-varying channel among different receivers to improve system performance. Maximizing overall throughput and satisfying QoS requirements for transmission flows are two important objectives that need to be considered. In literature, many opportunistic scheduling policies for ad hoc networks have been proposed, in which each transmitter schedules the transmission indepen-dently. However, due to co-channel interference, the decisions of neighboring transmitters are highly correlated. Moreover, to achieve the QoS requirements, nodes have to be cooperative to share the common wireless channel. In this paper, we formulate the opportunistic scheduling problem taking the interaction among the neighboring transmitters into account. We present an optimal scheduling policy which maximizes the overall network performance while satisfying QoS requirements of individual flows. We also proposed COS, a distributed Cooperative and Opportunistic Scheduling algorithm, which modifies IEEE 802.11 protocol to implement the optimal scheduling policy by exchanging average channel conditions and QoS factors among 2-hop neighboring nodes. Simulation results indicate that our implementation achieves higher network throughput and provides better QoS support than existing work.

Consider a wireless ad hoc network with random access channel. We present a model that takes into account topology, routing, random access in MAC layer (governed by IEEE 802.11orslotted aloha) and forwarding probability. In this paper , we are focusing to study the effect of cooperation on the stability and throughput of ad-hoc network. Forwarding packets of other nodes is an example of activity that requires such a collaboration. Hence, it may not be in interest of a node to always forward the requesting packet. We propose a new approach (based on ) to derive throughput of multi-hop routes and stability of forwarding queues. With this cycle approach, we correct the analytical expressions derived in [2] and discover that their results are valid only in particular cases such as symmetric networks. However, in this paper, we get extended results for general network case. Moreover, we confirm that (i) the forwarding queues in a system of weighted fair queues has a special property and (ii) the end to end throughput of a connection does not depend on the load of the intermediate forwarding queues between a source and a destination. We perform extensive simulations and verify that the analytical results exactly match the results obtained from simulations.

The paper addresses mobility control routing in robotic sensor wireless networks, where either mobile sensors or mobile actuators assist in wireless data intensive transmissions from sensors. Given a communication request between a source destination pair, the problem is to find a route and move each node on the route to its desired location, such that total transmission power is minimized. We study the optimal number of hops and optimal distance of adjacent nodes on the route. We propose OHCR algorithm that is based on the optimal number of hops on the route. We further propose MPoPR algorithm that minimizes transmission power over progress. After finding an initial path, strategies of move in rounds and move directly are both considered to move nodes to desirable locations. Extensive simulations are conducted to evaluate our proposed routing algorithms.

Multicast for ad hoc networks becomes popular. However, along with increase of the network size and node mobility, it is difficult to design the multicast routing for large ad hoc networks with high mobility. In order to cope with this difficulty, this paper proposes a new hierarchical multicast routing for such ad hoc networks. The characteristics of the proposed scheme are introduction of the inter-cluster group mesh in the autonomous clustering. The states of clusters are changed according to join and leave of multicast members and the group mesh among clusters is dynamically constructed. The simulation results show that the proposed scheme is scalable for the network size and adaptable to node mobility.

Directed diffusion (DD) is a data-centric routing protocol based on purely local interactions between individual network nodes. This protocol uses application-specific context for data aggregation and dissemination. Therefore, it can be completely matched to the application requirements in a large distributed sensor network. Many work have been recently done to improve the energy efficiency of this protocol. In this paper, an extension to DD is presented in order to construct multiple paths between the sink and the sensor sources. Using this method, load-balancing is implemented to increase the life-time of the sensor nodes collaborating in the routing process. The proposed protocol, Multi-path directed diffusion (MDD), can produce more than one disjoint or braided paths and spread the data collected in the sources, properly between the paths. In this way, an efficient load balancing mechanism has been implemented. The simulation results show that through using MDD, the lifetime of the network connections between the sources and the sink will be increased and the interest flooding rate which is proved to be an expensive operation can be reduced.

In this paper, we are interested in simple routing protocols, for wireless sensor networks. One of the major challenging issue in sensor networks is that nodes rely on batteries with limited capacity. Therefore, to extend the life of the sensors and the reliability of the network, an efficient routing algorithm can reduce/optimize energy consumption by reducing routing overhead while maintaining a high delivery rate. Moreover, a good routing method should compute a routing path that is as close as possible to the shortest path. We propose , a localized energy efficient greedy routing scheme. To minimize the gap between the shortest path and the computed path, ORouting computes its routing path by locally selecting the next hop node based on its orthogonal distance to the direction of the source/destination pair. Moreover, in its routing decision, ORouting is biased towards its neighbors in the forward direction towards the destination. We compare ORouting to several routing protocols. We show, through simulation, that our protocol improves significantly energy consumption and achieves a high percentage of successful routings.

Gossiping is a lightweight and simple technique for information dissemination in many application domains, be it in Wireless Sensor Networks (WSNs), Mobile Ad-hoc Networks (MANETs), or Vehicular Ad-hoc Networks (VANETs). Much research has been conducted in the past on probabilistic dissemination methods because of their efficiency compared with simple flooding and their simple application. However most work was focused on static gossiping, i.e., the gossiping probability cannot be adapted according to topology changes. Thus, topology characteristics have to be known in advance.

In this work the use of position information for building up a neighborhood relationship is proposed. Based on this information, a forwarding hierarchy is constructed and the protocol is capable to adjust the dissemination probability dynamically in a distributed manner. The protocol is evaluated in a highway scenario, where the network characteristic varies from sparse networks with highly mobile nodes to a traffic jam with very high node density and low node velocities. The applicability of the proposed protocol for such scenarios is shown by simulations.

To balance energy consumption for nodes is an important factor considered in wireless sensor networks (WSN). In the paper, we research a heterogeneous wireless sensor networks with two different type sensors which have different initial energy as well as have different length data message to transmit. Each sensor node in such a network is systematically gathering and transmission sensed data to a base station for further processing. We develop and analyze the protocol based on residual energy and energy consumption rate (REECR), which is an energy efficient routing protocol we proposed previously for heterogeneous wireless sensor networks. Although REECR protocol is more energy efficient than Low-Energy Adaptive Clustering Hierarchy (LEACH) protocol, it is not very ideal to balance the energy consumption of nodes, namely, from the first node dies to the last node dies, the time span is long. This paper proposed a zone-based REECR (ZREECR) routing protocol to balance the energy consumption of nodes in the networks, simulation results show that all nodes die from start to end become shorter, the balance of energy consumption between nodes is improved.

To alleviate the problem of sub-optimal routing and multiple encapsulations in nested mobile network, a hierarchical based route optimization scheme (HBRO) is proposed. Root mobile router (MR) works as mobility anchor point (MAP) and transmits packets for the mobile network nodes (MNNs). Following advantages can be achieved with this scheme: 1) correspondent node (CN) transmits packet directly to MAP, bypassing all of the home agents (HA) of MRs; 2) tunnel exists only inside the mobile network. We build an analysis model for for performance evaluation, and the analysis results indicate that packet delivery overhead decreases significantly, and the delivery efficiency is beyond 80 percent when packet size is larger than 200 bytes and network resources are utilized to a great extent.