Catálogo de publicaciones - libros

For active, probing-based bandwidth measurements performed on top of the unifying IP layer, it may seem reasonable to expect the measurement problem in wireless networks, such as ad-hoc networks, to be no different than the one in wired networks. However, in networks with 802.11 wireless links we show that this is not the case.

Our experiments show that the measured available bandwidth is dependent on the probe packet size (contrary to what is observed in wired networks). Another equally important finding is that the measured link capacity is dependent on the probe packet size on the cross-traffic intensity.

The study we present has been performed using a bandwidth measurement tool, DietTopp, that is based on the previously not implemented TOPP method. DietTopp measures the end-to-end available bandwidth of a network path along with the capacity of the congested link.

In an ad hoc network, mobile nodes communicate with each other using multi-hop wireless links. There is no stationary infrastructure; for instance, there are no base stations. Each node in the network also acts as a router, forwarding data packets for other nodes. A central challenge in the design and evaluation of ad hoc networks is the estimation and the monitoring of network resources such as available bandwidth. Considering the dynamic routing protocols that can efficiently find routes between two communicating nodes, the high level of loss and the interference between node transmissions, the available bandwidth measurement techniques already provided by the literature may be inaccurate. Our goal is to carry out a systematic performance study of SloPS’s [1] behavior when acting with three dynamic routing protocols for ad hoc networks: the Dynamic Source Routing protocol (DSR) [2, 3], the Ad Hoc On-Demand Distance Vector protocol (AODV) [4, 5] and the Destination-Sequenced Distance Vector (DSDV) [6].

This article focuses on technical issues related to quality of service provisioning at routing layer in ad hoc networks. It describes the design and implementation of a unified support for quality of service (QoS) metrics within the routing protocol OLSR. This is achieved by the extension of both signalling messages and route calculation process. Major benefits of the proposed approach are to allow dynamic enforcing and adaptation of QoS metrics according to policies defined into the network. QoS metrics information are inserted in a generic way within OLSR signalling messages taking advantage of Linux kernel plugins. These messages are used by the routing process in order to compute routes with respect to the chosen QoS metrics.

In this paper, we consider routing in large wireless multihop networks with possibly irregular topologies. Existing position-based routing protocols have deficiencies in such scenarios as they always forward packets directly towards the destination. Greedy routing frequently fails and costly recovery mechanisms have to be applied. We propose the Ants-based Mobile Routing Architecture (AMRA) for optimized routing, which combines position-based routing, topology abstraction, and swarm intelligence. AMRA routes packets along paths with high connectivity and short delays by memorizing past traffic and by using ant-like packets to discover shorter paths. The geographic topology abstraction allows AMRA to cope with high mobility and large networks. Simulative evaluation indicate that compared to other position-based routing AMRA finds significantly shorter paths with only marginal overhead protocols.

To become realistically untethered, wireless communication networks need to be self-organised, rapidly deployable, infrastructureless and mobile. Existing protocols are efficient in routing data dynamically between mobile nodes that belong to the same connected component. Concrete applications such as Defence and Disaster-Relief cannot always assume that the network is connected (, not partitionned). However, even if the network is continuously partitioned, a “communication path” may be available through time and mobility using intermediate mobile nodes (temporally within reach of each other) — we have coined these “Extremely Mobile Networks”. We consider the problem of routing in a highly mobile network which, possibly, may never be fully connected. We introduce new algorithms that always allow to route a packet toward a remote destination. The packet bounces from connected components to connected components, thanks to node mobility.

This paper presents a transparent multi-level routing scheme, named MORHE, that improves the scalability of the OLSR protocol by exploiting the heterogeneous nature of nodes in the network. In our work we try to take an approach that focuses on scenarios where ad hoc technology can be applied, but where we also find nodes in the network with varying capacity. The MORHE protocol makes use of nodes which have a large capacity (e.g. more energy, larger transmission range) to build something that could best be described as an ad hoc infrastructure. Nodes are grouped in clusters that need to be interconnected by specific nodes. This implies that a node no longer needs to know the entire network topology as is the case of the OLSR protocol, but only needs to maintain routes to the nodes inside its own cluster and to the other clusters. Using this approach the signalling overhead — which is one of the main reasons why OLSR is not scalable — is greatly reduced. We also introduce a simple mobility management scheme to allow nodes to roam the different ad hoc clusters.

The mobility characteristic of Ad Hoc and Sensor networks implies that the topological information contained in the traditional IP address can no longer reflect the position of a node in the network. In this paper we propose a new approach to resolve the location-identification coupling contained in the IP address. The approach uses a Virtual Regular Structure (VRS) to describe the addressing space of the entire network; such a structure provides additional desired properties such as robustness and multi-paths. Our approach explores a distributed implementation of the VRS based on a trellis graph description of routing tables instead of the traditional trees. We show that the construction of the optimal structure is an NP-Complete problem; in this paper, we propose a heuristic and evaluate its performance via simulations.

We study the connectivity properties of an ad hoc network consisting of nodes each moving according to the Random Waypoint mobility model. In particular, we focus on estimating two quantities, the probability that the network is connected, and the mean durations of the connectivity periods. The accuracy of the approximations is compared against numerical simulations. For the probability of connectivity, an approximation is given that is remarkably accurate. By numerical examples we also show that in sparse network the mobility has a positive effect on connectivity, whereas in dense network the situation becomes the opposite. For the mean length of the connectivity periods results are also accurate in the important region where the probability of connectivity rises rapidly.

Games, by definition, offer the challenge of the presence of an opponent, to which a playing strategy should respond. In finite-timed zero-sum games, the strategy should enable to win the game within a limited playing time. Motivated by robot soccer, in this talk, we will present several approaches towards learning to select team strategies in such finite-timed zero-sum games. We will introduce an adaptive playbook approach with implicit opponent modeling, in which multiple team strategies are represented as variable weighted plays. We will discuss different plays as a function of different game situations and opponents. In conclusion, we will present an MDP-based learning algorithm to reason in particular about current score and game time left. Through extensive simulated empirical studies, we will demonstrate the effectiveness of the learning approach. In addition, the talk will include illustrative examples from robot soccer. The major part of this work is in conjunction with my PhD student Colin McMillen.

We define, for an overlay built on top of an ad hoc network, a simple criterion for neighbourhood: two overlay nodes are neighbours if and only if there exists a path between them of at most hops, and is called the (overlay) neighbourhood range. A small may result in a disconnected overlay, while an unnecessarily large would generate extra control traffic. We are interested in the minimum ensuring overlay connectivity, the so-called critical .

We derive a necessary and sufficient condition on to achieve asymptotic connectivity of the overlay almost surely, i.e. connectivity with probability 1 when the number of overlay nodes tends to infinity, under the hypothesis that the underlying ad hoc network is itself asymptotically almost surely connected.

This condition, though asymptotic, sheds some light on the relation linking the critical to the number of nodes , the normalized radio transmission range and the overlay density (i.e., the proportion of overlay nodes). This condition can be considered as an approximation when the number of nodes is large enough. Since is considered as a function of , we are able to study the impact of topology control mechanisms, by showing how the shape of this function impacts the critical .