Antoine Van de Capelle

Performance analysis of unslotted CSMA/CA in wireless networks

In this paper a novel analytical model for the saturation throughput of unslotted Carrier Sensing Multiple Access with Collision Avoidance (CSMA/CA) in wireless networks is proposed. A fixed point procedure is developed based on the interaction of the Physical layer (PHY) and the Medium Access Control sub-layer (MAC). The output of the Clear Channel Assessment (CCA), i.e. idle or busy medium in the neighborhood of a node, serves as a feedback mechanism for the dynamical scheduling rate controlled by the back-off procedure. The PHY is described by a renewal process between successful transmissions with failed attempts and collided packets in between. A semi-Markov process of the internal states of a node is used as a model for the MAC. An event-driven simulator for the non-beacon enabled IEEE Std 802.15.4™MAC is developed to verify the numerical results of the analytical method. A detailed analysis of the idle period after a transmission is carried out based on the proposed analytical approach. The probability that the CCA senses the channel idle depends clearly on the actual back-off stage and the first back-off expiration after a transmission cannot be modeled by a exponential distribution when a finite number of nodes are in contention. The output of the event-driven simulations confirms both statements in great detail and the saturated throughput evaluated with the analytical procedure is verified by event-driven simulations.

Direction of a arrival (DOA) parameter estimation with the SAGE algorithm

In this study, we investigate the estimation of the relative delay, the azimuth and elevation angle and the complex amplitude of a known received signal with the space alternating generalised expectation (SAGE) maximisation algorithm. This optimisation algorithm is used to replace the high-dimensional optimisation procedure, necessary to compute the joint maximum likelihood estimate of the parameters using several separate maximisation processes, which can be performed sequentially. Spherical arrays provide wide scan coverage with low grating lobe levels. Because a uniform antenna element distribution on this spherical surface gives rise to the smallest variations in received signal levels over the whole angular space, a uniform distribution of antennas on a sphere is used. For this completely general three-dimensional array, we have for the fitst time investigated the resolution of the SAGE algorithm, the convergence and the bit error rate as a function of the signal to noise ratio. We will show that this algorithm is a powerful tool that can be applied successfully for high-resolution parameter extraction.

Split Formulation of the Charge and Current Integral Equations for Arbitrarily Shaped Dielectrics

The low-frequency instability in the method of moments can be resolved by reformulating the underlying integral equations to explicitly include charge as an unknown. A new technique based on this charge and current formulation that reduces the computational complexity, both in time and memory, will be introduced. Considerations when using a charge and current technique include the practical problems of basis functions and the evaluation of the singular surface integrals, well as more theoretical problems such as whether the continuity equation still holds. The technique is implemented using a simple set of basis and testing functions and has been verified quantitatively, checking the charge and current distributions against analytic solutions where existing, as well as qualitatively against the boundary conditions for the scattered field. These results show the method to be both stable and accurate.

Improving Slow-start based probing mechanisms for flow adaptation after handovers

In a heterogeneous mobile networking environment, when a mobile user needs to perform a handover and switch to a new network, besides all the mobility management procedures to keep its connections alive and divert its on-going flows to the new location, the transmission rate of these flows should also be correctly adapted to match the conditions of the new network. One of the common practices is to let the flows to go through a fresh new Slow-start (or its variants) phase to probe along the path between the new location of the mobile user and its correspondent for the appropriate rate. Unfortunately, in many situations packets from these flows will be leaked into the new path at the old rate already before the Slow-start probing is performed, which not only interferes with the probing process but also disturbs the new network and affects the cross traffic along the new path. In this paper we propose to introduce some intelligence into the network, more specifically onto the mobility agent, so that this packet leaking issue is well taken care of before and during the Slow-start probing. On the other hand, with the introduced intelligence the probing process can also be improved in the sense of promptness, speed and data throughput, and these improvements are clearly demonstrated by simulation results.

A study of the effect of the top patch in rectangular dual patch microstrip antennas

A moment method procedure is used to solve the integral equations describing rectangular dual patch microstrip antennas. After checking the accuracy of the procedure by comparing calculated with measured results, it is used to analyse theoretically the effect of the top patch, especially the variation of the impedance curve and the gain of the configuration with two parameters, the distance between the two patches and the size of the top patch. Finally, a first attempt is made to outline a technique for the determination of these two parameters in order to obtain a maximum bandwidth.RésuméLa méthode des moments est utilisée pour résoudre l’équation intégrale d’une antenne microruban rectangulaire à plaques empilées. Après le contrôle de la précision de la procédure par une comparaison des résultats calculés et des résultats mesurés, la procédure est utilisée pour analyser théoriquement l’effet de la plaque supérieure, en particulier la variation de l’impédance et du gain de la configuration avec deux paramètres, la distance entre les deux plaques et la dimension de la plaque supérieure. Finalement, un premier effort est fait pour construire une technique pour déterminer la valeur des deux paramètres afin d’obtenir la bande la plus large.

A comparison of implementations of a combined charge and current formulation of the method of moments

The Method of Moments (MoM, also referred to as the Boundary Element Method outside of electromagnetics [1]) is a well-known and oft-used technique for complex electromagnetics simulations based on a numerical approach to Maxwells equations. Results can be very accurate, but as it requires solving a dense matrix equation it is very computationally intensive. MoM has traditionally been used for narrow-band antenna simulations, but thanks to advances in processing power and in the technique itself, it is slowly becoming feasible to perform simulations of electrically large objects and of diverse scenarios over a broad frequency band.

Computing the influence of wind turbines on RF systems taking into account terrain

The main effects of wind turbines on radars have been described in previous papers [1–3]. They include shadowing by the larger parts and generation of false echoes. Previous studies have been focusing on the influence of moving objects, such as wind turbines, on aeronautical and maritime radars, usually working in the L/S band and in the X-band respectively. Here, we will not only take into account terrain properties, but also compute systems very close to each other and at lower frequencies, where UTD is not valid any more. Also, particular attention will be paid to the computations of systems that are in the near-field of the antenna.

Bridging the Gap between Mathematical Traffic Models and Operational Parameters

Over the past decade and more, several accurate but complex traffic models have been developed by just as many researchers. Among the most accurate models are those based on mathematical principles that are able to model the multifractal nature of network traffic. Unfortunately these models are hardly even usable by network engineers because they lack a connection to operational parameters that can easily be estimated based on the knowledge of the network. In this paper, we try to bridge the gap between high-level parameters describing the concerning network and the inputs the modern traffic models need to generate artificial traffic. We first build a model of the behavior of TCP over the duration of a flow, and we then approximate this behavior based on simple parameters such as for example the packet transmission time and the RTT.

Queuing analysis of multi-hop CSMA/CA wireless networks handling many traffic flows

In this paper a semi-analytical model for the evaluation of the queuing performance in CSMA-CA multi-hop wireless networks with arbitrary topology is proposed. The intrinsic fairness of single-hop wireless networks can be used to model the network as a processor sharing system with multiple queues, one for each node, and a fair scheduling regime. In the multi-hop scenario, the queuing performance depends on the access protocol, the network topology and the traffic profiles on all links. In a previous paper a model is introduced which allows the calculation of the saturation throughput of nodes in multi-hop CSMA-CA networks with RTS/CTS enabled access. A simplified version of the method is used to estimate the resources of the servers in a many-sources large deviations analysis of a queuing system with multiple coupled servers. The queuing performance of wireless networks can be evaluated by mapping the different servers to the nodes in the network. The cumulative complementary distribution function of the buffer occupation of a node for different topologies is calculated by a novel method and compared to the results of an event-driven simulation with the same settings. A good fit between the semi-analytical model and the simulation is obtained.

A plane-polar near-field to far-field transformation with the fast Hankel transform method

It is known that a direct radial integration, used to compute the far-field from uniformly spaced plane-polar near-field measurements requires the evaluation of a large amount of Bessel functions and hence CPU time. Up to 1985 only unequally spaced fast Hankel algorithms were available. Hansen [3] developed an algorithm that was usable for equally spaced measurements points, but only for order zero. His theory is generalised in this paper and applied to a plane-polar near-field to far-field transformation.RésuméIl a été établi que les transformations de champs proches plans-polaires requièrent l’évaluation de maintes fonctions de Bessel et donc un temps de calcul considérable. Jusqu’en 1985 les algorithmes rapides de Hankel se limitaient à des points espacés de manière inégale. Hansen [3] a développé un algorithme qui se prêtait à des points de mesures uniformément espacés. mais pour l’ordre zéro seulement. Sa théorie est généralisée dans cet article et appliquée à une transformation de champs plans-polaires proches en champs lointains.