Network Topology and Protocol Design for Efficient Consensus in Sensor Networks
Year: 2017
Type: dissertation
Abstract: In the new era of Internet of Things, complex sensor networks are becoming\ncrucial to link the physical world to the Internet. These sensor networks,\ncomposed by hundreds or thousands of nodes, provide many important services\nto promote a heightened level of awareness about the area of interest\nsuch as in predictive maintenance, intelligent buildings, enhanced security\nsystems, etc. In order to make these services possible, several signal processing\ntasks are needed to support their operation, some widely used examples\nof these tasks are parameter estimation, signal detection and target tracking.\nThese tasks allow to improve the services by inferring missing data, reducing\nsamples noise, etc., at the cost of some collaboration of the network nodes\nthat implies their repeated communication over time. Most of these solutions\nare consensus-based strategies, which have recently attracted a great deal of\nresearch work because of its simplicity. These are in-network algorithms,\nwhere each node only exchanges information with its immediate neighbors\nand these are able to obtain global information as a function of some sensed\ndata. A relevant example is the average consensus algorithm, which its goal\nis to obtain, in a distributed way, the average of the initial data. These algorithms\navoid the need of performing all the computations at one or more sink\nnodes, thus, reducing congestion around them and incrementing the robustness\nof the network.\nIn this dissertation, we focus on improving consensus algorithms in terms\nof different parameters and under different types of communications and network\nconfigurations. Each setting considered requires its own assumptions\nand methodologies, since the convergence conditions for each of them are\nrelated but different in general. In particular, in this work, all of the methodologies\nproposed are based on designing how the underlying communications\nare performed.\nIn static networks, where the asymptotic convergence to the average value\nis easily ensured, a topology optimization can be a priori performed in terms\nof several relevant parameters. In particular, we optimize the network topology\nto make consensus algorithms fast and energy efficient. In this setting\nand for continuous systems, we derive a general framework to minimize several\nenergy related functions under different network and nodes constraints To solve it, we propose a fractional convex-concave optimization problem\nwith different constraints that leads to obtain the optimal topology in terms\nof the energy function considered. As a significant variation of the previous\nresults, we also optimize the network topology in discrete systems. The\ndiscretization of the system introduces a weight matrix and certain step-size\n(related to the discrete increments of time) in the process. We show that if\nthis step-size is small enough, the energy related problems stated before can\nbe still casted as convex-concave fractional problems with the weight matrix\nas a unique optimization variable. As the step-size of the process increases\nin size, a discrete system requires a different approach. To solve it, we aim to\nfind another formulation based on adding a constraint on the connectivity and\nsolving the problem several times (for different values of the step-size). In\naddition, two low-complex methodologies with different computational requirements\nare proposed to a posteriori redesign an existing topology by the\ncollaboration of the network nodes.\nOn the contrary, in time-varying (random) networks, it is needed to guarantee\na minimum accuracy of the algorithm, while maximizing the number of\nsimultaneous exchanges of data to ensure fast convergence. Regarding random\nand asymmetric communications, we propose a novel gossip algorithm,\nwhich is based on the residual information that is generated when an asymmetric\ncommunication is performed. We exploit this information to preserve\nthe summation of the process and accelerate it. Moreover, our proposal is\nuseful in the case of having both unicast and broadcast communications, presenting\nfaster convergence in both schemes than existing approaches in the related\nliterature. When the problem of wireless interferences constraining the\ncommunications is additionally taken into account, we propose a novel and\ncomputationally efficient link scheduling protocol that correctly operates in\nthe presence of secondary interference. Our protocol is easily implementable\nand does not require global knowledge of the network. The main objective\nof this new protocol is to be suitable for a cross-layer scheme in which the\nexecution of the average consensus algorithm is favoured, ensuring necessary\nconditions for its convergence with certain accuracy. Additionally, the\nnumber of simultaneous links is additionally considered in order to make the\nconvergence of the consensus process as fast as possible.
Source: NVA (Nasjonalt vitenarkiv)
Author: César Asensio Marco
Institution: University of Agder
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Keywords: network topology optimization, consensus algorithms, average consensus, sensor networks, Internet of Things, energy efficiency, discrete-time consensus, step size, convergence rate, network protocols, distributed signal processing, in-network processing, topology optimization, networked control systems, distributed detection, distributed tracking
Open Access status: green
OpenAlex ID: W2594889901
