Knowledge Of Topology Research Articles

Localization of Noise in Biochemical Networks.

Noise, or uncertainty in biochemical networks, has become an important aspect of many biological problems. Noise can arise and propagate from external factors and probabilistic chemical reactions occurring in small cellular compartments. For species survival, it is important to regulate such uncertainties in executing vital cell functions. Regulated noise can improve adaptability, whereas uncontrolled noise can cause diseases. Simulation can provide a detailed analysis of uncertainties, but parameters such as rate constants and initial conditions are usually unknown. A general understanding of noise dynamics from the perspective of network structure is highly desirable. In this study, we extended the previously developed law of localization for characterizing noise in terms of (co)variances and developed noise localization theory. With linear noise approximation, we can expand a biochemical network into an extended set of differential equations representing a fictitious network for pseudo-components consisting of variances and covariances, together with chemical species. Through localization analysis, perturbation responses at the steady state of pseudo-components can be summarized into a sensitivity matrix that only requires knowledge of network topology. Our work allows identification of buffering structures at the level of species, variances, and covariances and can provide insights into noise flow under non-steady-state conditions in the form of a pseudo-chemical reaction. We tested noise localization in various systems, and here we discuss its implications and potential applications. Results show that this theory is potentially applicable in discriminating models, scanning network topologies with interesting noise behavior, and designing and perturbing networks with the desired response.

Network intrusion detection system for DDoS attacks in ICS using deep autoencoders

Anomaly detection in industrial control and cyber-physical systems has gained much attention over the past years due to the increasing modernisation and exposure of industrial environments. Current dangers to the connected industry include the theft of industrial intellectual property, denial of service, or the compromise of cloud components; all of which might result in a cyber-attack across the operational network. However, most scientific work employs device logs, which necessitate substantial understanding and preprocessing before they can be used in anomaly detection. In this paper, we propose a network intrusion detection system (NIDS) architecture based on a deep autoencoder trained on network flow data, which has the advantage of not requiring prior knowledge of the network topology or its underlying architecture. Experimental results show that the proposed model can detect anomalies, caused by distributed denial of service attacks, providing a high detection rate and low false alarms, outperforming the state-of-the-art and a baseline model in an unsupervised learning environment. Furthermore, the deep autoencoder model can detect abnormal behaviour in legitimate devices after an attack. We also demonstrate the suitability of the proposed NIDS in a real industrial plant from the alimentary sector, analysing the false positive rate and the viability of the data generation, filtering and preprocessing procedure for a near real time scenario. The suggested NIDS architecture is a low-cost solution that uses only fifteen network-based features, requires minimal processing, operates in unsupervised mode, and is straightforward to deploy in real-world scenarios.

Flocking motion of multi-agent systems: A proximity digraph case

We investigate the coordinated motion of a multi-agent system with heterogeneous distance-dependent communication constraints. In this setup, the underlying interaction network is dynamic since edges appear or disappear as the agents navigate their workspace. Inspired by the gradient-descent method, we provide a distributed controller which preserves the position-dependent communication network connectivity properties. We use a distributed connectivity measure based on the entries of the first-left eigenvector of the network's associated Laplacian matrix to provide the agents with local knowledge of the overall network topology and reveal its dynamics properties. We illustrate our result with a numerical simulation.

Structure-Enriched Topology Learning For Cross-Domain Multi-Person Pose Estimation

Human pose estimation has been widely studied with much focus on supervised learning. However, in real applications, a pretrained pose estimation model usually needs be adapted to a novel domain without labels or with sparse labels. Existing domain adaptation methods cannot well deal with it since poses have flexible topological structures and need fine-grained local features. Aiming at the characteristics of human pose, we propose a novel domain adaptation method for multi-person pose estimation (MPPE) to alleviate the human-level shift. Firstly, the training samples of human poses are clustered into groups according to the posture similarity. Within the clustered space, we conduct three adaptation modules: Cross-Attentive Feature Alignment (CAFA), Intra-domain Structure Adaptation (ISA) and Adaptive Human-Topology Adaptation (AHTA). The CAFA adopts a bidirectional spatial attention mechanism to explore fine-grained local feature correlation between two humans, and thus to adaptively aggregate consistent features for adaptation. ISA only works in semi-supervised domain adaptation (SSDA) to exploit semantic relationship of corresponding keypoints for reducing the intra-domain bias. Importantly, we creatively propose an AHTA to enrich human topological knowledge for reducing the inter-domain discrepancy. Specifically, the pose structure and the cross-instance topological relations are modeled via graph networks. This flexible topology learning benefits the occluded or extreme pose inference. Extensive experiments are conducted on two popular benchmarks and additional two challenging datasets. Results demonstrate the competency of our method, which works in unsupervised or semi-supervised modes, compared with the existing supervised approaches.

Flowthrough Centrality: A Stable Node Centrality Measure

This paper introduces flowthrough centrality, a node centrality measure determined from the hierarchical maximum concurrent flow problem (HMCFP). Based upon the extent to which a node is acting as a hub within a network, this centrality measure is defined to be the fraction of the flow passing through the node to the total flow capacity of the node. Flowthrough centrality is compared to the commonly-used centralities of closeness centrality, betweenness centrality, and flow betweenness centrality, as well as to stable betweenness centrality to measure the stability (i.e., accuracy) of the centralities when knowledge of the network topology is incomplete or in transition. Perturbations do not alter the flowthrough centrality values of nodes that are based upon flow as much as they do other types of centrality values that are based upon geodesics. The flowthrough centrality measure overcomes the problem of overstating or understating the roles that significant actors play in social networks. The flowthrough centrality is canonical in that it is determined from a natural, realized flow universally applicable to all networks.

DC-Side Impedance Estimation of a Modular Multilevel Converter Through System Identification of a Partially Black-Boxed Control System

The stability of a power electronicssystem can be assessed by means of the impedance-based stability criterion. Impedance modeling is a useful tool to analyze the effect of different circuit parameters and control schemes on the behavior of a converter. Modeling the input impedance of a power electronics converter is often successful when having full knowledge of the converter topology, the circuit parameters, and the parameters and implementation of the control system. However, due to the proprietary nature of voltage source converter-based high voltage direct current systems, their exact control structure is often concealed. This complicates the calculation of the impedance of a modular multilevel converter, known for its complex internal dynamics. This paper proposes a method to estimate the impedance of a modular multilevel converter with partially black-boxed converter control. A discussion on partitioning the control system into open and closed parts is made, and the results are verified with simulations in time and frequency domains.

Power-Balancing Dual-Port Grid-Forming Power Converter Control for Renewable Integration and Hybrid AC/DC Power Systems

In this work, we investigate grid-forming (GFM) control for dc/ac power converters in emerging power systems that contain ac and dc networks, renewable generation, and conventional generation. We propose a change in control paradigm to a universal dual-port GFM control strategy that simultaneously forms the converter ac and dc voltage (i.e., dual-port GFM), unifies standard functions of grid-following (GFL) and GFM (e.g., primary frequency control, maximal power point (MPP) tracking) in a single universal controller, and is backward compatible with conventional machine-based generation. Notably, in contrast to state-of-the-art control architectures that use a mix of grid-forming and grid-following control, dual-port GFM control can be used independently of the converter power source or network configuration. Our main contributions are stability conditions that cover emerging hybrid ac/dc networks as well as machines and converters with and without controlled power source, which only require partial knowledge of the network topology. Finally, a detailed case study is used to illustrate and validate the results.

Extended real-time voltage instability identification method based on synchronized phasor measurements

This paper presents an extended adaptive approach designed to accurately estimate the Thévenin equivalent parameters using phasor measurements at a given bus for measurements lying in any of four quadrants of the PQ-plane. The improvement is achieved by using a new condition to properly update the estimated parameters after an initial guess. Based on an adaptive philosophy, the proposed approach can correctly account for the intrinsic nonlinearities of a power system, can provide a real-time estimation of Thévenin parameters, and does not require network topology knowledge. The method is validated using the Kundur 2-area system, showing estimation improvements compared to the current adaptive approach and the classical recursive least-squares method. The proposed approach is able to estimate both sides of the system with respect to the measurement bus. In addition, a data-driven voltage stability index is developed. To illustrate the performance of the proposed approach in a larger power system, a voltage stability assessment is carried out on the IEEE 39-bus system, considering the action of overexcitation limiters of generators and nonlinear loads. The proposed approach is suitable for applications that require an accurate Thévenin equivalent estimation in real-time, such as for voltage stability assessment. The new approach provides a reliable tool for the system operators to make proper and timely decisions.

Topology Inference With Multivariate Cumulants: The Möbius Inference Algorithm

Many tasks regarding the monitoring, management, and design of communication networks rely on knowledge of the routing topology. However, the standard approach to topology mapping—namely, active probing with traceroutes—relies on cooperation from increasingly non-cooperative routers, leading to missing information. Network tomography, which uses end-to-end measurements of additive link metrics (like delays or log packet loss rates) across monitor paths, is a possible remedy. Network tomography does not require that routers cooperate with traceroute probes, and it has already been used to infer the structure of multicast trees. This paper goes a step further. We provide a tomographic method to infer the underlying routing topology of an arbitrary set of monitor paths using the joint distribution of end-to-end measurements, without making any assumptions on routing behavior. Our approach, called the Möbius Inference Algorithm (MIA), uses cumulants of this distribution to quantify high-order interactions among monitor paths, and it applies Möbius inversion to “disentangle” these interactions. In addition to MIA, we provide a more practical variant called Sparse Möbius Inference, which uses various sparsity heuristics to reduce the number and order of cumulants required to be estimated. We show the viability of our approach using synthetic case studies based on real-world ISP topologies.

Connectivity-preserving distributed algorithms for removing links in directed networks

AbstractThis article considers the link removal problem in a strongly connected directed network with the goal of minimizing the dominant eigenvalue of the network’s adjacency matrix while maintaining its strong connectivity. Due to the complexity of the problem, this article focuses on computing a suboptimal solution. Furthermore, it is assumed that the knowledge of the overall network topology is not available. This calls for distributed algorithms which rely solely on the local information available to each individual node and information exchange between each node and its neighbors. Two different strategies based on matrix perturbation analysis are presented, namely simultaneous and iterative link removal strategies. Key ingredients in implementing both strategies include novel distributed algorithms for estimating the dominant eigenvectors of an adjacency matrix and for verifying strong connectivity of a directed network under link removal. It is shown via numerical simulations on different type of networks that in general the iterative link removal strategy yields a better suboptimal solution. However, it comes at a price of higher communication cost in comparison to the simultaneous link removal strategy.

Resource-Efficient Platooning Control of Connected Automated Vehicles Over VANETs

This paper is concerned with distributed cooperative longitudinal platooning control of connected automated vehicles over resource-constrained vehicular ad-hoc networks. First, a general platoon-based control framework is constructed, which simultaneously accounts for unknown leader control input, external disturbances, intermittent data samplings and transmissions, various spacing policies, and different vehicle-to-vehicle communication topologies. To alleviate excessive occupancy of the limited communication resources, an adaptive Zeno-free event-triggered communication mechanism is developed to determine when the sampling and transmission actions of vehicular data should be performed among interacting vehicles. Then, a scalable co-design approach for the desired adaptive platoon controller as well as event triggering condition is developed. One promising feature of the proposed platooning control strategy is that it is independent on the platoon scale or any global knowledge of the associated communication topology, and thus implementable on each vehicle in a scalable manner. Finally, several numerical case studies are conducted to verify the efficacy of the derived approach.

A parallel based evolutionary algorithm with primary-auxiliary knowledge

The development of hybrid algorithms or the application of multiple strategies is one of the focal points for research on improving evolutionary algorithms. However, since most of the evolutionary equations of different algorithms can be transformed into each other, it is difficult to change the established properties of the given algorithm through a hybrid algorithm based on a mixture of evolutionary equations. In addition, multi-strategy methods tend to adopt the best strategy for the current local domain through greedy strategies in the solution process, which does not ensure validity in the global domain. Recently, Federated Learning has achieved remarkable results in machine learning, where the idea of model independence, parallelism and data sharing can essentially compensate for the weaknesses of hybrid and multi-strategy algorithms. Inspired by the idea of Federated Learning, this paper proposes an evolutionary algorithm named as parallel based Evolutionary Algorithm with primary-auxiliary knowledge. Specifically, a Spark-based primary-auxiliary knowledge model is developed, with different evolutionary algorithms used on each parallel sub-model. Then, an effective topological knowledge (individual) migration method is devised, which enables the primary knowledge model to learn the best knowledge from different auxiliary knowledge models through a topological structure. In this way, the best knowledge on the auxiliary knowledge models can be transferred to the primary knowledge model. Through a test conducted on the CEC2013 test set, it can be found out that the proposed algorithm clearly outperforms the 10 algorithms compared, which demonstrates the excellent performance of our proposed parallel based evolutionary algorithm with primary-auxiliary knowledge.

EEG-GAT: Graph Attention Networks for Classification of Electroencephalogram (EEG) Signals.

Graph neural networks (GNN) are an emerging framework in the deep learning community. In most GNN applications, the graph topology of data samples is provided in the dataset. Specifically, the graph shift operator (GSO), which could be adjacency, graph Laplacian, or their normalizations, is known a priori. However we often have no knowledge of the grand-truth graph topology underlying real-world datasets. One example of this is to extract subject-invariant features from physiological electroencephalogram (EEG) to predict a cognitive task. Previous methods use electrode sites to represent a node in the graph and connect them in various ways to hand-engineer a GSO e.g., i) each pair of electrode sites is connected to form a complete graph, ii) a specific number of electrode sites are connected to form a k-nearest neighbor graph, iii) each pair of electrode site is connected only if the Euclidean distance is within a heuristic threshold. In this paper, we overcome this limitation by parameterizing the GSO using a multi-head attention mechanism to explore the functional neural connectivity subject to a cognitive task between different electrode sites, and simultaneously learn the unsupervised graph topology in conjunction with the parameters of graph convolutional kernels.

Arc fault detection and identification via non-intrusive current disaggregation

The series arc faults causing increasing incidences of household fires are difficult to be detected due to the low fault current amplitude. The existing research on arc fault detection is at the power socket level with simple appliances circumstances, but hardly on the whole residential household. As series arc fault current waveform holds unique information of downstream appliance in operation, this paper proposes an arc fault detection and identification method via supervised non-intrusive current disaggregation for low-voltage consumers in the service entrance. In the proposed method, a NILM module initially identifies the appliances in operation in the residential household while an arc fault detection module screens the simultaneous total current signal for an arc fault. In case of arc fault detected, an arc current disaggregation module will find the best-fit arc fault waveform of one or more specific appliances by total current harmonic disaggregation. Then combining the prior knowledge of household power line topology, the consumer can be alerted with the arc resided branch indicated by the operation appliances. By collecting current signals of appliances with and without arc fault, a set of features is extracted from the measurements for total current feature threshold selection and appliance-specific template acquisition. Besides arc fault detection at the service wire entrance, it also infers the specific circuit branch on which arc faults occur. This approach contributes to timely location and elimination of fire hazards. Experiments in consumer scenarios are carried out with an arc fault generator, and the results confirm the effectiveness of the proposed method.

Comparison of different methods of spatial disaggregation of electricity generation and consumption time series

Energy system models involve various input data sets representing the generation, consumption and transport infrastructure of electricity. Especially energy system models with a focus on the transmission grid require time series of electricity feed-in and consumption in a high spatial resolution. In general, there are two approaches to obtain regionalized time series: top-down and bottom-up. In many cases, both methodologies may be combined to aggregate or disaggregate input data. Furthermore, there exist various approaches to assign regionalized feed-in of renewable energy sources and electrical load to the model’s grid connection points. The variety in the regionalization process leads to significant differences on a regional scope, even if global values are the same.We develop a Methodology to compare regionalization techniques of input data for photovoltaics, wind and electrical load between various models as well as data assignment techniques to the power grid nodes. We further define two invariants to evaluate the outcome of the regionalization process at the NUTS 3 level, one invariant for the annual profiles and one for the installed capacities. This Methodology enabled us to compare different regionalization and assignment workflows using simple parameters, without explicit knowledge of grid topology. Our results show that the resolution of the input data and the use of a top-down or a bottom-up approach are the most determinant factors in the regionalization process.

Secure and ultra-reliable provenance recovery in sparse networks: Strategies and performance bounds

Provenance embedding algorithms are well known for tracking the footprints of information flow in wireless networks. Recently, low-latency provenance embedding algorithms have received traction in vehicular networks owing to strict deadlines on the delivery of packets. While existing low-latency provenance embedding methods focus on reducing the packet delay, they assume a complete graph on the underlying topology due to the mobility of the participating nodes. We identify that the complete graph assumption leads to sub-optimal performance in provenance recovery, especially when the vehicular network is sparse, which is usually observed outside peak-hour traffic conditions. As a result, we propose a two-part approach to design provenance embedding algorithms for sparse vehicular networks. In the first part, we propose secure and practical topology-learning strategies, whereas in the second part, we design provenance embedding algorithms that guarantee ultra-reliability by incorporating the topology knowledge at the destination during the provenance recovery process. Besides the novel idea of using topology knowledge for provenance recovery, a distinguishing feature for achieving ultra-reliability is the use of hash-chains in the packet, which trade communication-overhead of the packet with the complexity-overhead at the destination. We derive tight upper bounds on the performance of our strategies, and show that the derived bounds, when optimized with appropriate constraints, deliver design parameters that outperform existing methods. Finally, we also implement our ideas on OMNeT++ based simulation environment to show that their latency benefits indeed make them suitable for vehicular network applications.

Satellite Formation Flying Control by Using Only Angle Measurements

This paper investigates a practical formation control problem of multi-satellite system, in which a group of small satellites are required to reach and maintain a desired geometric configuration under communication and measurement constraints. Toward this goal, we consider each satellite uses its local coordinate systems and senses its neighboring satellites to obtain angle measurements. A formation control scheme using only the angle measurements is proposed to form the desired configurations. Moreover, the triangular and polygonal configuration cases are developed, in which the sum of interior angles of polygons is utilized to guarantee the stability of the determined final configuration. In particular, motivated by formation rigidity principles which mean that satellites' formation is seen as a multi-point framework which need to be rigid to guarantee the uniqueness of the shape, we propose a formation determination approach to construct a rigid square configuration, and proves that, compared with existing schemes, the absolute position of each satellite and the knowledge of the communication topology are not needed for the proposed formation control law. Moreover, since the angle measurements are scalars and sensed with respect to the local coordinate system, the proposed angle-based formation control law is more robust against the misalignment of measurement coordinates. Simulation results are presented to illustrate the performance of the proposed control laws.

Hyper-Embedder: Learning a Deep Embedder for Self-Supervised Hyperspectral Dimensionality Reduction

Hyperspectral imaging has attracted growing interest among the researchers from the geoscience and remote sensing fields owing to its very rich spectral information. However, the high spectral dimensionality of hyperspectral images tends to suffer from information redundancy. Manifold embedding is a mainstream strategy of nonlinear hyperspectral dimensionality reduction. The sensitivity to the noise and the inflexibility to the out-of-sample problem (i.e., new samples) are the main drawbacks of the manifold embedding based methods. To this end, we propose to learn a deep embedder in a self-supervised fashion for hyperspectral dimensionality reduction, called Hyper-Embedder. Hyper-Embedder effectively reduces the computational complexity and storage-costing compared to conventional embedding models, and improves the robustness against various noises, e.g., spectral variabilities. More significantly, Hyper-Embedder is capable of learning an explicit nonlinear mapping to make a one-to-one match between each original pixel (spectral signature) in the hyperspectral image and its dimension-reduced representation. These low-dimensional representations can be generated and given by existing and classic nonlinear manifold embedding methods. In this paper, we attempt to learn the correspondence or mapping by optimizing a deep regression network. The to-be-developed network can not only capture the local topological knowledge graph of all spectral signatures of hyperspectral data but be applicable to fast prediction and inference of samples from other hyperspectral scenes. The proposed Hyper-Embedder outperforms existing state-of-the-art hyperspectral dimensionality reduction algorithms on two commonly-used hyperspectral datasets, i.e., Indian Pines and Augsburg scenes.

Towards resilient and smart urban road networks: Connectivity restoration via community structure

In the time of global warming, smart and resilient cities need to have the ability to give quick responses to today’s frequent natural disasters. Building a resilient urban transportation system that can timely recover the accessibility of the rescue facilities is vital to the survival chances of citizens. From the perspective of selectively road restoration, we proposed a mixed integer programming model that is based on community hierarchy to provide quick disaster response for damaged urban road networks. This model can efficiently find roads that should give repairing priority to maximize the connectivity in a limited repairing scale. Network community structure has been innovatively introduced into our model, which can remarkably simplify the searching process by extracting critical connectivity information. The model has been tested on square-shaped road networks and Istanbul road networks. The results show that our community structure model gives superior solutions in significantly reduced computational time, and is practical in solving complex real-world large-scale connectivity repairing problems. These findings provide new insights into the understanding of the important role played by mesoscopic topological knowledge of road networks when seeking strategies and solutions for connectivity emergency restoration, which is significant for the development of resilient and sustainable cities. • We propose a community-based MIP approach for connectivity restoration. • The restoration strategies take the advantage of the mesoscopic structure. • The community structure markedly facilitates the optimization process. • The developed algorithms operate on large-scale urban road networks.

Model-free voltage control of active distribution system with PVs using surrogate model-based deep reinforcement learning

Accurate knowledge of the distribution system topology and parameters is required to achieve good voltage control performance, but this is difficult to obtain in practice. This paper proposes a physical-model-free voltage control method based on a surrogate-model-enabled deep reinforcement learning approach. Specifically, a surrogate model is trained in a supervised manner using the recorded limited number of historical data to learn the relationship between the power injections and voltage fluctuations of each node. Then, the deep reinforcement learning algorithm is applied to learn an optimal control strategy from the experiences obtained by continuous interactions with the surrogate model. The proposed method can achieve physical-model-free control of unbalanced distribution network and inform real-time decisions to deal with fast voltage fluctuations caused by the rapid variation of PV generation. Simulation results on an unbalance IEEE 123-bus system show that the proposed method can achieve similar performance as that of perfect physical-model-based approaches while being advantageous over other traditional methods.