State Propagation Research Articles

FaaS execution models for edge applications

In this paper, we address the problem of supporting stateful workflows following a Function-as-a-Service (FaaS) model in edge networks. In particular we focus on the problem of data transfer, which can be a performance bottleneck due to the limited speed of communication links in some edge scenarios and we propose three different schemes: a pure FaaS implementation, StateProp, i.e., propagation of the application state throughout the entire chain of functions, and StateLocal, i.e., a solution where the state is kept local to the workers that run functions and retrieved only as needed. We then extend the proposed schemes to the more general case of applications modeled as Directed Acyclic Graphs (DAGs), which cover a broad range of practical applications, e.g., in the Internet of Things (IoT) area. Our contribution is validated via a prototype implementation. Experiments in emulated conditions show that applying the data locality principle reduces significantly the volume of network traffic required and improves the end-to-end delay performance, especially with local caching on edge nodes and low link speeds.

Trajectory-optimized cluster-based network model for the sphere wake

We propose a novel trajectory-optimized cluster-based network model (tCNM) for nonlinear model order reduction from time-resolved data following Li et al. [“Cluster-based network model,” J. Fluid Mech. 906, A21 (2021)] and improving the accuracy for a given number of centroids. The starting point is k-means++ clustering, which minimizes the representation error of the snapshots by their closest centroids. The dynamics is presented by “flights” between the centroids. The proposed trajectory-optimized clustering aims to reduce the kinematic representation error further by shifting the centroids closer to the snapshot trajectory and refining state propagation with trajectory support points. Thus, curved trajectories are better resolved. The resulting tCNM is demonstrated for the sphere wake for three flow regimes, including the periodic, quasi-periodic, and chaotic dynamics. The representation error of tCNM is five times smaller as compared to the approximation by the closest centroid. Thus, the error is at the same level as proper orthogonal decomposition (POD) of same order. Yet, tCNM has distinct advantages over POD modeling: it is human interpretable by representing dynamics by a handful of coherent structures and their transitions; it shows robust dynamics by design, i.e., stable long-time behavior; and its development is fully automatable, i.e., it does not require tunable auxiliary closure and other models.

Concept of radial slippage propagation triggering self-loosening and optimisation design of novel anti-loosening structures

Self-loosening of bolted joints can occur in a vibration environment, and it may induce bolt fatigue fracture with catastrophic consequences. It is essential to clarify the self-loosening mechanism, based on which novel anti-loosening thread structures can be developed. In this paper, we propose the concept of radial slippage propagation and provide new insights into the self-loosening process. The new theory states that the slippage along the radial direction of the thread surface induces more slippage areas (slippage propagation), and self-loosening occurs due to the dynamic evolution and propagation of contact states on the thread and bearing surfaces with an increase in the number of vibration cycles. Finite element analysis (FEA) was used to validate the propagation process of slippage areas on the thread surface. A novel bolted joint with step thread engagement was developed, which could prevent the occurrence of relative motion of the external and internal threads in the radial direction and thus block slippage propagation. A three-dimensional (3D) finite element model (FEM) of the novel thread structure was established, and a test specimen was manufactured using two special tools. FEA and experiments validated its superior anti-loosening and anti-fatigue performances, and the convenience of installation and removal. Experimental validation of the radial slippage propagation theory and the performance optimisation of the step-thread structure should be performed in the future.

A VFM-based identification method for the dynamic anisotropic plasticity of sheet metals

The anisotropic plasticity of sheet metals at high strain rates is a critical factor that can influence the fabrication and service safety of engineering parts under dynamic loading conditions. The characterization of dynamic anisotropic plasticity using conventional testing methods usually suffers from the limitations resulted from the homogeneous deformation state and one-dimensional wave propagation preconditions. In this work, a highly efficient approach is proposed for the simultaneous identification of the anisotropic yielding and hardening constitutive parameters at high strain rates based on the heterogeneous impact test and the virtual fields method. First, the identification framework that combines the selected anisotropic plasticity constitutive models and the principle of virtual work concerning the dynamic kinematics term is introduced. In order to validate the proposed identification algorithm, simulated virtual impact tests were performed to provide reference parameters and unpolluted inertial acceleration/strain/strain rate field data. These data were then processed to extract the target dynamic anisotropic parameters. The results show that using the double-notched impact configuration and the proposed identification algorithm the selected anisotropic yielding and hardening parameters can be reasonably identified from a single impact test, significantly simplifying the testing procedures compared to conventional dynamic testing methods. Also, the effects of different processing parameters and influential factors on the identification accuracy have been investigated. From the sensitivity studies, the preferred ranges of the processing parameters have been determined and the robustness of the proposed method within these verified.

A lightweight hardware‐efficient recurrent network for video super‐resolution

Recent studies in the field of video super-resolution (VSR) have taken great advantage of neural networks and achieved remarkable performance. However, current neural network-based VSR methods have the following two limitations: (1) sophisticated models usually have high computational overhead and require large caches to store intermediate feature maps and parameters. (2) Bidirectional RNN-based methods cannot perform VSR until a complete video sequence is available because of the back-propagation of hidden states. The above two shortcomings make the existing VSR models hard to achieve real-time VSR. To reduce the computational complexity of the network, the authors propose a lightweight hardware-efficient recurrent neural network model, LERN, which is combined with depthwise separable convolution and pixel unshuffle. An SRA block is also introduced to help the model get better VSR performance. In order to wisely exploit video inter-frame information without causing the back propagation of hidden states, the authors introduce a new recurrent architecture that is optimized for video sequences. Besides, the authors design a quantization algorithm to find the optimal fixed-point representations for each layer. Experimental results demonstrate that LERN can achieve excellent performance with low overhead.

Noisy propagation of Gaussian states in optical media with finite bandwidth

We address propagation and entanglement of Gaussian states in optical media characterised by nontrivial spectral densities. In particular, we consider environments with a finite bandwidth J(omega ) = J_0 left[ theta (omega -Omega ) - theta (omega - Omega - delta )right] , and show that in the low temperature regime Tll Omega ^{-1}: (i) secular terms in the master equation may be neglected; (ii) attenuation (damping) is strongly suppressed; (iii) the overall diffusion process may be described as a Gaussian noise channel with variance depending only on the bandwidth. We find several regimes where propagation is not much detrimental and entanglement may be protected form decoherence.

A medium-scale study of using machine learning fusion to improve TLE prediction precision without external information

This paper presents a machine learning (ML) approach that can improve orbit prediction precision and reduce uncertainties with a properly designed fusion strategy. The ML models are trained based on historical TLE sets and then ML-modifications are generated for the future states and uncertainty predictions. The ML-modification is first interpreted as a pseudo-measurement of the true orbit prediction error and then fused with the conventional orbital state and uncertainty propagation result. The regularized particle filter (PF) with progressive correction and systematic sampling is adopted for uncertainty propagation while using the Simplified General Perturbations #4 (SGP4) model. Comprehensive experiments are conducted and analyzed on over 100 resident space objects (RSOs) in different orbit types. The results demonstrate that the fusion strategy enables the ML approach to work with the advanced PF prediction method. The prediction precision can be significantly improved for the majority of the cases. The prediction accuracy is observed to be improved in some cases but without a clear pattern. The results and analysis also suggest that future studies of a monitoring system facilitating the fusion process would be helpful in practical applications.

An elastic framework for ensemble-based large-scale data assimilation

Prediction of chaotic systems relies on a floating fusion of sensor data (observations) with a numerical model to decide on a good system trajectory and to compensate non-linear feedback effects. Ensemble-based data assimilation (DA) is a major method for this concern depending on propagating an ensemble of perturbed model realizations. In this paper, we develop an elastic, online, fault-tolerant and modular framework called Melissa-DA for large-scale ensemble-based DA. Melissa-DAallows elastic addition or removal of compute resources for state propagation at runtime. Dynamic load balancing based on list scheduling ensures efficient execution. Online processing of the data produced by ensemble members enables to avoid the I/O bottleneck of file-based approaches. Our implementation embeds the PDAF parallel DA engine, enabling the use of various DA methods. Melissa-DAcan support extra ensemble-based DA methods by implementing the transformation of member background states into analysis states. Experiments confirm the excellent scalability of Melissa-DA, propagating 16,384 members for a regional hydrological critical zone assimilation relying on the ParFlow model on a domain with about 4 M grid cells. The same use case was ported to the PDAF state-of-the-art DA framework relying on a MPI approach. A comparison with Melissa-DA at 2500 members on 20,000 cores shows our approach is about 50% faster per assimilation cycle.

Parameters controlling of vortex solitons in nonlocal nonlinear medium with gradually characteristic length

This paper discusses the dynamics of vortex beam in nematic liquid crystals with the tunable characteristic length (CL) of response function. It is found that the transition of the soliton states with different widths during propagations can be controlled by changing the CL under the condition that the nonlocal degree is constant. Furthermore, the transitions propagation of different soliton state is stable and can resist noise. In addition, the numerical results prove that the parameters of vortex solitons, such as rotating periods of angles, orbital angular momentum (OAM), gradient force upon micro-particles all can be controlled conveniently at the desired positions. The controllable density of OAM and gradient force have important academic value and application prospect in the field of optical tweezers and spanner.

General model of counterpropagating continuous-variable entangled states in lossy waveguides

We present a general theoretical approach to model an integrated source of counterpropagating continuous-variable entangled states in lossy waveguides pumped by a classical pulsed source incident from above the waveguide. We use a backward Heisenberg approach to model the generation of the entangled state and then solve the adjoint master equation to model the propagation of the state in the lossy waveguide. We employ a numerical method to implement the Schmidt decomposition method for the biphoton wave function rather than pursuing analytical methods. This approach allows us to model a wide variety of waveguide systems and pump configurations. We apply our model to the nonlinear generation and propagation of continuous-variable entangled states in a coupled resonator optical waveguide under a variety of different pump conditions and derive the optimal pumping conditions for our system.

Intermediate filaments are natural energy conductors in live cells

Two possible mechanisms describing intra- and inter-cell energy transfer in biological systems were analyzed. The first one is based on Davydov vibration soliton (DVS) theory, implying CO vibrational energy transfer along α-helix polypeptides. According to Davydov, a certain vibration of one of the CO groups somehow receives the entirety of ATP hydrolysis energy within an enzyme molecule. Next, dipole–dipole interactions of the CO groups of neighboring amino acid residues should ensure propagation of the DVS along the polypeptide chain, transporting it to the site of catalytic reaction. Strong limitations of this theory when applied to energy transfer in living systems were underlined, accompanied by total lack of experimental evidence of DVS existence.The second, much more viable mechanism, based on electronic excited state (exciton) propagation along individual protein molecules and their assemblies – intermediate filaments (IFs) – was considered and discussed in detail. Excitons in IFs may be generated by photon absorption or by ATP hydrolysis energy transfer to IFs. Infrared (IR) excitons were generated in the latter case, which propagated along IFs, enabling energy transfer within and between cells, and inter-cellular communications. Earlier is has been noted that high-contrast vision of vertebrates is based on photon energy propagation along Müller cell (MC) IFs in the form of excitons, from the inner limiting membrane retinal layer to the outer fragments of cone cells, located in the outer limiting membrane retinal layer. Therefore, MC IFs operate as photon energy guides, transferring excitons from MC to cone cells, and thus communicating external visual information to the retinal cones and the brain. We finally conclude that apparently the mechanism based on the properties of IFs as natural energy guides plays the main role in communications within and between cells of live organisms.

Disorder‐Driven Collapse of Topological Phases in Photonic Topological Insulator

By embedding the position disorder of rods in 2D gyromagnetic photonic crystal, the influences of increasing random disorder on the photonic topological phases (PTPs) in a photonic topological insulator are numerically investigated. When the disorder is small, the PTPs almost have no change. With the increase in disorder, the PTPs at the edge of the original topological bandgap are first affected, and those at center of the bandgap have the best robustness. As disorder increases to a critical value, the PTPs collapse ultimately. During the collapse of the PTPs, disorders can induce the Anderson localization of electromagnetic (EM) wave, and produce localized hot spots in the system. When the disorder is not too large, the hot spots are excited at the edge area near the excitation source, and have little influence on the one‐way propagation of the topological edge state (TES). As disorder increases continuously, the hot spots would penetrate the system deeply and affect the propagation of EM waves significantly, which leads to the collapse of the PTPs and destruction of one‐way propagation of TES. Even so, there are still remaining long‐range and short‐range orders in the system. This research provides a theoretical model and an experiment platform for further studying the relationship between topological phases and disorders.

Role of stochastic noise and generalization error in the time propagation of neural-network quantum states

Neural-network quantum states (NQS) have been shown to be a suitable variational ansatz to simulate out-of-equilibrium dynamics in two-dimensional systems using time-dependent variational Monte Carlo (t-VMC). In particular, stable and accurate time propagation over long time scales has been observed in the square-lattice Heisenberg model using the Restricted Boltzmann machine architecture. However, achieving similar performance in other systems has proven to be more challenging. In this article, we focus on the two-leg Heisenberg ladder driven out of equilibrium by a pulsed excitation as a benchmark system. We demonstrate that unmitigated noise is strongly amplified by the nonlinear equations of motion for the network parameters, which causes numerical instabilities in the time evolution. As a consequence, the achievable accuracy of the simulated dynamics is a result of the interplay between network expressiveness and measures required to remedy these instabilities. We show that stability can be greatly improved by appropriate choice of regularization. This is particularly useful as tuning of the regularization typically imposes no additional computational cost. Inspired by machine learning practice, we propose a validation-set based diagnostic tool to help determining optimal regularization hyperparameters for t-VMC based propagation schemes. For our benchmark, we show that stable and accurate time propagation can be achieved in regimes of sufficiently regularized variational dynamics.

Semiclassical propagation of coherent states and wave packets: Hidden saddles.

Semiclassical methods are extremely important in the subjects of wave-packet and coherent-state dynamics. Unfortunately, these essentially saddle-point approximations are considered nearly impossible to carry out in detail for systems with multiple degrees of freedom due to the difficulties of solving the resulting two-point boundary-value problems. However, recent developments have extended the applicability to a broader range of systems and circumstances. The most important advances are first to generate a set of real reference trajectories using appropriately reduced dimensional spaces of initial conditions, and second to feed that set into a Newton-Raphson search scheme to locate the exposed complex saddle trajectories. The arguments for this approach were based mostly on intuition and numerical verification. In this paper, the methods are put on a firmer theoretical foundation and then extended to incorporate saddles hidden from Newton-Raphson searches initiated with real trajectories. This hidden class of saddles is relevant to tunneling-type processes, but a hidden saddle can sometimes contribute just as much as or more than an exposed one. The distinctions between hidden and exposed saddles clarifies the interpretation of what constitutes tunneling for wave packets and coherent states in the time domain.

Influence of random media on orbital angular momentum quantum states of optical vortex beams

Photonic orbital angular momentum (OAM) carrying orthogonal modes exhibits fascinating nature for information encoding. Here we focus on the propagation of OAM states in random media to investigate the turbulence-induced OAM degradation. We utilize the quadratic approximation of the phase structure function to derive a global continuous expression of OAM probability at the single-photon level. The expression provides a general insight into the energy transition between signal and crosstalk modes in random media. Especially for each crosstalk mode, our expression is accurate in predicting the second-order energy transition. The effectiveness of our model is validated via experimental data and simulations in various random media. We also reveal the universality of entangled optical vortex beams and obtain an expression of concurrence. It is demonstrated that utilizing high-order optical vortex beams can effectively improve the antijamming ability and enhance the propagation distance of entanglement. Our results provide a universal framework for OAM propagation through different random media.

Optimal low-thrust linearized elliptic orbit rendezvous considering the communication window

The optimal low-thrust linearized elliptic orbit rendezvous problems are studied for both continuous thrust and piecewise constant thrust considering the communication window constraint. Based on the analytical state propagation expressions for linear relative motion under on–off constant thrust, the analytical energy-optimal solution using piecewise constant thrust is derived by the Lagrange multiplier method for given switching time points. This solution is further extended to obtain the analytical energy-optimal solution using continuous thrust. For the piecewise constant thrust with free switching time points, the fuel-optimal problem is formulated as a nonlinear programming problem, which is solved by a modified differential evolution algorithm. In addition, the derived analytical energy-optimal solution for given switching time points is used to reduce the number of optimization variables, and it is numerically proven to effectively reduce the optimization time. Several examples are presented to show the effectiveness of the proposed method for the low-thrust orbit rendezvous problem considering the communication window constraint.

A Tightly Coupled LiDAR-Inertial SLAM for Perceptually Degraded Scenes.

Realizing robust six degrees of freedom (6DOF) state estimation and high-performance simultaneous localization and mapping (SLAM) for perceptually degraded scenes (such as underground tunnels, corridors, and roadways) is a challenge in robotics. To solve these problems, we propose a SLAM algorithm based on tightly coupled LiDAR-IMU fusion, which consists of two parts: front end iterative Kalman filtering and back end pose graph optimization. Firstly, on the front end, an iterative Kalman filter is established to construct a tightly coupled LiDAR-Inertial Odometry (LIO). The state propagation process for the a priori position and attitude of a robot, which uses predictions and observations, increases the accuracy of the attitude and enhances the system robustness. Second, on the back end, we deploy a keyframe selection strategy to meet the real-time requirements of large-scale scenes. Moreover, loop detection and ground constraints are added to the tightly coupled framework, thereby further improving the overall accuracy of the 6DOF state estimation. Finally, the performance of the algorithm is verified using a public dataset and the dataset we collected. The experimental results show that for perceptually degraded scenes, compared with existing LiDAR-SLAM algorithms, our proposed algorithm grants the robot higher accuracy, real-time performance and robustness, effectively reducing the cumulative error of the system and ensuring the global consistency of the constructed maps.

Propagation graph estimation from individuals\u2019 time series of observed states

Various things propagate through the medium of individuals. Some individuals follow the others and take the states similar to their states a small number of time steps later. In this paper, we study the problem of estimating the state propagation order of individuals from the real-valued state sequences of all the individuals.We propose a method of constructing a state propagation graph from individuals’ time series of observed states. The propagation order estimated by our proposed method is demonstrated to be significantly more accurate than that by a baseline method (optimal constant delay model) for our synthetic datasets, and also to be consistent with visually recognizable propagation orders for the dataset of Japanese stock price time series and biological cell firing state sequences.

An Enhanced Hybrid Visual-Inertial Odometry System for Indoor Mobile Robot.

As mobile robots are being widely used, accurate localization of the robot counts for the system. Compared with position systems with a single sensor, multi-sensor fusion systems provide better performance and increase the accuracy and robustness. At present, camera and IMU (Inertial Measurement Unit) fusion positioning is extensively studied and many representative Visual–Inertial Odometry (VIO) systems have been produced. Multi-State Constraint Kalman Filter (MSCKF), one of the tightly coupled filtering methods, is characterized by high accuracy and low computational load among typical VIO methods. In the general framework, IMU information is not used after predicting the state and covariance propagation. In this article, we proposed a framework which introduce IMU pre-integration result into MSCKF framework as observation information to improve the system positioning accuracy. Additionally, the system uses the Helmert variance component estimation (HVCE) method to adjust the weight between feature points and pre-integration to further improve the positioning accuracy. Similarly, this article uses the wheel odometer information of the mobile robot to perform zero speed detection, zero-speed update, and pre-integration update to enhance the positioning accuracy of the system. Finally, after experiments carried out in Gazebo simulation environment, public dataset and real scenarios, it is proved that the proposed algorithm has better accuracy results while ensuring real-time performance than existing mainstream algorithms.

Photoluminescence of Cis-Polyacetylene Semiconductor Material

Photoluminescence (PL) is one of the key experimental characterizations of optoelectronic materials, including conjugated polymers (CPs). In this study, a simplified model of an undoped cis-polyacetylene (cis-PA) oligomer was selected and used to explain the mechanism of photoluminescence (PL) of the CPs. Using a combination of the ab initio electronic structure and a time-dependent density matrix methodology, the photo-induced time-dependent excited state dynamics were computed. We explored the phonon-induced relaxation of the photoexcited state for a single oligomer of cis-PA. Here, the dissipative Redfield equation of the motion was used to compute the dissipative excited state dynamics of electronic degrees of freedom. This equation used the nonadiabatic couplings as parameters. The computed excited state dynamics showed that the relaxation rate of the electron is faster than the relaxation rate of the hole. The dissipative excited-state dynamics were combined with radiative recombination channels to predict the PL spectrum. The simulated results showed that the absorption and emission spectra both have a similar transition. The main result is that the computed PL spectrum demonstrates two mechanisms of light emission originating from (i) the inter-band transitions, corresponding to the same range of transition energies as the absorption spectrum and (ii) intra-band transitions not available in the absorption spectra. However, the dissipative Redfield equation of the motion was used to compute the electronic degrees of freedom of the nonadiabatic couplings, which helped to process the time propagation of the excited dynamic state. This excited dynamic state shows that the relaxation rate of the electron is faster than the relaxation rate of the hole, which can be used for improving organic semiconductor materials for photovoltaic and LED applications.