Discovering Latent Variables for the Tasks With Confounders in Multi-Agent Reinforcement Learning

Efficient exploration in complex coordination tasks has been considered a challenging problem in multi-agent reinforcement learning(MARL). It is significantly more difficult for those tasks with latent variables that agents cannot directly observe. However, most of the existing latent variable discovery methods lack a clear representation of latent variables and an effective evaluation of the influence of latent variables on the agent. In this paper, we propose a new MARL algorithm based on the soft actor-critic method for complex continuous control tasks with confounders. It is called the multi-agent soft actor-critic with latent variable(MASAC-LV) algorithm, which uses variational inference theory to infer the compact latent variables representation space from a large amount of offline experience.Besides, we derive the counterfactual policy whose input has no latent variables and quantify the difference between the actual policy and the counterfactual policy via a distance function. This quantified difference is considered an intrinsic motivation that gives additional rewards based on how much the latent variable affects each agent. The proposed algorithm is evaluated on two collaboration tasks with confounders, and the experimental results demonstrate the effectiveness of MASAC-LV compared to other baseline algorithms.

Variable Curvature Modeling Method of Soft Continuum Robots with Constraints

The inherent compliance of continuum robots holds great promise in the fields of soft manipulation and safe human–robot interaction.This compliance reduces the risk of damage to the manipulated object and its surroundings.However,continuum robots possess theoretically infinite degrees of freedom,and this high flexibility usually leads to complex deformations when subjected to external forces and positional constraints.Describing these complex deformations is the main challenge in modeling continuum robots.In this study,we investigated a novel variable curvature modeling method for continuum robots,considering external forces and positional constraints.The robot configuration curve is described using the developed mechanical model,and then the robot is fitted to the curve.A ten-section continuum robot prototype with a length of 1 m was developed in order to validate the model.The feasibility and accuracy of the model were verified by the ability of the robot to reach target points and track complex trajectories with a load.This work was able to serve as a new perspective for the design analysis and motion control of continuum robots.

Influence of varied drought types on soil conservation service within the framework of climate change:insights from the Jinghe River Basin,China

Severe soil erosion and drought are the two main factors affecting the ecological security of the Loess Plateau,China.Investigating the influence of drought on soil conservation service is of great importance to regional environmental protection and sustainable development.However,there is little research on the coupling relationship between them.In this study,focusing on the Jinghe River Basin,China as a case study,we conducted a quantitative evaluation on meteorological,hydrological,and agricultural droughts(represented by the Standardized Precipitation Index(SPI),Standardized Runoff Index(SRI),and Standardized Soil Moisture Index(SSMI),respectively)using the Variable Infiltration Capacity(VIC)model,and quantified the soil conservation service using the Revised Universal Soil Loss Equation(RUSLE)in the historical period(2000-2019)and future period(2026-2060)under two Representative Concentration Pathways(RCPs)(RCP4.5 and RCP8.5).We further examined the influence of the three types of drought on soil conservation service at annual and seasonal scales.The NASA Earth Exchange Global Daily Downscaled Projections(NEX-GDDP)dataset was used to predict and model the hydrometeorological elements in the future period under the RCP4.5 and RCP8.5 scenarios.The results showed that in the historical period,annual-scale meteorological drought exhibited the highest intensity,while seasonal-scale drought was generally weakest in autumn and most severe in summer.Drought intensity of all three types of drought will increase over the next 40 years,with a greater increase under the RCP4.5 scenario than under the RCP8.5 scenario.Furthermore,the intra-annual variation in the drought intensity of the three types of drought becomes smaller under the two future scenarios relative to the historical period(2000-2019).Soil conservation service exhibits a distribution pattern characterized by high levels in the southwest and southeast and lower levels in the north,and this pattern has remained consistent both in the historical and future periods.Over the past 20 years,the intra-annual variation indicated peak soil conservation service in summer and lowest level in winter;the total soil conservation of the Jinghe River Basin displayed an upward trend,with the total soil conservation in 2019 being 1.14 times higher than that in 2000.The most substantial impact on soil conservation service arises from annual-scale meteorological drought,which remains consistent both in the historical and future periods.Additionally,at the seasonal scale,meteorological drought exerts the highest influence on soil conservation service in winter and autumn,particularly under the RCP4.5 and RCP8.5 scenarios.Compared to the historical period,the soil conservation service in the Jinghe River Basin will be significantly more affected by drought in the future period in terms of both the affected area and the magnitude of impact.This study conducted beneficial attempts to evaluate and predict the dynamic characteristics of watershed drought and soil conservation service,as well as the response of soil conservation service to different types of drought.Clarifying the interrelationship between the two is the foundation for achieving sustainable development in a relatively arid and severely eroded area such as the Jinghe River Basin.

Temporally Preserving Latent Variable Models:Offline and Online Training for Reconstruction and Interpretation of Fault Data for Gearbox Condition Monitoring

Latent variable models can effectively determine the condition of essential rotating machinery without needing labeled data.These models analyze vibration data via an unsupervised learning strategy.Temporal preservation is necessary to obtain an informative latent manifold for the fault diagnosis task.In a temporalpreserving context,two approaches exist to develop a condition-monitoring methodology:offline and online.For latent variable models,the available training modes are not different.While many traditional methods use offline training,online training can dynamically adjust the latent manifold,possibly leading to better fault signature extraction from the vibration data.This study explores online training using temporal-preserving latent variable models.Within online training,there are two main methods:one focuses on reconstructing data and the other on interpreting the data components.Both are considered to evaluate how they diagnose faults over time.Using two experimental datasets,the study confirms that models from both training modes can detect changes in machinery health and identify faults even under varying conditions.Importantly,the complementarity of offline and online models is emphasized,reassuring their versatility in fault diagnostics.Understanding the implications of the training approach and the available model formulations is crucial for further research in latent variable modelbased fault diagnostics.

Exploration of the Impact Mechanism of Government Credibility Based on Variable Screening Method

Government credibility is an important asset of contemporary national governance, an important criterion for evaluating government legitimacy, and a key factor in measuring the effectiveness of government governance. In recent years, researchers’ research on government credibility has mostly focused on exploring theories and mechanisms, with little empirical research on this topic. This article intends to apply variable selection models in the field of social statistics to the issue of government credibility, in order to achieve empirical research on government credibility and explore its core influencing factors from a statistical perspective. Specifically, this article intends to use four regression-analysis-based methods and three random-forest-based methods to study the influencing factors of government credibility in various provinces in China, and compare the performance of these seven variable selection methods in different dimensions. The research results show that there are certain differences in simplicity, accuracy, and variable importance ranking among different variable selection methods, which present different importance in the study of government credibility issues. This study provides a methodological reference for variable selection models in the field of social science research, and also offers a multidimensional comparative perspective for analyzing the influencing factors of government credibility.

Modeling of microstructure evolution of AZ80 magnesium alloy during hot working process using a unified internal state variable method

In this paper, a unified internal state variable(ISV) model for predicting microstructure evolution during hot working process of AZ80 magnesium alloy was developed. A novel aspect of the proposed model is that the interactive effects of material hardening, recovery and dynamic recrystallization(DRX) on the characteristic deformation behavior were considered by incorporating the evolution laws of viscoplastic flow, dislocation activities, DRX nucleation and boundary migration in a coupled manner. The model parameters were calibrated based on the experimental data analysis and genetic algorithm(GA) based objective optimization. The predicted flow stress, DRX fraction and average grain size match well with experimental results. The proposed model was embedded in the finite element(FE) software DEFORM-3 D via user defined subroutine to simulate the hot compression and equal channel angular extrusion(ECAE) processes. The heterogeneous microstructure distributions at different deformation zones and the dislocation density evolution with competitive deformation mechanisms were captured.This study can provide a theoretical solution for the hot working problems of magnesium alloy.

MODELING, VALIDATION AND OPTIMAL DESIGN OF THE CLAMPING FORCE CONTROL VALVE USED IN CONTINUOUSLY VARIABLE TRANSMISSION

Associated dynamic performance of the clamping force control valve used in continuously variable transmission （CVT） is optimized. Firstly, the structure and working principle of the valve are analyzed, and then a dynamic model is set up by means of mechanism analysis. For the purpose of checking the validity of the modeling method, a prototype workpiece of the valve is manufactured for comparison test, and its simulation result follows the experimental result quite well. An associated performance index is founded considering the response time, overshoot and saving energy, and five structural parameters are selected to adjust for deriving the optimal associated performance index. The optimization problem is solved by the genetic algorithm （GA） with necessary constraints. Finally, the properties of the optimized valve are compared with those of the prototype workpiece, and the results prove that the dynamic performance indexes of the optimized valve are much better than those of the prototype workpiece.

FeatureMatching Combining Variable Velocity Model with Reverse Optical Flow

The ORB-SLAM2 based on the constant velocity model is difficult to determine the search window of the reprojection of map points when the objects are in variable velocity motion,which leads to a false matching,with an inaccurate pose estimation or failed tracking.To address the challenge above,a new method of feature point matching is proposed in this paper,which combines the variable velocity model with the reverse optical flow method.First,the constant velocity model is extended to a new variable velocity model,and the expanded variable velocity model is used to provide the initial pixel shifting for the reverse optical flow method.Then the search range of feature points is accurately determined according to the results of the reverse optical flow method,thereby improving the accuracy and reliability of feature matching,with strengthened interframe tracking effects.Finally,we tested on TUM data set based on the RGB-D camera.Experimental results show that this method can reduce the probability of tracking failure and improve localization accuracy on SLAM(Simultaneous Localization and Mapping)systems.Compared with the traditional ORB-SLAM2,the test error of this method on each sequence in the TUM data set is significantly reduced,and the root mean square error is only 63.8%of the original system under the optimal condition.

On the Cosmic Evolution of the Quantum Vacuum Using Two Variable G Models and Winterberg’s Thesis

We work within a Winterberg framework where space, i.e., the vacuum, consists of a two component superfluid/super-solid made up of a vast assembly (sea) of positive and negative mass Planck particles, called planckions. These material particles interact indirectly, and have very strong restoring forces keeping them a finite distance apart from each other within their respective species. Because of their mass compensating effect, the vacuum appears massless, charge-less, without pressure, net energy density or entropy. In addition, we consider two varying G models, where G, is Newton’s constant, and G-1, increases with an increase in cosmological time. We argue that there are at least two competing models for the quantum vacuum within such a framework. The first follows a strict extension of Winterberg’s model. This leads to nonsensible results, if G increases, going back in cosmological time, as the length scale inherent in such a model will not scale properly. The second model introduces a different length scale, which does scale properly, but keeps the mass of the Planck particle as, ± the Planck mass. Moreover we establish a connection between ordinary matter, dark matter, and dark energy, where all three mass densities within the Friedman equation must be interpreted as residual vacuum energies, which only surface, once aggregate matter has formed, at relatively low CMB temperatures. The symmetry of the vacuum will be shown to be broken, because of the different scaling laws, beginning with the formation of elementary particles. Much like waves on an ocean where positive and negative planckion mass densities effectively cancel each other out and form a zero vacuum energy density/zero vacuum pressure surface, these positive mass densities are very small perturbations (anomalies) about the mean. This greatly alleviates, i.e., minimizes the cosmological constant problem, a long standing problem associated with the vacuum.

Model Reference Variable Structure Autopilot Design for Homing Missile

To design the control system for some homing missile so that the autopilot can transfer guidance command correctly and be robust to disturbances, such as the measurement noises and parameter variation caused by areodynamic floating. The model reference adaptive control was combined with the variable structure control to design a model reference variable structure (MRVS) control system whose control structure is simple and can be realized easily. The simulation results indicate that MRVS can complete the task of transferring guidance command and suppress the distrubances effectively.

VARIABLE MODEL FOR BUILDING CASE BASE AND ITS APPLICATION

The case variable model is put forward by analyzing the system model and the IDEFO model. The component variables of the machining system are classified into four types, I. E uncontrolled variables, process variables, controlled variables and output variables. The process of building the case base is given. The high-speedcutting data base system is developed based on the presented variable model.

Transition Features from Simplicity-Universality to Complexity-Diversification Under UHNTF

A large unified hybrid network model with a variable speed growth （LUHNM-VSG） is proposed as third model of the unified hybrid network theoretical framework （UHNTF）. A hybrid growth ratio vg of deterministic linking number to random linking number and variable speed growth index a are introduced in it. The main effects of vg and a on topological transition features of the LUHNM-VSC are revealed. For comparison with the other models, we construct a type of the network complexity pyramid with seven levels, in which from the bottom level-1 to the top level-7 of the pyramid simplicity-universality is increasing but complexity-diversity is decreasing. The transition relations between them depend on matching of four hybrid ratios （dr, fd, gr, vg）. Thus the most of network models can be investigated in the unification way via four hybrid ratios （dr, fd, gr, vg）. The LUHNM-VSG as the level-1 of the pyramid is much better and closer to description of real-world networks as well as has potential application.

Estimation Method of State-of-Charge For Lithium-ion Battery Used in Hybrid Electric Vehicles Based on Variable Structure Extended Kalman Filter

Since the main power source of hybrid electric vehicle（HEV） is supplied by the power battery,the predicted performance of power battery,especially the state-of-charge（SOC） estimation has attracted great attention in the area of HEV.However,the value of SOC estimation could not be greatly precise so that the running performance of HEV is greatly affected.A variable structure extended kalman filter（VSEKF）-based estimation method,which could be used to analyze the SOC of lithium-ion battery in the fixed driving condition,is presented.First,the general lower-order battery equivalent circuit model（GLM）,which includes column accumulation model,open circuit voltage model and the SOC output model,is established,and the off-line and online model parameters are calculated with hybrid pulse power characteristics（HPPC） test data.Next,a VSEKF estimation method of SOC,which integrates the ampere-hour（Ah） integration method and the extended Kalman filter（EKF） method,is executed with different adaptive weighting coefficients,which are determined according to the different values of open-circuit voltage obtained in the corresponding charging or discharging processes.According to the experimental analysis,the faster convergence speed and more accurate simulating results could be obtained using the VSEKF method in the running performance of HEV.The error rate of SOC estimation with the VSEKF method is focused in the range of 5% to 10% comparing with the range of 20% to 30% using the EKF method and the Ah integration method.In Summary,the accuracy of the SOC estimation in the lithium-ion battery cell and the pack of lithium-ion battery system,which is obtained utilizing the VSEKF method has been significantly improved comparing with the Ah integration method and the EKF method.The VSEKF method utilizing in the SOC estimation in the lithium-ion pack of HEV can be widely used in practical driving conditions.

Application of Muskingum routing method with variable parameters in ungauged basin

This paper describes a flood routing method applied in an ungauged basin, utilizing the Muskingum model with variable parameters of wave travel time K and weight coefficient of discharge x based on the physical characteristics of the river reach and flood, including the reach slope, length, width, and flood discharge. Three formulas for estimating parameters of wide rectangular, triangular, and parabolic cross sections are proposed. The influence of the flood on channel flow routing parameters is taken into account. The HEC-HMS hydrological model and the geospatial hydrologic analysis module HEC-GeoHMS were used to extract channel or watershed characteristics and to divide sub-basins. In addition, the initial and constant-rate method, user synthetic unit hydrograph method, and exponential recession method were used to estimate runoff volumes, the direct runoff hydrograph, and the baseflow hydrograph, respectively. The Muskingum model with variable parameters was then applied in the Louzigou Basin in Henan Province of China, and of the results, the percentages of flood events with a relative error of peak discharge less than 20% and runoff volume less than 10% are both 100%. They also show that the percentages of flood events with coefficients of determination greater than 0.8 are 83.33%, 91.67%, and 87.5%, respectively, for rectangular, triangular, and parabolic cross sections in 24 flood events. Therefore, this method is applicable to ungauged basins.

Application of Road Information in Ground Moving Target Tracking

A new algorithm is developed to achieve accurate state estimation in ground moving target tracking by means of using road information. It is an adaptive variable structure interacting multiple model estimator with dynamic models modification （DMM VS-IMM for short）. Firstly, road information is employed to modify the target dynamic models used by filter, including modification of state transition matrix and process noise. Secondly, road information is applied to update the model set of a VS-IMM estimator. Predicted state estimation and road information are used to locate the target in the road network on which the model set is updated and finally IMM filtering is implemented. As compared with traditional methods, the accuracy of state estimation is improved for target moving not only on a single road, but also through an intersection. Monte Carlo simulation demonstrates the efficiency and robustness of the proposed algorithm with moderate computational loads.

An Algorithm of the Adaptive Grid and Fuzzy Interacting Multiple Model

This paper studies the algorithm of the adaptive grid and fuzzy interacting multiple model （AGFIMM） for maneuvering target tracking, while focusing on the problems of the fixed structure multiple model （FSMM） algorithm＇s cost-efficiency ratio being not high and the Markov transition probability of the interacting multiple model （IMM） algorithm being difficult to determine exactly. This algorithm realizes the adaptive model set by adaptive grid adjustment, and obtains each model matching degree in the model set by fuzzy logic inference. The simulation results show that the AGFIMM algorithm can effectively improve the accuracy and cost-efficiency ratio of the multiple model algorithm, and as a result is suitable for enineering apolications.

Prediction method of reservoir pore types based on pore shape substitution

Carbonate,tight sandstone,and shale reservoirs have many pore types,and the relationship between the porosity and elastic parameters is extremely discrete due to the complex pore shape.This paper presents a method for predicting reservoir pore types based on pore shape substitution.The pore shape substitution allows for accurately characterizing the changes in the elastic properties of the rock with the changes in pore shape,assuming there are no changes in terms of minerals,porosity,or fl uids.By employing a multiple-porosity variable critical porosity model,the eff ective pore aspect ratio could be inverted from the velocities of the rock.To perform pore shape substitution,we could replace the eff ective pore aspect ratio with another pore aspect ratio or increase/decrease the volume content of diff erent pore shapes.The reservoir pore types could be evaluated by comparing the differences in the reservoir velocities before and after the substitution of the pore shape.The test results pertaining to the theoretical model and the well logging data indicated that the pore shape substitution method could be applied to characterize pore types in terms of separating the eff ects of the pore shapes from the eff ects of the minerals,porosity,or fl uids on the velocities.

DYNAMIC MODEL AND SIMULATION OF A NOVEL ELECTRO-HYDRAULIC FULLY VARIABLE VALVE SYSTEM FOR FOUR-STROKE AUTOMOTIVE ENGINES

In modem four-stroke engine technology, variable valve timing and lift control offers potential benefits for making a high-performance engine. A novel electro-hydraulic fully variable valve train for four-stroke automotive engines is introduced. The construction of the nonlinear mathematic model of the valve train system and its dynamic analysis are also presented. Experimental and simulation results show that the novel electro-hydraulic valve train can achieve fully variable valve timing and lift control. Consequently the engine performance on different loads and speeds will be significantly increased. The technology also permits the elimination of the traditional throttle valve in the gasoline engines and increases engine design flexibility.

Closed-loop identification of systems using hybrid Box–Jenkins structure and its application to PID tuning

The paper describes a closed-loop system identification procedure for hybrid continuous-time Box–Jenkins models and demonstrates how it can be used for IMC based PID controller tuning. An instrumental variable algorithm is used to identify hybrid continuous-time transfer function models of the Box–Jenkins form from discretetime prefiltered data, where the process model is a continuous-time transfer function, while the noise is represented as a discrete-time ARMA process. A novel penalized maximum-likelihood approach is used for estimating the discrete-time ARMA process and a circulatory noise elimination identification method is employed to estimate process model. The input–output data of a process are affected by additive circulatory noise in a closedloop. The noise-free input–output data of the process are obtained using the proposed method by removing these circulatory noise components. The process model can be achieved by using instrumental variable estimation method with prefiltered noise-free input–output data. The performance of the proposed hybrid parameter estimation scheme is evaluated by the Monte Carlo simulation analysis. Simulation results illustrate the efficacy of the proposed procedure. The methodology has been successfully applied in tuning of IMC based flow controller and a practical application demonstrates the applicability of the algorithm.

Design of Model Following Variable Structure Controller for Three-axle Vehicle

An optimal control procedure is developed for the front and rear wheels of a three-axle vehicle moving on a complex typical road based on model following variable structure control strategy. The actual vehicle may be considered as an uncertain system. Cornering stiffness of front and rear wheels and external disturbances are varied in a limited range. The model-following variable structure control method is used to control both front and rear wheels steering operations of the vehicle, so that steering responses of the vehicle follow from those of the reference model. By numerical results obtained from computer simulation, it is demonstrated that the control system model can cope with the effects of parameter perturbations and outside disturbances.