基于机动预测的强化学习无人机空中格斗研究

在无人机空中格斗过程中,由于无人机自身状态以及空战态势,敌我双方机动动作及行为策略的选择具有极强的不确定性。针对这个问题,将强化学习方法引入无人机空中格斗过程,建立无人机机动模型及动作集;将空战态势评估函数作为强化学习中的信号函数;采用概率神经网络(PNN)作为对敌机动预测单元;在敌我双方战场信息完全感知条件下,该算法能够不断学习,使无人机通过与环境的交互来掌握其最佳机动行为策略,实现无人机的一对一空中对抗。

基于目标机动预测的空空导弹战术攻击区仿真研究

在空战中往往存在目标会根据当前战场态势提前做出机动的情况,使得传统空空导弹战术攻击区并不能准确地、完全地反映出当前时刻攻防态势,提出了一种基于目标机动预测的战术攻击区仿真方法,通过建立导弹三自由度数学模型、有限外推的目标机动预测模型以及导引律模型进行攻击区仿真分析。结果表明,通过预测目标机动方式对攻击区进行提前修正能够使飞行员及早掌握敌我态势的变化,做出决策判断,尽快完成OODA环,也为后续决策提供了有效的辅助信息。

远程联合打击下防空火力单元机动路线预测研究

红蓝双方演习进攻作战中,为全面夺取战场制空权,通常将蓝方防空火力单元作为首波重点打击目标,因此如何准确预测蓝方防空火力单元活动规律,科学设计联合侦察打击行动具有重要意义。首先,系统分析防空火力单元所采用的机动保障模式,结合防御能力定义给出防御能力计算公式并引入栅格网络模型完成定量计算,基于保障范围约束、阵地资源约束以及机动流程约束,面向实际保障需求,以最短机动路线和最高防御能力为优化目标,建立了适用于地面防空火力单元机动路线的预测模型。其次,针对所构建的非线性模型求解难题,提出了模型线性化转换和求解方法。然后,在最优转移路线的基础上,设计了一种构建机动路线库的算法,从而生成蓝方机动路线的概率预测方案,解析蓝方防空火力单元的机动规律。最后,基于案例仿真,验证了所设计模型和方法的可行性和全局高效性,可实现对蓝方防空火力单元机动路线的有效预测。

基于自注意力机制和CNN-LSTM的空战目标机动轨迹预测

空战目标机动轨迹是有丰富时空特征的多维时间序列,具有高度复杂性和不确定性。针对现阶段轨迹预测运动学模型建立困难、时序预测的方法难以提取时空特征且只能单一的从T到T+1时刻的顺序式训练的问题,文中提出了一种自注意力机制(self-attention,ATT)和卷积神经网络(convolutional neural network,CNN)-长短期记忆网络(long short-term memory,LSTM)结合的模型(CNN-LSTM-ATT)。离线状态下训练模型,获得的最优模型可以实现目标机动轨迹的高精度预测。文中模型与CNN-LSTM、LSTM模型进行单步预测对比分析,具有良好的单步预测和不同过载机动预测的能力。考虑到电磁干扰和复杂环境导致传输数据的误差和缺失,进行了目标轨迹的5步预测,预测结果和评价指标均优于CNN-LSTM、LSTM模型。

路网服务水平影响下的机动车规模预测

本文在“时空消耗”的路网广义容量计算模型的基础上,引入了路网服务水平的概念,并建立了路网服务水平影响下的路网容量计算模型。在此基础上,本文根据公共客运机动车(公交车、出租车)和其他机动车的不同出行特点,对路网容量下各类机动车保有量进行了预测,为实现机动车发展规模的宏观控制提供了依据。

基于支持向量回归的机动车价格指数预测

用R语言构建支持向量回归(SVR)模型进行预测,采用控制变量的方法,改变模型参数(C,L,σ)对数据进行实验,最终得出最优参数组合。应用均方根误差(RMSE)和平均绝对误差(MAE)2个度量标准分析实验结果,发现支持向量回归对机动车价格指数预测精度高,性能良好,适宜对价格指数数据进行预测。

Mechanism,prevention,and control of mining-induced dynamic disasters in underground metal mines in China:Challenges and solutions

Metal mineral resources play an indispensable role in the development of the national economy.Dynamic disasters in underground metal mines seriously threaten mining safety,which are major scientific and technological problems to be solved urgently.In this article,the occurrence status and grand challenges of some typical dynamic disasters involving roof falling,spalling,collapse,large deformation,rockburst,surface subsidence,and water inrush in metal mines in China are systematically presented,the characteristics of mining-induced dynamic disasters are analyzed,the examples of dynamic disasters occurring in some metal mines in China are summarized,the occurrence mechanism,monitoring and early warning methods,and prevention and control techniques of these disasters are highlighted,and some new opinions,suggestions,and solutions are proposed simultaneously.Moreover,some shortcomings in current disaster research are pointed out,and the direction of efforts to improve the prevention and control level of dynamic disasters in China’s metal mines in the future is prospected.The integration of forward-looking key innovative theories and technologies in the abovementioned aspects will greatly enhance the cognitive level of disaster prevention and mitigation in China’s metal mining industry and achieve a significant shift from passive disaster relief to active disaster prevention.

Pose prediction based on dynamic modeling and virtual prototype simulation of shield tunnelling machine

Compared with traditional feedback control,predictive control can eliminate the lag of pose control and avoid the snakelike motion of shield machines.Therefore,a shield pose prediction model was proposed based on dynamic modeling.Firstly,the dynamic equations of shield thrust system were established to clarify the relationship between force and movement of shield machine.Secondly,an analytical model was proposed to predict future multistep pose of the shield machine.Finally,a virtual prototype model was developed to simulate the dynamic behavior of the shield machine and validate the accuracy of the proposed pose prediction method.Results reveal that the model proposed can predict the shield pose with high accuracy,which can provide a decision basis whether for manual or automatic control of shield pose.

综合客运枢纽机动车流量预测方法研究

在综合客运枢纽规划中,机动车交通流量预测是枢纽对外交通组织的核心部分,是有效控制枢纽对城市交通影响的前提基础,然而目前还没有针对枢纽机动车交通流量预测的系统方法。鉴于此,在前人的研究基础上,结合现阶段综合客运枢纽建设特征,充分考虑枢纽内交通换乘、商业及办公后,提出一种较为系统的预测方法。首先分析综合客运枢纽的基本功能和换乘类型,并以此为基础解析了枢纽内机动车交通需求的构成和特征;接着给出了机动车交通流量预测的基本流程,并系统地提出预测模型;最后针对模型中的参数给出了保障模型适用性和准确性的建议与措施。

Dynamic recrystallization behavior of burn resistant titanium alloy Ti-25V-15Cr-0.2Si

Dynamic recrystallization (DRX) behavior in β phase region for the burn resistant titanium alloy Ti?25V?15Cr?0.2Si was investigated with a compression test in the temperature range of 950?1100 °C and the strain rate of 0.001?1 s?1. The results show that deformation mechanism of this alloy in hot deformation is dominated by DRX, and new grains of DRX are evolved by bulging nucleation mechanism as a predominant mechanism. DRX occurs more easily with the decrease of strain rate and the increase of deformation temperature. Grain refinement is achieved due to DRX during the hot deformation at strain rate range of 0.01?0.1 s?1 and temperature range of 950?1050 °C. DRX grain coarsening is observed for the alloy deformed at the higher temperatures of 1100 °C and the lower strain rates of 0.001 s?1. Finally, in order to determine the recrystallized fraction and DRX grain size under different deformation conditions, the prediction models of recrystallization kinetics and recrystallized grain sizes were established.

滑翔导弹末段多约束智能弹道规划

滑翔导弹末段飞行时空复杂度高、不确定性强、约束多,给弹道规划与制导算法带来了较大的建模和求解难度。针对这一问题,同时增大末段机动范围并提高弹道规划效率,本文提出一种利用连续型深度置信神经网络(ConvolutionalDeep Brief Networks,CDBN)预测机动能力、设计经由点状态实现末段多约束智能弹道规划的方法。过程中采用CDBN对机动能力进行在线预测,快速判定经由点状态的可行性,并且通过经由点状态智能设计,实现前后段能量的优化分配,扩大弹道机动包络;通过设计三角函数型弹目视线角实现末段弹道摆动机动,推导机动弹道最优末制导律对视线角进行跟踪,并调节机动频率以满足速度约束。仿真结果表明,CDBN相对BP网络具有更高的机动能力预测精度;本文所提智能弹道规划方法在满足末端速度约束的前提下,可以实现弹道摆动机动并大幅增加飞行包络。弹道规划能够在0.5s内完成,满足工程应用的快速性要求。

NUMERICAL PREDICTION OF AIRCRAFT HYDRAULIC SYSTEM BASED ON THERMAL DYNAMIC ANALYSIS

A mathematical model of principal elements of the aircraft hydraulic system is presented based on the heat transfer theory. The dynamic heat transfer process of the hydraulic oil and the pump shells within an aircraft hydraulic system are analyzed by the difference method. A kind of means for the prediction to variational trends of the aircraft hydraulic system temperature is provided during operation. The numerical prediction and simulation under the operational conditions are presented for ground trial running and the decelerated operation in flight. Computational results show that there is a good coincidence between the experimental data and the numerical predictions.

军用高级教练机嵌入式训练系统设计

军用高级教练机的嵌入式训练是通过在教练机上嵌装训练模拟系统的功能,以模拟战斗机的特性和作战环境,来实施飞行员的飞行训练任务。笔者围绕军用高级教练机嵌入式训练系统设计问题,以实现空战对抗训练为目标,着重研究了战斗机分系统数学建模、基于交互式多模型算法(IMM)的机动预测、嵌入式训练系统的软硬件设计等内容。

基于机载单目视觉的无人机移动目标实时跟踪

传统的无人机主要是通过机载惯导、GPS实现自主跟踪,然而受到飞行环境和成本等因素的限制无法满足情况,考虑到实用性和低成本,论文主要研究使用单目机载相机解决无人机对移动目标的自主跟踪问题。论文提出了基于ORB特征点的八参数全局运动补偿方法,结合TLD算法实现动态场景下的移动目标跟踪;在跟踪控制部分,论文引入目标面积的图像特征,提出了基于图像的视觉伺服控制器,实现图像雅克比矩阵伪逆的快速求解;同时在飞行控制器中使用串级PID姿态控制器,设计跟踪仿真实验,并在真实平台上完成跟踪实验。

Elastodynamic modeling and joint reaction prediction for 3-PRS PKM

To gain a thorough understanding of the load state of parallel kinematic machines(PKMs), a methodology of elastodynamic modeling and joint reaction prediction is proposed. For this purpose, a Sprint Z3 model is used as a case study to illustrate the process of joint reaction analysis. The substructure synthesis method is applied to deriving an analytical elastodynamic model for the 3-PRS PKM device, in which the compliances of limbs and joints are considered. Each limb assembly is modeled as a spatial beam with non-uniform cross-section supported by lumped virtual springs at the centers of revolute and spherical joints. By introducing the deformation compatibility conditions between the limbs and the platform, the governing equations of motion of the system are obtained. After degenerating the governing equations into quasi-static equations, the effects of the gravity on system deflections and joint reactions are investigated with the purpose of providing useful information for the kinematic calibration and component strength calculations as well as structural optimizations of the 3-PRS PKM module. The simulation results indicate that the elastic deformation of the moving platform in the direction of gravity caused by gravity is quite large and cannot be ignored. Meanwhile, the distributions of joint reactions are axisymmetric and position-dependent. It is worthy to note that the proposed elastodynamic modeling method combines the benefits of accuracy of finite element method and concision of analytical method so that it can be used to predict the stiffness characteristics and joint reactions of a PKM throughout its entire workspace in a quick and accurate manner. Moreover, the present model can also be easily applied to evaluating the overall rigidity performance as well as statics of other PKMs with high efficiency after minor modifications.

Seat inventory control methods for Chinese passenger railways

Railway seat inventory control strategies play a crucial role in the growth of profit and train load factor. The railway passenger seat inventory control problem in China was addressed. Chinese passenger railway operation features and seat inventory control practice were analyzed firstly. A dynamic demand forecasting method was introduced to forecast the coming demand in a ticket booking period. By clustering, passengers' historical ticket bookings were used to forecast the demand to come in a ticket booking period with least squares support vector machine. Three seat inventory control methods: non-nested booking limits, nested booking limits and bid-price control, were modeled under a single-fare class. Different seat inventory control methods were compared with the same demand based on ticket booking data of Train T15 from Beijing West to Guangzhou. The result shows that the dynamic non-nested booking limits control method performs the best, which gives railway operators evidence to adjust the remaining capacity in a ticket booking period.

Vector control of permanent magnet synchronous motor based on dynamic matrix control

Aiming at the control problem of strongly nonlinear and coupled permanent magnet synchronous motor(PMSM)oil rig,this paper presents a predictive control method based on dynamic matrix model.In this method,the dynamic matrix algorithm using multistep prediction technique is applied to the speed loop control of the motor vector control.And its control effect is compared with the traditional proportional integral(PI)control of the motor.By comparing the initial dynamic response and the steady-state recovery under load interference of the two methods,it is shown that the dynamic response and the robustness of the motor controlled by the new method is better than that controlled by conventional PI method.And the feasibility of new control in the application of PMSM oil rig is verified.

Prediction of Planing Craft Motion Based on Grey System Theory

In order to minimize the harm caused by the instability of a planing craft, a motion prediction model is essential. This paper analyzed the feasibility of using an MGM（1,N） model in grey system theory to predict planing craft motion and carried out the numerical simulation experiment. According to the characteristics of planing craft motion, a recurrence formula was proposed of the parameter matrix of an MGMfl,N） model. Using this formula, data can be updated in real-time without increasing computational complexity significantly. The results of numerical simulation show that using an MGM（1,N） model to predict planing motion is feasible and useful for prediction. So the method proposed in this study can reflect the planing craft motion mechanism successfully, and has rational and effective functions of forecasting and analyzing trends.

Evaluation and prediction of blast induced ground vibration using support vector machine

We present the application of Support Vector Machine (SVM) for the prediction of blast induced ground vibration by taking into consideration of maximum charge per delay and distance between blast face to monitoring point. To investigate the suitability of this approach, the predictions by SVM have been compared with conventional predictor equations. Blast vibration study was carried out at Magnesite mine of Pithoragarh, India. Total 170 blast vibrations data sets were recorded at different strate-gic and vulnerable locations in and around to mine. Out of 170 data sets, 150 were used for the training of the SVM network as well as to determine site constants of different conventional predictor equations, whereas, 20 new randomly selected data sets were used to compare the prediction capability of SVM network with conventional predictor equations. Results were compared based on Co-efficient of Determination (CoD) and Mean Absolute Error (MAE) between monitored and predicted values of Peak Particle Veloc-ity (PPV). It was found that SVM gives closer values of predicted PPV as compared to conventional predictor equations. The coef-ficient of determination between measured and predicted PPV by SVM was 0.955, whereas it was 0.262, 0.163, 0.337 and 0.232 by USBM, Langefors-Kihlstrom, Ambraseys-Hendron and Bureau of Indian Standard equations, respectively. The MAE for PPV was 11.13 by SVM, whereas it was 0.973, 1.088, 0.939 and 1.292 by USBM, Langefors-Kihlstrom, Ambraseys-Hendron and Bureau of Indian Standard equations respectively.

Aeroengine Performance Parameter Prediction Based on Improved Regularization Extreme Learning Machine

Performance parameter prediction technology is the core research content of aeroengine health management,and more and more machine learning algorithms have been applied in the field.Regularized extreme learning machine(RELM)is one of them.However,the regularization parameter determination of RELM consumes computational resources,which makes it unsuitable in the field of aeroengine performance parameter prediction with a large amount of data.This paper uses the forward and backward segmentation(FBS)algorithms to improve the RELM performance,and introduces an adaptive step size determination method and an improved solution mechanism to obtain a new machine learning algorithm.While maintaining good generalization,the new algorithm is not sensitive to regularization parameters,which greatly saves computing resources.The experimental results on the public data sets prove the above conclusions.Finally,the new algorithm is applied to the prediction of aero-engine performance parameters,and the excellent prediction performance is achieved.