基于图卷积深度强化学习的协同空战机动决策方法

针对多无人机智能协同空战对抗决策问题,提出了一种基于长短期记忆与竞争图卷积深度强化学习的多机协同空战机动对抗决策方法.首先,对多机协同空战对抗问题进行描述;其次,在竞争Q网络中,引入长短期记忆网络用于处理带有强时序相关性的空战信息,接着,搭建图卷积网络作为多机之间的通信基础,提出基于长短期记忆与竞争图卷积深度强化学习算法的协同空战训练框架,并对协同空战决策训练算法进行了设计.二对一空战仿真结果验证了本文所提出的协同智能对抗决策方法的有效性,其具有决策速度快、学习过程稳定的特点以及适应空战环境快速变化下的协同策略学习能力.

国外有人/无人机协同作战研究现状与发展趋势

针对现阶段无人机智能化程度相对有限的问题,分析有人/无人机协同作战。梳理有人/无人机协同作战的概念、特点需求及历史概况;描述美国、俄罗斯、英国等国有人/无人机协同作战的研究现状;结合科技进步与战争形态发展,重点梳理有人/无人机协同作战的关键技术,对有人/无人机协同作战的发展前景进行展望。结果表明,该分析可为我军发展有人/无人机混合编队提供参考。

基于T-S模糊故障树的多态导航系统性能可靠性

考虑复杂多态导航系统在不同故障状态具有不同的性能可靠性,将T-S模糊故障树模型应用到多态导航系统性能分析中,以T-S模糊故障树表示系统的性能变量,映射为系统性能指标值,结合统计学中的期望思想,计算不同层次事件发生故障时系统的性能可靠性,并以典型的全球定位系统(GPS)和惯性定向定位导航系统(INS)相组合的导航系统进行分析建树,求解系统在多故障状态下的性能可靠性指标,并进行了实例分析。结果表明:无论底事件处于轻微还是严重故障状态,对系统性能都会产生影响,中间事件是导航系统的薄弱环节,即便是中间层轻微故障,也会引起系统性能可靠性较大的下降。

凹多边形区域覆盖无人机航迹规划方法

覆盖航迹规划技术对于提高无人机的目标搜索能力和正确完成任务能力具有重要的意义。针对凹多边形区域形状,本文提出一种无人机覆盖算法,该算法旨在协助无人机自主完成无遗漏地覆盖搜索给定区域。首先,描述了如何选取无人机的飞行方向;其次,给出了如何将凹多边形转变为凸多边形的流程,并详细说明无人机覆盖区域的过程;最后,通过仿真示例验证了算法的有效性与正确性。

Dynamic affine formation control of networked under-actuated quad-rotor UAVs with three-dimensional patterns

This paper focuses on the solution to the dynamic affine formation control problem for multiple networked underactuated quad-rotor unmanned aerial vehicles(UAVs)to achieve a configuration that preserves collinearity and ratios of distances for a target configuration.In particular,it is investigated that the quad-rotor UAVs are steered to track a reference linear velocity while maintaining a desired three-dimensional target formation.Firstly,by integrating the properties of the affine transformation and the stress matrix,the design of the target formation is convenient and applicable for various three-dimensional geometric patterns.Secondly,a distributed control method is proposed under a hierarchical framework.By introducing an intermediary control input for each quad-rotor UAV in the position loop,the necessary thrust input and the desired attitude are extracted.In the attitude loop,the desired attitude represented by the unit quaternion is tracked by the designed torque input.Both conditions of linear velocity unavailability and mutual collision avoidance are also tackled.In terms of Lyapunov theory,it is prooved that the overall closed-loop error system is asymptotically stable.Finally,two illustrative examples are simulated to validate the effectiveness of the proposed theoretical results.

联用雷贝拉唑和替普瑞酮治疗慢性胃炎的临床效果及对复发率的影响

目的:探讨联用雷贝拉唑和替普瑞酮治疗慢性胃炎的临床效果及对复发率的影响。方法:选取2018年3月-2020年2月来笔者所在医院就诊的慢性胃炎患者108例,随机分成两组,每组54例,对照组采用雷贝拉唑治疗,观察组联用雷贝拉唑和替普瑞酮治疗,观察分析两组患者临床疗效、临床症状消失时间、治疗前后胃黏膜修复状况、不良反应发生状况及复发率的差异。结果:观察组患者临床疗效为88.89%,明显高于对照组的74.07%(P<0.05)。观察组患者上腹痛、腹胀、反酸消失时间均短于对照组(P<0.05)。观察组患者胃黏膜修复状况较对照组患者更优(P<0.05)。治疗期间,两组患者均未发生血尿常规异常、严重肝肾异常等不良反应。随访发现,观察组患者复发率为6.25%,显著低于对照组的22.50%(P<0.05)。结论:联用雷贝拉唑和替普瑞酮治疗慢性胃炎,临床疗效更好,可促进患者临床症状的早期缓解,促进胃黏膜修复,且不易复发,安全性较高,值得临床推广。

Multiscale feature learning and attention mechanism for infrared and visible image fusion

Current fusion methods for infrared and visible images tend to extract features at a single scale,which results in insufficient detail and incomplete feature preservation.To address these issues,we propose an infrared and visible image fusion network based on a multiscale feature learning and attention mechanism(MsAFusion).A multiscale dilation convolution framework is employed to capture image features across various scales and broaden the perceptual scope.Furthermore,an attention network is introduced to enhance the focus on salient targets in infrared images and detailed textures in visible images.To compensate for information loss during convolution,jump connections are utilized during the image reconstruction phase.The fusion process utilizes a combined loss function consisting of pixel loss and gradient loss for unsupervised fusion of infrared and visible images.Extensive experiments on the dataset of electricity facilities demonstrate that our proposed method outperforms nine state-of-theart methods in terms of visual perception and four objective evaluation metrics.

Segment differential aggregation representation and supervised compensation learning of ConvNets for human action recognition

With more multi-modal data available for visual classification tasks,human action recognition has become an increasingly attractive topic.However,one of the main challenges is to effectively extract complementary features from different modalities for action recognition.In this work,a novel multimodal supervised learning framework based on convolution neural networks(Conv Nets)is proposed to facilitate extracting the compensation features from different modalities for human action recognition.Built on information aggregation mechanism and deep Conv Nets,our recognition framework represents spatial-temporal information from the base modalities by a designed frame difference aggregation spatial-temporal module(FDA-STM),that the networks bridges information from skeleton data through a multimodal supervised compensation block(SCB)to supervise the extraction of compensation features.We evaluate the proposed recognition framework on three human action datasets,including NTU RGB+D 60,NTU RGB+D 120,and PKU-MMD.The results demonstrate that our model with FDA-STM and SCB achieves the state-of-the-art recognition performance on three benchmark datasets.

联合连通有向切换拓扑条件下无人机集群鲁棒编队控制

针对联合连通有向切换拓扑网络条件下存在外界扰动的无人机集群鲁棒时变编队与轨迹跟踪控制问题进行了研究.首先,依据无人机真实飞行状态信息、期望队形信息与需要跟踪的轨迹信息之间的误差,以及联合连通通信拓扑网络条件下能够相互通信无人机期望飞行状态与实际飞行状态之间误差信息设计出了无人机集群系统的编队控制方法.然后通过一类特殊的矩阵分解将集群系统时变编队与轨迹跟踪控制问题转换成联合连通快速时间平均系统的渐近稳定控制问题,进一步提出了系统渐近稳定的充分条件,并通过建立的分段连续Lyapunov泛函进行了证明.最后仿真结果验证了本文所提方法能够实现联合连通有向切换拓扑网络条件下无人机集群的时变编队与轨迹跟踪控制飞行.

2018年无人机领域热点评述

无人机集群和反无人机技术作为未来无人机发展的重要方向,已经成为研究热点,科技创新和相关政策的推动为无人机产业的蓬勃发展提供了持续动力。本文从无人机系统发展新路线、实战化应用、集群自主控制、创新驱动产业等多个角度,总结了2018年无人机领域的科技热点及发展态势。

Affine formation maneuver tracking control of multiple second-order agents with time-varying delays

This paper considers an affine maneuver tracking control problem for leader-follower type second-order multi-agent systems in the presence of time-varying delays, where the interaction topology is directed. Using the property of the affine transformation,this paper gives the sufficient and necessary conditions of achieving the affine localizability and extends it to the second-order condition. Under the(n + 1)-reachable condition of the given n-dimensional nominal formation with n + 1 leaders, a formation of agents can be reshaped in arbitrary dimension by only controlling these leaders. When the neighboring positions and velocities are available, a formation maneuver tracking control protocol with time-varying delays is constructed with the form of linear systems, where the tracking errors of the followers can be specified. Based on Lyapunov-Krasovskii stability theory, sufficient conditions to realize affine maneuvers are proposed and proved, and the unknown control gain matrix can be solved only by four linear matrix inequalities independent of the number of agents. Finally, corresponding simulations are carried out to verify the theoretical results, which show that these followers can track the time-varying references accurately and continuously.

New advances in multiple autonomous aerial robots formation control technology

Autonomous aerial robotics has become a hot direction ofresearch inside the community of robotics and control. Theprimary problem addressed by formation control is to steermultiple aerial robots to form desired geometric patterns and,at the same time, realize desired collective swarming behaviorsin a decentralized or distributed manner. In contrast toground vehicles, aerial robots have the ability to work inthree-dimensional (3D) airspace. Equipped with electric orhydraulic motors, the vertical take-off and landing (VTOL)capability is a typical performance of aerial robots. Formationcontrol technology for such aerial robots is incessantlyspringing up to satisfy the requirements of highly intelligentautonomous systems, which affects both military and civilareas, including missile defense, battlefield surveillance,satellite network construction, fire suppression, power gridinspection, commercial show, etc. [1–5]. Such a problem ofmultiple aerial robots formation control is exceptionallychallenging to analyze if practical constraints such as complexdynamics, motion constraints, and imperfect measurementsare incorporated.

基于增益矩阵的二阶集群系统鲁棒分布式立体编队控制

本文研究解决了在三维空间内针对二阶集群系统的鲁棒分布式编队生成与保持控制问题,所设计的基于增益矩阵的控制方法仅需测量局部坐标系下智能体间的相对位置信息而不依赖于交互通信.首先设计了二阶集群系统的立体编队生成控制策略,通过构造李雅普诺夫函数证明了全局稳定性,并利用凸优化求解得出增益矩阵.所设计的生成控制律对扰动、输入饱和等具有较好的鲁棒性,同时在生成过程中引入了启发式的避碰算法.在此基础上,设计了带有期望距离的编队保持控制策略,使得集群系统能够保持固定大小的队形.然后,将所设计的控制律进行改进,拓展并应用于领航者-跟随者网络下的二阶集群系统立体编队控制.最后,通过相关的仿真实验验证了理论结果的正确性.

A guidance law for UAV autonomous aerial refueling based on the iterative computation method

The rendezvous and formation problem is a significant part for the unmanned aerial vehicle（UAV） autonomous aerial refueling（AAR） technique. It can be divided into two major phases： the long-range guidance phase and the formation phase. In this paper, an iterative computation guidance law（ICGL） is proposed to compute a series of state variables to get the solution of a control variable for a UAV conducting rendezvous with a tanker in AAR. The proposed method can make the control variable converge to zero when the tanker and the UAV receiver come to a formation flight eventually. For the long-range guidance phase, the ICGL divides it into two sub-phases： the correction sub-phase and the guidance sub-phase. The two sub-phases share the same iterative process. As for the formation phase, a velocity coordinate system is created by which control accelerations are designed to make the speed of the UAV consistent with that of the tanker.The simulation results demonstrate that the proposed ICGL is effective and robust against wind disturbance.

Affine transformation based formation maneuvering for discrete-time directed networked systems

For the tracking control problems of affine formation maneuvers with discrete leader-following multi-agent systems, various distributed control schemes based on the directed graphs are designed in this paper. Consider the single and double-integrator models, different types of formation maneuvers can be achieved subject to the states of leaders with the novel characteristic of affine localizability. Meanwhile, the properties of affine transformation and signed Laplacian are introduced, which simplify the structures of controllers with distinct constraints to a great extent. Moreover, the convergence analysis of proposed protocols is proven. Static, constant and time-varying target formations can be tracked successfully. In the end, corresponding numerical simulations are carried out to certify the validity of theoretical results.

Distributed planar formation maneuvering of leader-follower networked systems via a barycentric coordinate-based approach

This paper presents a distributed planar leader-follower formation maneuver control strategy for multi-agent systems with different agent dynamic models.This method is based on the barycentric coordinate-based(BCB)control,which can be performed in the local coordinate frame of each agent with required local measurements.By exploring the properties of BCB Laplacians,a time-varying target formation can be BCB localizable by a sufficient number of leaders uniquely,and this formation is converted from a given nominal formation with geometrical similarity transformation.The proposed control laws can continuously maneuver collective single-and double-integrator agents to achieve a translation,scale,rotation,or even their compositions in various directions.For the formation shape control problem of multi-car systems with/without saturation constraints,the obtained control performance can preserve good robustness.Global stability is also proven by mathematical derivations and verified by numerical simulations.