基于改进YOLOv5s的机坪特种车辆检测算法研究

机坪特种车辆作为航班保障服务的重要一环,其种类多,形状各异;目前已有的车辆检测算法在识别机坪特种车辆时检测精度不高,在遮挡时无法检测;针对于此问题,提出了一种基于改进YOLOv5s的机坪特种车辆检测算法;为了在机坪特种车辆检测中快速、准确的定位感兴趣区域,在主干网络中融合协同注意力机制;考虑到机坪监控场景下特种车辆尺度差别较大的情况,为了能够增强对不同尺度特种车辆的检测能力,提出了四尺度特征检测网络结构;为了提高检测网络多尺度特征融合能力,结合加权双向特征金字塔结构对网络的Neck部分进行改进;将改进后的算法在自建的机坪特种车辆数据集上进行训练、测试,实验结果表明,与YOLOv5s相比,改进后算法的精确度提升了1.6％,召回率提升了3.5％,平均精度mAP0.5和mAP0.5:0.95分别有2.3％和3.3％的提升。     

舰机协同下多基地声纳阵应召搜潜效能研究

针对传统单基地声纳搜潜难以满足反潜作战要求的问题,提出了一种拖曳声纳与反潜直升机布放浮标阵联合起来形成多基地声纳搜潜的方法。结合舰机实际运动特点,建立舰艇运动模型与直升机布放圆形、方形、三角形被动浮标阵模型。采用Monte Carlo方法,对搜潜概率进行仿真,比较多基地与非多基地声纳的搜潜概率,并分析不同的初始条件对搜潜概率的影响。仿真结果表明,多基地声纳搜潜概率高于非多基地声纳,圆形多基地阵搜潜效能相对最佳,三种阵型的搜潜概率随着初始距离、初始位置分布、航速均值的增大而减小,并能针对不同搜潜要求与实际作战条件选择合适的布放半径和被动浮标布放个数。提出的方法适用于一定距离范围内的应召搜潜,对反潜作战具有一定的军事意义。     

Image-Based Visual PID Control of a Micro Helicopter Using a Stationary Camera

This paper proposes an image-based visual servo control method for a micro helicopter. The helicopter does not have any sensors to measure its position or posture on the body. A stationary camera is placed on the ground and it obtains image features of the helicopter. The differences between current features and given reference features are computed. PID controllers then make control input voltages for helicopter control and they drive the helicopter. The proposed controller can avoid some major difficulties in computer vision such as numerical instability due to image noise or model uncertainties, since the reference is defined in the image frames. An experimental result demonstrates that the proposed controller can keep the helicopter in a stable hover.     

Nonlinear dynamic modeling and control of a small-scale helicopter

A test bench for experimental testing of the attitude control of a small-scale helicopter is constructed. A nonlinear model with 10 states is developed for this experimental setup. The unknown model parameters are estimated using the extended Kalman filter with flight test data of the helicopter operating on the test bench. In this work, it is proved that the nonlinear helicopter dynamic model may be globally feedback linearized using the dynamic feedback linearization technique. In order to satisfy multiple closed-loop performance specifications simultaneously, a controller is proposed by applying the Convex Integrated Design (CID) method to the feedback linearized model. Finally, the controller is tested in simulation demonstrating the closed-loop performance of the proposed controller.     

基于实时灰度Hough变换的超声图像针状物体检测

实时针状物体分割、跟踪技术是图像导引手术治疗过程中的一项关键技术,针分割结果直接关系到手术的成败.以灰度Hough变换为基础,提出了一种在二维超声图像中检测插入软组织中针的自动跟踪技术--实时灰度Hough变换技术.该技术不需要二值化,它由粗-精搜索策略和改进的基于相位编组方法的灰度Hough变换两部分组成.用患者乳腺活检超声图像进行了实验,其结果表明,方法在没有外部硬件加速的情况下,在PC机上就能够实时、准确地检测出活检针,其方向误差为1°左右,位置误差在0.5mm以内,完全可以满足超声图像导引手术治疗与活检的需要.     

无人直升机的姿态增强学习控制设计与验证

针对小型无人直升机的姿态控制问题,考虑到现有基于模型的控制方法对直升机动力学模型的先验信息依赖较大,以及未建模动态系统的影响等问题,设计了一种基于增强学习(RL)的飞行控制算法.仅利用直升机的在线飞行数据,补偿了未建模不确定性的影响.同时为了抑制外界扰动,提高系统的鲁棒性,设计了一种基于误差符号函数积分的鲁棒(RISE)控制算法.将两种算法结合,并利用基于Lyapunov分析的方法,证明了无人机姿态控制误差的半全局渐近收敛.最后在无人直升机飞行控制实验平台上,进行了姿态控制的实时实验验证.实验结果表明,本文提出的控制方法具有良好的控制效果,对系统不确定性和外界风扰具有良好鲁棒性.     

基于强化学习的小型无人直升机有限时间收敛控制设计

针对小型无人直升机精确动力学模型难以获取以及姿态控制易受未知外界风扰影响的问题,设计一种基于强化学习(reinforcement learning,RL)与super twisting相结合的非线性控制算法.利用直升机在线飞行数据,训练执行者-评价者(actor-critic,AC)网络以逼近系统建模不确定部分.为了抑制未知外界风扰,提高系统鲁棒性,同时补偿AC网络逼近误差,设计基于super twisting的鲁棒控制算法.进而,利用Lyapunov稳定性分析方法证明无人直升机姿态误差能在有限时间内收敛到零.最后对所提出的算法进行实验验证,实验结果表明,所提出算法具有良好的控制效果,对系统不确定性和外界扰动具有良好的鲁棒性.     

一种邻居动态调整的粒子群优化算法

为了达到全局寻优能力与寻优速度的平衡,提出一种邻居动态调整的粒子群优化算法。该算法依据粒子的多样性变化和进化状态,实现邻居结构的动态改变。算法引入种群熵评估粒子的多样性,定义粒子邻居扩充因子和局部影响因子来描述粒子的进化状态,并提出邻居扩充与约束策略来控制好粒子的影响力。实验结果表明,该算法具有较强的全局寻优能力和较好的寻优速度。     

Solving push-out conflicts in apron taxiways of airports by a network-based simulation

At busy airports, push-out conflicts occur simultaneously in apron taxiways and affect gate assignment operations in apron control. In current manual operations, it is not easy for apron controllers to consider and solve such conflicts in an exact and efficient manner, due to the uncertainties involved in the operations, the difficulties in efficiently controlling and solving push-out conflicts, and also the fact that too many more important factors should be considered in these real-time operations. At most airports today, a buffer time between the use of gates is introduced in gate assignment operations to reduce the impact of push-out conflicts to the delay of aircraft getting into, or leaving from, a gate. From the standpoint of gate assignment operations in apron control, this study analyzes aircraft ground movements and conflicts in apron taxiways. For coping with increasing air traffic demand and minimizing delays caused by inefficient scheduling of aircraft ground movements, and their effect on the gate assignment operations, a network-based model is proposed to perform simulation to support the analysis of the effect of aircraft ground movements in apron taxiways to gate assignment operations. By the proposed method, the delay characteristics of aircraft ground movements in apron taxiways and the activities on gates can be analyzed and studied in detail. Furthermore, the results of the simulation are applied to solve push-out conflicts and minimize delays to aircraft on gates in apron control. Computational results of experiments are also presented and discussed.     

基于随机森林的直升机飞行状态识别方法

针对直升机飞行状态识别训练样本数据少而导致识别率不高的问题,提出一种基于随机森林的直升机飞行状态识别方法。首先利用去野点、限幅、平滑处理对飞行数据进行预处理,并根据特征参数将飞行状态分为8个小类;然后利用随机森林识别率较高的特点,对每一小类进行随机森林分类器设计;最后利用训练样本训练每个随机森林分类器,并将训练好的随机森林分类器识别直升机全起落飞行状态。以某型直升机实飞数据作为实验数据,将该方法与RBF神经网络法和SVM法进行对比实验,结果表明在小样本情况下该方法识别率有明显提高,识别速度也有所提高,可为直升机寿命预测提供依据。     

Biologically adaptive robust mean shift algorithm with Cauchy predator-prey BBO and space variant resolution for unmanned helicopter formation

Visual tracking technology can provide measurement information for unmanned helicopter formation and thus, more attention is being paid to this research area. We propose a novel mean shift （MS） algorithm that is both adaptive and robust for unmanned helicopter formation and apply it to the leading unmanned helicopter tracking. The movement of an unmanned helicopter is very flexible and changeable, which makes the tracking there of more difficulty than for common targets. In creating an algorithm that can adapt to the acceleration of the unmanned helicopter and estimates both the scale and orientation of the movement changes, we combine the traditional MS with the bio-inspired Cauchy predator-prey biogeography-based optimization （CPPBBO） evolutionary algorithm, and also the space variant resolution （SVR） mechanism of the human visual system （MS-CPPBBO-SVR）. To demonstrate the effectiveness and robustness of the proposed method and justify the importance of the CPPBBO algorithm and SVR mechanism at the same time, a series of comparative exper- iments were carried out. The experimental results of the proposed MS-CPPBBO-SVR method are compared with other competitive tracking methods, such as MS, MS with SVR （MS-SVR）, MS-SVR with several other optimization algorithms, and the robust particle filter algorithm. The experimental results demonstrate that our proposed tracking approach, MS-CPPBBO-SVR, is more adaptive, robust and efficient in target tracking than the other methods.     

Trajectory linearization control for a miniature unmanned helicopter

Trajectory linearization control (TLC) approach is applied to design a nonlinear trajectory tracking controller for a miniature unmanned helicopter. The controller is based on nonlinear model derived from Newton-Euler equations, with collective and cyclic pitches being regarded as actual controls. Via model simplifications, the nonlinear helicopter model can be transformed into a cascaded form including four subsystems. Corresponding with the simplified cascaded structure, the proposed controller is constructed by synthesizing TLC designs for subsystems. Each of them includes designs of a pseudo-inversion and an error regulator. A summarized TLC algorithm for the unmanned helicopter is listed briefly after the detailed controller design process. Simulation results are provided to demonstrate performances of the closed-loop system.     

Attitude estimation and control of a 3-DOF lab helicopter only based on optical flow

This paper proposes an attitude estimation method based on optical flow to solve the attitude tracking control problem for a three degree-of-freedom (3-DOF) lab helicopter. First, the relationship between optical flow and the motion of a general unmanned aerial vehicle is derived from the transformation between the image coordinate frame and the world coordinate frame. Then, an expression for the angular velocity of the 3-DOF helicopter is deduced only based on optical flow, and the attitude information is acquired by solving nonlinear equations. Finally, using visual feedback, a linear quadratic regulation (LQR) controller is designed for hovering and tracking, which consists of a feedforward controller and a LQR state feedback controller. Closed-loop experimental results on the lab helicopter demonstrate the effectiveness of the proposed estimation and control methods.     

基于CMAC和PID复合控制的电动伺服加载算法研究

针对直升机在海上执行搜救、探测任务时,海浪噪声会干扰高度信息的测量,引起直升机高度波动的问题,提出一种有效的用于处理海浪噪声的卡尔曼滤波方案。通过建立海浪模型,分析无线电高度表和垂直加速度计测量信号的特点,给出了融合两种高度传感器信息的海浪滤波算法,并应用于直升机高度控制系统进行检验。仿真结果表明,根据海况调整滤波器参数,用所提出的算法可以有效滤除海浪噪声,给出满意的高度信息。将此滤波方案应用于直升机高度控制系统,可以明显改善飞机高度的稳定效果。     

可见光图像中的高压线缺陷自动诊断方法

研究直升机巡检系统中,基于可见光图像的电力线缺陷诊断方法。对直升机机载系统采集到的可见光图像进行实时分析,自动诊断电力线上可能存在的断股、异物附着缺陷。研究了缺陷在可见光图像中的特征,并提出了一种基于亮度和空间信息的线对象检测方法,自动识别电力线。对线对象进行分析,获取其位置、方向、宽度信息,并进行对象分类。结合缺陷特征和对象与分类信息,设计缺陷诊断流程,诊断断股及异物缺陷。实验证明,该方法对电力线检测准确,能够发现电力线存在的可疑异物和断股缺陷,成功应用于直升机巡检系统中。     

Nonlinear adaptive control system design and experiment for a 3-DOF model helicopter

This paper deals with model following control of a model helicopter with three degree-of-freedom. Since the decoupling matrix is singular, a nonlinear structure algorithm is used to design the controller. Furthermore, since the model dynamics are described linearly by unknown system parameters, a well-known parameter estimation technique is introduced. The integral type of estimation model is proposed here since the use of the derivative type of model cannot obtain the desired estimation result. The experimental results show the effectiveness of the proposed method. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

Hierarchical hybrid modelling and control of an unmanned helicopter

In this paper, we propose a hybrid modelling and control design scheme for an unmanned helicopter. This control structure has a hierarchical form with three layers: the regulation layer, the motion planning layer, and the supervision layer. For each layer, a separate hybrid controller has been developed. Then, a composition operator is adopted to capture the interactions between these layers. The resulting closed-loop system can flexibly command the helicopter to perform different tasks, autonomously. The designed controller is embedded in the avionic system of an unmanned helicopter, and actual flight test results are presented to demonstrate the effectiveness of the proposed control structure.     

直升机反潜模拟中的吊放声纳系统仿真

吊放声纳系统仿真是直升机反潜作战模拟的关键。分析了吊放声纳发现判决方程、各种搜潜方法的共性特征,在此基础上将探测、发现目标的问题逐步转换成易于求解的数学模型。以单机和单潜作战对抗过程为典型实例,研究了仿真系统功能需求及总体结构。讨论了基于水声资料库、Direct Sound技术、实时3D效果控制,实现音频信号的合成和回放的方法。针对一个合理的数据结构对于有效的数据处理的重要性,将空间矩阵法引入吊声数据处理仿真。运行和使用表明,仿真系统为直升机反潜综合战术模拟训练提供了实用的平台。     

Nonlinear sliding sector design for multi‐input systems with application to helicopter control

The ability of helicopters to hover and land vertically has spurred an interesting field of research on the development of autonomous flight for these rotatory wing aircrafts. Linear control theory with gain scheduling, which is based on linearizing the system at the equilibrium points, dominated the helicopter autopilot design. Unlike the linear cascaded autopilot structure used in the existing literature, this paper uses state‐dependent linear like structure, including rate‐limited actuator dynamics, with cascaded autopilot topology. This approach allows nonlinear control laws to be implemented throughout the entire flight envelope, providing satisfactory robustness and stability over the various parameter uncertainties and time delays. The cascaded autopilot topology with nonlinear dynamical equations contains a new sliding sector control (SSC) mechanism which is derived for multi‐input nonlinear dynamical systems. The proposed SSC structure for multi‐input nonlinear systems is used in the inner loop of the cascaded autopilot system where the fastest dynamics are required to be controlled for rapid changes in the helicopter dynamical characteristics which enables one to stabilize the helicopter over a wide range of flight conditions. The proposed cascaded autopilot topology with the new SSC mechanism is tested in simulations to assess its robustness and stability properties. To establish its feasibility, the proposed control method is replaced with a suboptimal control method, namely state‐dependent differential Riccati equation (SDDRE) method, for the inner loop and the results of the proposed control architecture are compared with those of SDDRE method.     

挖掘Web日志降低信息搜寻的时间费用

如何根据用户的行为信息优化站点的设计是一个重要的研究问题．提出了一种新的支持站点设计优化的Web使用挖掘方案．此方案基于Web日志中的搜寻路径统计用户寻找目标花费的平均时间，用以量化Web页面的搜寻费用．在此基础上提出了一种高效的数据挖掘方法，寻找一组能够有效压缩搜寻路径(降低时间费用)的超链接．实验表明，挖掘的结果能够提供许多有用的信息，帮助管理者及时发现站点设计中存在的问题．