基于准最大最小模型预测控制的AUV视觉对接

针对六自由度自主式水下机器人(autonomous underwater vehicle, AUV)视觉对接这一重要课题,提出一种基于融合深度信息的改进准最大最小模型预测控制(quasi-min-max model predictive control, QMM-MPC)方法,有效提高复杂水下视觉伺服对接系统性能.首先,针对水下AUV视觉由于能见度低导致深度信息存在不确定性的影响,建立新的六自由度AUV视觉伺服模型;然后,结合AUV运动和图像特征运动的测量数据,设计在线深度估计器,同时提出结合多李雅普诺夫函数的QMM-MPC算法,通过求取凸多面体中各顶点不同上界值,降低传统QMM-MPC算法中单李雅普诺夫函数上界所带来的强保守性;最后,通过仿真验证所提出方法的有效性和优越性.     

基于迭代优化算法的AUV水下运动目标航行参数估计

为了解决自主水下机器人（autonomous underwater vehicle,AUV）对水下运动目标进行实时动态追踪的技术难题,本文将渐消记忆递推最小二乘算法与平方根算法相结合,提出一种迭代优化算法。该算法充分利用渐消记忆递推最小二乘算法的快速收敛性能,利用平方根算法解决迭代过程中的数值不稳定问题。迭代优化算法能够快速解算出运动目标的初始距离、航向角及运动方向,数值收敛时间约为3\min,目标运动速度信息也能够在5\min左右收敛。该算法的收敛时间短、计算速度快,甚至AUV无需进行任何形式的机动即可令其保持悬停,这些优点使本算法适用于AUV水下运动目标追踪的工程实际问题。     

基于智能视觉的车辆事故预警监控系统设计

针对车辆行驶中驾驶员视觉盲区或注意力不集中而造成的交通事故问题,提出采用智能视觉技术对车辆行驶过程中四周的异常物体进行视频监控,对动态视频场景中的运动目标进行检测、识别与实时测距,通过车辆智能视频监控系统最大程度的为驾驶员提供更多预警信息,预防交通事故的发生;文中介绍了车辆智能视觉监控系统的硬件设计方法与软件工作流程,并研究了运动目标检测算法与单目视觉测距算法,通过仿真实验,验证了该智能视觉车辆监控系统对于运动目标进行检测、识别与智能测距判断的实验结果.     

交会对接航天员训练电视图像实时高动态范围渲染

应用基于人眼视觉系统的高动态范围实时渲染技术处理交会对接仿真场景,可以为航天员手控交会对接训练提供高逼真度的电视图像.文中结合交会对接电视图像的特点,分析了交会对接航天员训练电视图像高动态范围实时渲染的关键技术;根据两航天器相对距离,将交会对接过程分为远、中和近3个距离段,采用S-curve算法、自适应对数算法和Reinhard算法分别对每个距离段进行调和映射处理;改进了自适应对数算法和Reinhard算法,并基于GPU处理技术实现仿真电视图像的实时显示.实验结果表明,采用基于距离变化调整调和映射算法的方法处理交会对接仿真电视图像可以高逼真地模拟实录视频场景,为航天员手控交会对接训练提供了视景支持.     

一种基于反光带方案的ARV对接视觉定位算法

针对自主遥控潜水器（ARV）近距离水下对接面临的定位方法效果不佳、定位信息误差较大等问题,设计了一种基于反光带方案的水下视觉定位算法。首先提出一种基于Q学习的混合鲸鱼优化算法（QHWOA）来提高算法的收敛精度和收敛速度,并以中继器对接口外布置的反光带为目标,利用QHWOA算法优化Otsu目标函数分割对接图像;然后提出基于PCA（主成分分析）降维法的关键点提取算法来提取关键轮廓与关键点;最后采用SRPnP算法解算ARV与中继器的相对位姿。通过水下对接实验进行关键点像素提取误差和视觉定位误差的计算,结果表明定位误差满足水下对接的精度要求。该算法能够在ARV水下对接时输出有效定位信息,引导ARV与中继器对接。     

基于ANFIS的多AUV协同定位系统量测异常检测方法

针对异常水声测距信息对多自主水下航行器(Autonomous underwater vehicles, AUV)协同定位系统的不利影响,以及传统故障检测方法在多水声测距信息交替混淆的情况下检测效率低的问题,提出一种基于自适应神经模糊推理系统(Adaptive neuro-fuzzy inference system, ANFIS)的量测异常检测方法.首先,分别建立与各水声测距系统相对应的ANFIS模型;然后,基于自适应容积卡尔曼滤波(Adaptive cubature Kalman filter, ACKF)和马氏距离构造反映量测异常的特征信息作为ANFIS的输入;其次,基于预定义的量测异常信息建立了初始混合数据库以训练ANFIS模型实现对量测异常的在线实时检测与隔离;最后,利用湖水实验数据进行了AUV协同定位仿真验证.实验结果表明该方法可以准确识别异常水声测距信息,与传统故障检测方法相比,误报率(False positive rate, FPR)与漏检率(False negative rate, FNR)均减少70%以上.     

FMM与改进GBNN模型相结合的多AUV实时围捕算法

多自主水下机器人(AUV)实时围捕是一个综合的研究课题,包括联盟生成和目标追捕等阶段.首先,基于快速行进算法(FMM)预估围捕时间,有效形成多AUV的动态围捕联盟;然后,在追捕阶段,AUV需要立即跟踪智能逃逸机器人以防止其逃跑.为了实现这一目标,在GBNN(Glasius biological inspired neural network)模型中使用反比例函数替换指数函数计算神经元连接权值,加入额外的衰减项,并提出两点加快神经元活性传播的改进措施,使其适用于实时追捕路径规划.仿真研究表明,围捕联盟形成机制和反比例权值GBNN模型实时路径规划策略都显示出其优越性.在水下环境的多AUV协作围捕中,所提出的围捕控制算法可以提高围捕效率,减少AUV所花费的追捕距离和逃逸机器人的逃逸距离.     

基于单目视觉的实时测距方法研究

为了利用单目视觉实时监测本车与前方障碍物之间的距离,在比较了现有的几类用于车辆控制的道路深度信息获取方法的基础上,首先研究了较为适用于汽车自动驾驶的几何关系推导法,进而提出了基于单目视觉的实时测距算法.通过试验可知,由于摄像机的俯仰角是影响实时测距算法的关键因素,因此又提出了基于道路边界平行约束条件的实时计算摄像机俯仰角算法.静态实车试验的结果显示,该基于单目视觉的实时测距算法具有较高的准确性,可以满足测距要求,而动态实车试验的结果则显示,此算法还可以满足汽车智能化控制的实时性要求.     

利用FastSLAM框架的多自治水下航行器同时定位与跟踪算法

协同定位是共融机器人研究领域的重要问题.协同定位方案的制定受限于机器人间信息交互的能力.针对长时间通讯中断时多自治水下航行器(AUV)协同定位精度明显下降的问题,借鉴同时定位与制图(SLAM)方法,提出了基于FastSLAM框架的同时定位与跟踪(SLAT)算法.将主AUV视为非合作目标,在从AUV上建立起一个关于主AUV的运动估计器,利用从AUV上声呐传感器实时获取的相对量测信息,在对主AUV运动状态估计的同时,完成对从AUV自定位精度的提升.仿真实验结果表明,在长时间通讯中断发生的条件约束下,相比于传统的航位推算方法,所提出的SLATF1.0和2.0算法能够有效减小定位误差,2.0算法对于探测精度变化等因素的影响具有更好适应性.     

基于机器视觉的机器人测距研究

机器人测距技术是机器人导航、路径规划以及协作编队关键技术之一。基于透视投影几何原理,提出了单目视觉测距模型,实现物体测距。单目视觉测距模型实现了在摄像机与目标平面平行或者有夹角情况下,对目标物体距离的测量。实验结果表明,所提出的测距模型和和所用的算法,不仅避免了复杂的摄像机标定,而且测量速度快,能够实现对目标物体空间距离的精确测量。     

Dual-eyes Vision-based Docking System for Autonomous Underwater Vehicle: an Approach and Experiments

A critical challenge for autonomous underwater vehicles (AUVs) is the docking operation for applications such as sleeping under the mother ship, recharging batteries, transferring data, and new mission downloading. The final stage of docking at a unidirectional docking station requires the AUV to approach while keeping the pose (position and orientation) of the vehicle within an allowable range. The appropriate pose therefore demands a sensor unit and a control system that have high accuracy and robustness against disturbances existing in a real-world underwater environment. This paper presents a vision-based AUV docking system consisting of a 3D model-based matching method and Real-time Multi-step Genetic Algorithm (GA) for real-time estimation of the robot’s relative pose. Experiments using a remotely operated vehicle (ROV) with dual-eye cameras and a separate 3D marker were conducted in a small indoor pool. The experimental results confirmed that the proposed system is able to provide high homing accuracy and robustness against disturbances that influence not only the captured camera images but also the movement of the vehicle. A successful docking operation using stereo vision that is new and novel to the underwater vehicle environment was achieved and thus proved the effectiveness of the proposed system for AUV.     

一种基于信息熵约束的快速FCM聚类水下图像分割算法

智能水下机器人视觉识别系统的使命是快速、准确地处理获得水下目标的相关信息并及时反馈给计算机来指导机器人进行下一步的任务。为了在保证分割质量的前提下实现快速图像分割,结合梯度算子、图像的直方图特征和采样计算,并以图像的相对信息损耗为约束,提出了一种基于熵约束的快速FCM聚类水下图像分割算法,并依据水下图像分割效果和模糊划分的有效性评价指标,详尽研究了新算法中加权指数二的取值规律性。实验结果表明,这种算法能够获得较好的分割质量和时间效率,符合机器人对实时性的需求。     

基于单目视觉的水下机器人管道检测

以单目CCD摄像机为视觉传感器,利用视觉系统测量方法获得水下管道的导航信息,并在此基础上建立了一个用于水下机器人的水下管道检测系统． 按照数据结构的抽象程度,将系统中传递的数据信息分为由低至高4个层次,描述了各层次内容,详细介绍了水下机器人管道检测方法． 为了提高系统的准确性和实时性,采用了动态窗口管道检测方法．在室内实验水池中,以某型号水下机器人为试验载体,进行了多次管道跟踪试验,验证了系统的可行性和有效性．     

AUV-enabled adaptive underwater surveying for optimal data collection

A new adaptive strategy for performing data collection with a sonar-equipped autonomous underwater vehicle (AUV) is proposed. The approach is general in the sense that it is applicable to a wide range of underwater tasks that rely on subsequent processing of side-looking sonar imagery. By intelligently allocating resources and immediately reacting to the data collected in-mission, the proposed approach simultaneously maximizes the information content in the data and decreases overall survey time. These improvements are achieved by adapting the AUV route to prevent portions of the mission area from being either characterized by poor image quality or obscured by shadows caused by sand ripples. The peak correlation of consecutive sonar returns is used as a measure for image quality. To detect the presence of and estimate the orientation of sand ripples, a new innovative algorithm is developed. The components of the overall data-driven path-planning algorithm are purposely constructed to permit fast real-time execution with only minimal AUV onboard processing capabilities. Experimental results based on real sonar data collected at sea are used to demonstrate the promise of the proposed approach.     

Underwater docking of an under-actuated autonomous underwater vehicle: system design and control implementation

Underwater docking greatly facilitates and extends operation of an autonomous underwater vehicle (AUV) without the support of a surface vessel. Robust and accurate control is critically important for docking an AUV into a small underwater funnel-type dock station. In this paper, a docking system with an under-actuated AUV is presented, with special attention paid to control algorithm design and implementation. For an under-actuated AUV, the cross-track error can be controlled only via vehicle heading modulation, so both the cross-track error and heading error have to be constrained to achieve successful docking operations, while the control problem can be even more complicated in practical scenarios with the presence of unknown ocean currents. To cope with the above issues, a control scheme of a three-hierarchy structure of control loops is developed, which has been embedded with online current estimator/compensator and effective control parameter tuning. The current estimator can evaluate both horizontal and vertical current velocity components, based only on the measurement of AUV’s velocity relative to the ground; in contrast, most existing methods use the measurements of both AUV’s velocities respectively relative to the ground and the water column. In addition to numerical simulation, the proposed docking scheme is fully implemented in a prototype AUV using MOOS-IvP architecture. Simulation results show that the current estimator/compensator works well even in the presence of lateral current disturbance. Finally, a series of sea trials are conducted to validate the current estimator/compensator and the whole docking system. The sea trial results show that our control methods can drive the AUV into the dock station effectively and robustly.     

Docking at pool and sea by using active marker in turbid and day/night environment

Nowadays, autonomous underwater vehicle (AUV) is playing an important role in human society in different applications such as inspection of underwater structures (dams, bridges). It has been desired to develop AUVs that can work in a sea with a long period of time for the purpose of retrieving methane hydrate, or rare metal, and so on. To achieve such AUVs, the automatic recharging capability of AUVs under the sea is indispensable and it requires AUVs to dock itself to recharging station autonomously. Therefore, we have developed a stereo-vision-based docking methodology for underwater battery recharging to enable the AUV to continue operations without returning surface vehicle for recharging. Since underwater battery recharging units are supposed to be installed in a deep sea, the deep-sea docking experiments cannot avoid turbidity and low-light environment. In this study, the proposed system with a newly designed active—meaning self-lighting—3D marker has been developed to improve the visibility of the marker from an underwater vehicle, especially in turbid water. Experiments to verify the robustness of the proposed docking approach have been conducted in a simulated pool where the lighting conditions change from day to night. Furthermore, sea docking experiment has also been executed to verify the practicality of the active marker. The experimental results have confirmed the effectiveness of the proposed docking system against turbidity and illumination variation.     

自主水下航行器的回坞导引和入坞控制算法

针对军事侦察和海洋环境监测领域中对自主水下航行器（AUV）水下自主回收能力的需求,研究了AUV自主回收过程中回坞和入坞的导引和控制问题。将水下自主回收过程分为回坞导引和入坞控制两个连续的阶段,其中回坞阶段采用经典的视线（LOS）导引法,使AUV到达回收器正前方的回坞航路点;入坞阶段则采用非线性横向跟踪控制方法,使AUV精确跟踪沿回收器中轴线的入坞直线航路航行并最终进入回收器。采用REMUS AUV的模型参数对水下回收进行了仿真研究,结果表明该方法是有效的,具有良好的工程应用前景。     

考虑通信延迟的多自治水下航行器协同定位算法

自治水下航行器(AUV)协同定位中通信延迟具有常态性. 面对延迟到达的信息, 传统方法一般会有定位精 度或实时性的损失. 针对通信延迟的不利影响, 本文在建立水声探测和通信时延模型的基础上, 以扩展卡尔曼滤波 (EKF)为算法框架, 提出了信息顺序到达和信息出序到达2种协同定位算法, 并以建构面向信息出序情景的算法为 主要创新工作. 在信息顺序到达算法中, 将延迟信息进行序贯处理以减小定位误差. 在信息出序到达算法中, 以信 息出现一步滞后的延迟为背景, 使用出序信息直接对从AUV最新状态估计进行再更新, 信息无损地实时估计运动 状态. 计算机仿真实验结果表明, 本文算法相比于传统的航位推算、整周期滤波、量测丢弃等方法, 具有更高的估计 精度; 相比于数据缓存滤波、重新滤波等方法, 具有强实时性.