智能停车辅助系统的系统配置--外文资料翻译

毕业设计(论文)外文资料翻译院 ： 专 业： 姓 名： 学 号： (用外文写)外文出处： 附 件： 1.外文资料翻译译文；2.外文原文。 指导教师评语： 签名： 年 月 日附件1：外文资料翻译译文智能停车辅助系统的系统配置智能停车场管理系统采用先进技术和高度自动化的机电设备，将机械、电子计算机和自控设备以及智能IC卡技术有机地结合起来，通过电脑管理可实现车辆出入管理、自动存储数据等功能，实现脱机运行并提供种高效管理服务的系统。 新型的智能停车场将生活理念和建筑艺术、信息技术、计算机电子技术等现代高科技完美结合，提供的是一种操作简单、使用方便、功能先进的人性化系统。它依靠高科技，以人为本，采用图形人机界面操作方式，提供一种更加安全、舒适、方便、快捷和开放的智能化、信息化生活空间，促进了人文环境的健康发展。 本文介绍了目前开发的投资促进机构配置（智能停车辅助系统）。IPAS允许司机指定目标位置的三个免费方法:基于单眼视觉的停车位置标记识别,基于超声波传感器的停车位置识别,和拖放GUI(图形用户界面)。IPAS生成最优路径马赫指定的目标位置。在停车场的运作，并估计自我车辆姿势使用ESP（电子稳定程序）的传感器，如车轮需要银行脚踏开关和传感器，转向角传感器。IPAS自动控制制动和转向通过发送所需的促动ESP和轨迹通过可以EPS(电动助力转向)。IPAS通知当前驱动器通过对停车操作后视图的图像轨迹估计，这是通过实验验证了系统的车辆。关键词智能停车辅助系统，驾驶员辅助系统介绍我们都由六部分组成:自我车辆姿态估计，路径生成器和流浪汉；跟踪，主动制动系统，主动转向系统，和HMI（人机界面）。IPAS既可进行半自动的停车辅助系统，其中的转向操作自动化。而且自动泊车辅助系统，其中转向和制动操作自动化。指定目标位置定位的自动/半自动停车操作的目标位置。我们开发了三个互补的方法：基于单目视觉的停车插槽标记识别，基于超声传感器的平行泊车插槽识别，和拖放的GUI（图形用户界面）。自我车辆姿态估计实现通过利用各种传感器，包括轮速传感器的车辆姿态估计的阿克曼模型，转向角度传感器，制动踏板开关，和车轮角度传感器。这些传感器连接到ESP（或，ESC）和EPS和报告他们的测量通过能。轨迹发生器使最优路径到达指定的目标位置。路径跟踪控制转向和制动驱动使估计的自我车辆姿态按照规划好的路径。在停车时，IPAS显示后视影像通过人机界面驱动和重叠轨迹上的估计。目标位置的三维定位我们的三维目标定位位置的自动方法和手动1支持2的方法。基于单目视觉的停车插槽标记识别指定的停车插槽标记垂直停车的情况下的目标位置。超声波传感器为基础的平行泊车插槽识别指定的情况下，并行的目标位置在向前和向后的停车情况。当司机得到目标位置使用自动的方法，可以调整或人工方法细化目标位置。当然，如果目标自动定位方法不可用，用户可以指定目标位置的手动方法，我，如拖放的概念为基础的GUI。虽然有研究发现目标位置的识别三维信息的相邻车辆 1 2 3 ，这种方法被认为是昂贵的第一代。他们有望被应用于下一代IPAS。基于单目视觉的停车插槽标记识别 4 基于单目视觉的停车插槽标记识别使用单眼的CMOS摄像头安装在车辆的后端显示2。视觉系统采用了广角镜头盖宽视野（视野）。它对输入图像进行补偿，广角镜头的径向畸变，并将失真的图像在鸟瞰图像。鸟瞰图像是一种虚拟的图像由一个飞行的鸟了。视觉系统开始识别程序从种子点，这是由与触摸屏驱动程序指定。检测正峰识别标记线段所示。边缘跟踪可以拒绝虚假标志着我一段和完善该线段的方向。精致的线段的线段之间，靠近相机是公认的准则，我是边境之间的道路和停车位。一旦指标是公认的，种子点的投影指标成为参考点。通过对标记强度和非标记强度之间的强度比，“T”形连接可以很容易地检测到如图所示。因此，指定的停车插槽可以被识别。提出的方法效果很好，尽管在相邻车辆由于它只使用指南和T形路口。超声波传感器为基础的平行泊车插槽识别 5 在平行泊车的情况，主要是有趣的障碍，即停放的车辆，可以从摄像机覆盖。距离测量的超声波传感器的制备，平行停车的驾驶过程中收集的，可以指定免费停车场。O1和O2对象集群和F1和F2免费空间集群。在聚类，提取了最近的自由空间聚类的两个端点位置：P1和P2。相邻对象clursters相邻的终点，即C1和C2，提供免费空间的外边界，属车辆姿态估计消除了相对于相邻车辆的相对位置的影响。拖放RMT基于目标位置的设计 6 由于手工方法应该是在自动方式尽管发展是必然的，研究的重点应在效率和方便的改进。为了验证所提出的人工方法的效率，我们测量的操作时间和点击数，然后比较拖放多箭法为基础的方法。车库停车，据观察，手术时间17.6%和点击次数减少减少64.2%。平行停车，据观察，手术时间29.4%和点击次数减少75.1% 6 下降。路径规划与跟踪目标位置的定位后，停车场控制器的半自动导航路径生成停车。车库停车状况规划路径基本上由两个系：当前车辆的前后方向延伸的线。线延伸的目标位置的中心线。圆形部分的半径是最小的旋转半径安装转向系统连接两线描述过渡部分。目前全自动泊车辅助系统不考虑多转停车试验情况呢。系统检查是否可以跟踪规划路径的操作开始前的停车场。参考点的车辆运动被认为是位于后桥的中心。当系统引导车辆停放,操作模式R-EPS从正常的力量协助模式变为“主动转向控制”模式。在汽车发展到目标位置,转向角是由R-EPS控制。实验结果验证提出的系统配置的可行性。我们进行了各种各样的车辆测试。估计路径测量轨迹几乎是一样的车库停车情况,分别平行泊车的情况。引用意味着测量轨迹和估计方法估计车辆构成的轨迹。结论在本文中。我们提出了一个IPAS系统配置，其中包括基于停车插槽标记识别的单目视觉，超声波传感器为基础的空闲时隙检测和拖放的手工设计方法的概念。通过可行性测试，我们可以发现，所提出的配置，大大提高驾驶方便只有少量的额外成本。良好的企业形象是企业无价的无形资产，智能停车场系统现代化的高科技产品的使用无疑会提高企业的物业管理形象和知名度，采用电脑收费管理系统，无论从产品的造型方面，还是整套系统所带来的先进性及管理的科学性，都将给物业管理树立起良好的形象，使企业步人良性循环，不断开拓进取，成为科学管理的楷模。 产品只是帮助人类更快捷、更方便、更简单地解决问题。智能停车场系统作为一种先进的产品最终是为人服务，它不能脱离以人为本的原则，否则技术再先进也是一种浪费。因此停车场技术的应用也应因地制宜，充分发挥停车场的所有功能。参考文献1 Nico Kaempchen等（2002）基于立体视觉的停车场使用的3D车辆模型的姿态估计。2002届IEEE智能车辆，卷2，页17-21。2 K. Fintzel 等. (2004). 立体停车辅助系统。2004届IEEE智能车辆，881-886页。 3 HG Jung等（2006）为自由停车位置识别的三维视觉系统。国际汽车技术学报，7卷，3号，361367页。4Ho Gi Jung等（2006）结构分析为基础的停车插槽标记为半自动停车系统识别。联合系统国际研讨会上的结构，在模式识别的句法模式识别和统计技术（S + SSPR 2006）。5 Dong Suk Kim等(2006)长距离的超声波传感器平行停车场的检测，2006 ksae春季会议。第三，第1621-1626（韩国）。6 Ho Gi Jung,等（2006）半自动泊车辅助系统的新用户界面。2006 FISTA汽车世界大会附件2：外文原文SYSTEM CONFIGURATION OF INTELLIGENT PARKINGASSISTANT SYSTEMIntelligent parking management system uses advanced technology and highly automated mechanical and electrical equipment, mechanical, and electronic computer-controlled equipment, and intelligent IC card technology combine management can be achieved through the computer management of vehicle access, automatic data storage function, the realization of de machine operation and provide - efficient management of the system. Smart new car park will be living concepts and architecture, information technology, computer electronics technology, such as the perfect combination of modern high-tech to provide a simple, easy-to-use, advanced system of human nature. It relies on high-tech, people-oriented, graphical man-machine interface operation, providing a more safe, comfortable, convenient, fast and open intelligent, information-based living space, promoting the healthy development of human environment.This paper describes the configuration of currently developed IPAS （Intelligent Parking Assistant System). IPAS allows driver to designate target position by three complimentary methods:monocular vision based parking slot marking recognition,ultrasonic sensor based parallel parking slot recognition，and drag&drop GUI (Graphical User Interface). IPAS generates an optimal path to mach the designated target position. During parking operations, IPAS estimates ego-vehicle pose using ESP (Electric Stability Program) sensors such as wheel need sensors, banking pedal switch and steering angle sensor. IPAS automatically controls braking and steering to achieve planned trajectory by sending required actuations to ESP and EPS (Electric Power Steering) via CAN.IPAS informs driver of the on-going parking operation by showing estimated trajectory on the rear view image. Proposed system is validated by vehicle experimentsKEYWORDSIntelligent Parking Assistant System, Driver Assistant SystemINTRODUCTIONI Our IPAS consists of six components. target position localization. ego-vehicle pose estimation, path generator & tracker, active braking system, active steer ing system, and ITMI (Human Machine Interface) as depicted. IPAS can implement not only semi-automaticparking assistant system, in which steering operation is automated. but also automatic parking assist system,in which steering and braking operation s automated. Target position localization designates the target position of automatic/semi-automatic parking operation. We developed three complimentary methods: monocular vision based parking slot marking recognition, ultrasonic sensor based parallel parking slot recognition, and drag&drop GUI (Graphical User Interface). Ego-vehicle pose estimation implements Ackerman model based vehicle pose estimation by utilizing various sensors including Wheel speed sensors, steering angle sensor, braking pedal switch, and wheel angle sensor. These sensors are attached to ESP(or, ESC) and EPS and report their measurements via CAN. Path generator makes an optimal path to reach the designated target position. Path tracker controls steering and braking actuation to make the estimated ego-vehicle pose follow the planned path. During parking operation, IPAS shows rear-view image to driver through the HMI and overlaps estimated trajectory on it. Figure 2(a) shows vision and ultrasonic sensor installed on our experimental vehicle. Figure 2(b) shows touch screen based HMI installed at indoor audio-console position.3D Localization of Target PositionOur 3D localization of target position supports 2 automatic methods and 1 manual method as shown in Figure 3. Monocular vision based parking slot marking recognition designates target position in case of perpendicular parking situation with parking slot markings. Ultrasonic sensor based parallel parking slot recognition designates target position in case of parallel parking situation with both forward and backward vehicle. Once driver gets target position using automatic method, driver can adjust or refine the target position by manual method. Naturally, if automatic target position localization methods are not available, user can designate target position by manual method, i,e. drag&drop concept based GUI. Although there have been researches to detect target position by recognizing 3D information of adjacent vehicles 123, such kind of methods are supposed to be expensive for the first generation. They are expected to be applied to the next generation IPAS.MONOCULAR VISION BASED PARKING SLOT MARKING RECOGNITION 4Monocular vision based parking slot marking recognition uses a monocular CMOS camera installed on the backend of vehicle as shown in 2(a). The vision system uses a wideangle lens to cover wide FOV (Field Of View). It rectifies input image to compensate the radial distortion of wide-angle lens, and then transforms the undistorted image into birds eye view image. Birds eye view image is a kind of virtual image taken by a flying bird.Vision system starts recognition procedure from the seed point, which is designated by driver with touch screen.Detecting positive peaks recognizes marking line-segments as shown. Edge following can reject false marking I ine-segments and refine the direction of line-segment . Among the refined line-segments, the line-segment nearest to the camera is recognized as guideline, vvhich is the border between roadway and parking slots. Once guideline is recognized, the projection of seed point to the guideline becomes reference point. By investigating the intensity ratio between on-marking intensity and off-marking intensity, T-shape junction can be easily detected as shown.Consequently, designated parking slot can be recognized .Proposed method works well in spite of adjacent vehicles thanks to the fact that it uses only guideline and T-shape junctionsULTRASONIC SENSOR BASED PARALLEL PARKING SLOT RECOGNITION 5In parallel parking situation, primarily interesting obstacles, i.e. parked vehicles, can be out of camers coverage. Distance measurements, collected during parallel parking preparation driving by ultmsonic sensor, can designate free parking slot. O1 and O2 are object clusters and F1 and F2 are free-space clusters. After the clustering, two end positions of the nearest free-space cluster are extracted: p1 and p2. Adjacent end points of neighboring object clursters, i.e. c 1 and c2, provide the outer border of free space,l. Vehicle pose estimator eliminates the effect of relative position with respect to adjacent vehicles.DRAG&DROP RMT BASED TARGET POSITION DESIGNATION 6Captured scene of parking site is displayed on small touch screen monitor and current target position is displayed. The inside region of rectangular target position is used as a moving cursor. The outside region of rectangular target position is used as a rotating cursor. User can easily designate the target position by the moving and rotating cursor .Manual target designation method has two usages. First, it can refine the target position acquired by automatic target localization methods. Second. it can be used at anytime driver want to manually designate target position without automatic methods.Because manual method is supposed to be inevitable in spite of the development of automatic methods, research focus should bc on the improvement of efficiency and convenience. To verify the efficiency of the proposed manual method, we measure the operation time and clicking number, and then compare drag&drop based method with multiple-arrow based method. For garage parking, it is observed that operation time is reduced by 17.6% and clicking number is reduced by 64.2%. For parallel parking, it is observed that operation time is reduced by 29.4% and clicking number is reduced by 75.1 % 6.PATH PLANNING AND TRACKINGAfter the localization of target position, parking controller generates parking path for semi- automatic guidance. Planned path of garage parking situation basically consists of two lines；line extending current vehicles rear direction. line extending the central line of target position. A circular segment whose radius is the smallest rotating radius of installed steering system connects the transition part of two lines as depicted .Current semi-automatic parking assist system does not consider the multi-turn parking trial situation yet. System checks whether it can track the planned path before the start of parking operation.The reference point of the vehicle motion is assumed to be located at the centre of rear axle. When the system guides vehicle to park, the operation mode of R-EPS changes from normal power ssist mode to active steering control mode. During vehicle progresses to the target position,steering angle is controlled by the R-EPS. Parking control ler receives vehicle travel pulse from transmission and steering anglc from R-E-PS. The steering angle conunand at the moment is determined by rcfcrring to the vchiclc travel steering angle profile. With a serial I ink such its CAN channel, parking connoller sends steering command ter R-EPS controllcr. which performs angular position control. Although longitudinal control of vehiclc is drivers responsibility, visual and acoustic guidance informs the driver of over-speed warning and the completion of parking operation. Hard limitation of vchiclc spccd can be imposed by ESP active banking control. During parking operation, vehicle trajectory Jiredictcd by current steering angle is projected onto rear-vicv image to inform driver of ongoing parking situation.EXPERIMENTAL RESULTTo validate the feasibility of proposed system configuration. we conducted various kinds of vehicle tests. Measured trajectories are almost the same with estimated path in case of garage parking situation and parallel parking situation respectively. Reference means measured trajectory and estimated means the trajectory by estimated vehicle pose.CONCLUSIONIn this paper，we propose an IPAS system configuration, which incorporates monocular vision based parking slot marking recognition, ultrasonic sensor based free slot detection and drag&drop concept based manual designation method. Through feasibility test, we can find out that proposed configuration enormously enhance drivers convenience with only small additional cost.Good corporate image is priceless intangible assets, smart parking system, the modernization of the use of high-tech products will no doubt enhance the image and visibility of property management, fee management computer system, in terms of product design, or package system brought about by the nature and management of scientific, management will have to establish a good image, so that virtuous circle of companies who