【电信学】【2007.03】未知位置目标的光学跟踪导航
本文为美国俄亥俄州立大学（作者：Dustin P. Bates）的硕士论文，共104页。
近年来，智能地面车辆（IGV）、无人机（UAV）和行星际航天器的发展极大地增加了对鲁棒、精确和自主导航系统的需求。这些系统需要实时的位置、航向和速度信息，以最小的误差在广谱环境中运行，以便成功地完成所设计的复杂任务。全球定位系统（GPS）已经满足了这一需求，为全球覆盖提供了米级精度。不幸的是，全球定位系统并不是完美的，它有一些缺点，破坏了它在满足某些系统需求方面的应用。树叶、城市景观（通常被称为“城市峡谷”）或室内环境对信号的障碍会阻断GPS卫星发出的信号，如果不将这些障碍全部消除，将大大降低解决方案的精度。考虑到来自卫星的信号低于噪声电平，它们也很容易受到干扰，无论是否出于有意。由于IGV、无人机和行星际航天器极有可能处于这样的环境中，因此导航解决方案中必须包含其他传感器，这些传感器不具有GPS和其他类似系统的缺点。惯性测量单元（IMU）在这些环境中工作，但随着时间的推移会产生较大的漂移误差。激光测距扫描仪，通常被称为激光雷达，用来解决这些问题的历史已经超过二十年。使用激光雷达的主要好处是低传感器噪声、自主性、同时用于避障、绘制周围环境的能力以及限制IMU漂移误差的可能性。虽然在利用激光雷达进行定位和制图方面已经做了大量工作，但是还没有形成一个功能完善的高完整性导航解决方案。本文不仅对二维激光雷达导航的一些主要组成部分和方法进行了阐述和改进，而且还将提出全新的方法和数学解决方案，以改进其作为导航系统的应用。这些改进中最大的是在对象匹配/跟踪和协方差计算中发现的。
In recent years, the development ofintelligent ground vehicles (IGV), unmanned aerial vehicles (UAV) andinterplanetary spacecraft has drastically increased the need for robust,precise and autonomous navigation systems. These systems need real timeposition, heading, and velocity information in a broad spectrum of environmentswith minimal error in order to successfully accomplish the complex tasks whichthey have been designed for. The Global Positioning System (GPS) has come tofulfill a vast amount of this need, providing global coverage with meter-levelaccuracy. Unfortunately, GPS is not perfect and has a few shortcomings whichundermine its use in fulfilling the requirements that some systems demand.Minor obstruction of the signal by foliage, cityscape (typically called an“urban canyon”), or indoor environments block signals from the GPS satellites,greatly reducing the accuracy of the solution if not eliminating it alltogether. Considering that the signals from the satellites are below the noisefloor, they are also very susceptible to interference, whether intentional ornot. Since it is highly probable that IGV, UAV, and interplanetary spacecraftwill be in such environments, there must be other sensors involved in thenavigational solution that do not have the shortcomings of GPS and othersimilar systems. Inertial Measurement Units (IMUs) work in these environments,but have large drift errors over time. Laser ranging scanners, typically calledLADAR, have been employed to solve aspects of these problems in various degreesof complexity and integration for over twenty years. The major benefits ofusing LADAR are low sensor noise, autonomy, its simultaneous use for obstacleavoidance, its ability to map its surroundings, and potential to bound thedrift error of an IMU. Although much work has been done in localization andmapping using LADAR, the problem of creating a fully functional high integritynavigation solution has not been achieved. This thesis will not only addressand improve some of the major components and methods of 2D LADAR navigation,but will also propose entirely new methods and mathematical solutions toimprove its use as a navigational system. The largest of these improvements arefound in object matching/tracking, and covariance calculations.
- 引言
- 项目背景
- 系统设计
- 测试结果与分析
- 结论与未来工作展望
更多精彩文章请关注公众号：